Methane is a potent greenhouse gas, and its accurate detection is critical for addressing global climate change. Although remote sensing has been a crucial technique for understanding the spatial distribution and temporal dynamics of methane emissions, it is still urgently needed that automate the identification of methane emission plume and effectively deconvolve the signal from background noise. In this study, we propose the Methane Plume Segmentation UNet (MPSUNet) to achieve precise segmentation of methane plumes from remote sensing imagery. MPSUNet incorporates the Pyramid Squeeze Attention (PSA) module to enhance feature representation and employs a joint loss function combining Dice Loss and Focal Loss to address challenges such as class imbalance and noisy data. A novel dataset, MPDataset, was constructed using EMIT methane enhancement and RGB radiance data, providing 4172 high-quality samples for model training and evaluation. Our results show that MPSUNet achieves a mean intersection over union (MIoU) of 78.20%, mean precision of 80.78%, recall of 71.11%, and mean pixel accuracy (MPA) of 85.41% on the complete four-channel MPDataset. Compared with seven classical segmentation models, the most improvents of MPSUNet in MIoU, MPrecision, Recall and MPA reach up to 5.33%, 12.28%, 18.04% and 8.94%, respectively. Notably, the integration of RGB channels enhances the segmentation of small and intricate plume structures. Cross-dataset evaluation using Sentinel-2 data further validates the model’s robustness, achieving an MIoU of 77.65% and an MPA of 83.61%. Generally, the proposed MPSUNet model marks a substantial performance in methane detection, which provides a robust technical framework for global-scale methane emission monitoring as emphasized by global climate agreements.

Nitrogen dioxide (NO2) is a critical air pollutant with significant health and environmental implications, particularly in urban areas where high levels of emissions are prevalent. Accurate nowcasting of ground-level NO2 concentrations is essential for effective air quality management and timely public health interventions. Traditional methods often struggle with balancing the spatial accuracy of ensemble learning models and the temporal forecasting strengths of time-series models like long short-term memory (LSTM) networks. In this study, we propose an improved hybrid framework, GC-LSTM, to nowcast regional ground-level NO2 concentrations on an hourly scale based on satellite-derived NO2 vertical column densities (VCDs), meteorological data, and on-site observations. GC-LSTM integrates the spatial learning capabilities of grained cascade forest (gcForest) with the temporal prediction strengths of LSTM networks, leveraging the strengths of both spatial inference and time-series prediction. This study focuses on the Beijing-Tianjin–Hebei (BTH) region, one of China’s most polluted areas, as a case study. Our results indicate that the GC-LSTM framework performs a strong correlation between predicted and observed ground-level NO2 concentrations, with an$R^{2}$of 0.746 and a mean absolute percentage error (MAPE) of 18.4% at a 1-h prediction interval. Even as the prediction intervals extended to 2 and 3 h, the GC-LSTM consistently outperforms the gcForest model across all evaluated metrics, with$R^{2}$values higher by 0.097 and 0.117, and root mean square error (RMSE) values lower by 0.666 and$1.76~\mu \text {g/m}^{3}$than those nowcasted by using the standalone gcForest model, respectively, highlighting its robustness and adaptability. Furthermore, the capacity of the GC-LSTM framework for continual learning and adaptation ensures its effectiveness in dynamic environments, making it a valuable tool for real-time air quality forecasting and environmental management.

The multiangle polarimetric (MAP) measurement provides abundant information about aerosol microphysical properties, but its physical retrieval methods of aerosols usually rely on time-consuming optimal iterative calculations. This study introduces a robust and efficient MAP aerosol retrieval over eastern China based on a data-driven deep learning (DL) method. By directly training the function relationship between Polarization and Directionality of the Earth’s Reflectances (POLDER) measurements and matched aerosol products in typical Aerosol Robotic Network (AERONET) sites with the deep belief network (DBN) methods, aerosol optical depth (AOD), fine mode AOD (FAOD), coarse mode AOD (CAOD), and single scattering albedo (SSA) can be retrieved reliably. Ground validation shows very high accuracy for POLDER-3 DBN AOD (${R} = 0.917$) and FAOD (${R} = 0.942$) compared with AERONET results. Despite a decrease in retrieval accuracy, DBN CAOD and spectral SSA exhibit very consistent variations with ground inversions. In particular, POLDER-3 DBN retrievals over eastern China perform better than generalized retrieval of aerosol and surface properties (GRASP) products with optimized method. Our results demonstrate that DBN can well model the complex functional relationships between MAP measurements and aerosol optical/microphysical parameters. With the striking advantage in computational efficiency and modeling ability, the DL methods, such as DBN, have an enormous potential in operational aerosol retrieval of the emerging MAP satellite instruments.

Accurate satellite-derived aerosol optical depth (AOD) with high temporal resolution is crucial for monitoring diurnal aerosol variations and understanding their impacts on atmospheric processes and air quality. The Advanced Geostationary Radiation Imager (AGRI) aboard the Fengyun 4A (FY-4A) satellite offers high spatiotemporal resolution, making it suitable for continuous atmospheric aerosol monitoring. In this study, an improved AOD retrieval algorithm is proposed for FY-4A/AGRI based on the generalized retrieval of atmosphere and surface properties (GRASP) framework. The algorithm incorporates multitemporal and multispectral FY-4A/AGRI observations within a 30-min window to enhance observational constraints for AOD retrieval. Reasonable prior information from Moderate Resolution Imaging Spectroradiometer (MODIS) bidirectional reflectance distribution function (BRDF) products and Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) aerosol components is introduced, enabling hourly AOD retrievals with high accuracy and robustness. Compared with AOD derived from the single-temporal strategy with fixed BRDF and aerosol models, results of validation against aerosol robotic network (AERONET) AOD measurements over Beijing-Tianjin–Hebei (BTH) region indicate that our improved FY-4A/AGRI AOD retrievals increase the R from 0.543 to 0.864, and reduce root-mean-square error (RMSE) from 0.149 to 0.09, with the percentage of data falling within the expected error (EE) range rising from 46.1% to 69.9%. In Asia, such advancements led to significant improvements in AOD retrieval performance in 2021, with validation results demonstrating a strong correlation ($R =0.826$for hourly retrievals and$R =0.891$for daily means) and high accuracy (RMSE =0.118 for hourly retrievals and RMSE =0.09 for daily means) against ground-based AOD measurements from 32 AERONET sites. Comparative analyses reveal that FY-4A/AGRI AOD retrievals outperform Himawari-8/AHI products and are comparable to MODIS multiangle implementation of atmospheric correction (MAIAC) data, particularly in capturing diurnal variations and spatial distributions of aerosols. The algorithm also demonstrates robustness across diverse land cover types and vegetation densities. Our AOD retrieval strategy provides a scalable approach for geostationary satellite aerosol retrieval, with implications for regional air quality monitoring and climate studies.

Precisely measuring the concentration profile of ozone (O3) in the troposphere is a necessary prerequisite for studying its climatic and environmental effects and for effectively preventing photochemical pollution. Currently, thermal infrared satellite sensors mainly utilize the absorption characteristics of ozone near 9.6 μm to retrieve the O3profile. However, the retrieval of the O3profile is an ill-posed problem, and it is necessary to introduce a priori ozone information as a constraint to expand the solvable domain of the underdetermined problem. Therefore, the a priori ozone profile is crucial for the retrieval accuracy. Moreover, because tropospheric O₃ accounts for only about 10% of the total atmospheric ozone, the detectable signals and information content from thermal infrared observations are extremely limited, making it necessary to enhance the degrees of freedom and accuracy of tropospheric ozone retrievals. In this study, tropospheric ozone profiles are retrieved using the Cross-track Infrared Sounder (CrIS) thermal infrared hyperspectral imager aboard the Suomi-NPP satellite. By comparing ozone profiles obtained from a Long Short-Term Memory (LSTM) model, the Empirical Orthogonal Function (EOF) method, the ERA5 ozone reanalysis data, and the radiosonde data, the results show that the ozone profiles obtained from the model established by the LSTM-based a priori are closer to the true vertical distribution of tropospheric ozone. To further improve retrieval accuracy, the total ozone column amount from the Ozone Mapping and Profiler Suite (OMPS) ultraviolet payload on the same satellite platform is incorporated to redefine the cost function within the optimal estimation framework. The retrieval model is verified using the ozone radiosonde data from the World Ozone and Ultraviolet Radiation Data Centre (WOUDC), and results are compared with retrievals that exclude the OMPS constraint. The results indicate that the retrieval results with the constraint of the total ozone column amount have a smaller relative error compared to the retrieval results without the constraint.

In our study, we present bifurcated attention, a method developed for language model inference in single-context batch sampling contexts. This approach aims to reduce redundant memory IO costs, a significant factor in latency for high batch sizes and long context lengths. Bifurcated attention achieves this by dividing the attention mechanism during incremental decoding into two distinct GEMM operations, focusing on the KV cache from prefill and the decoding process. This method ensures precise computation and maintains the usual computational load (FLOPs) of standard attention mechanisms, but with reduced memory IO. Bifurcated attention is also compatible with multi-query attention mechanism known for reduced memory IO for KV cache, further enabling higher batch size and context length. The resulting efficiency leads to lower latency, improving suitability for real-time applications, e.g., enabling massively-parallel answer generation without substantially increasing latency, enhancing performance when integrated with post-processing techniques such as reranking.

Powerful foundation models, including large language models (LLMs), with Transformer architectures have ushered in a new era of Generative AI across various industries. Industry and research community have witnessed a large number of new applications, based on those foundation models. Such applications include question and answer, customer services, image and video generation, and code completions, among others. However, as the number of model parameters reaches to hundreds of billions, their deployment incurs prohibitive inference costs and high latency in real-world scenarios. As a result, the demand for cost-effective and fast inference using AI accelerators is ever more higher. To this end, our tutorial offers a comprehensive discussion on complementary inference optimization techniques using AI accelerators. Beginning with an overview of basic Transformer architectures and deep learning system frameworks, we deep dive into system optimization techniques for fast and memory-efficient attention computations and discuss how they can be implemented efficiently on AI accelerators. Next, we describe architectural elements that are key for fast transformer inference. Finally, we examine various model compression and fast decoding strategies in the same context.

The Multi-angle Imaging SpectroRadiometer (MISR) measurement with a large range of scattering angles provides valuable information about aerosol microphysical properties. The current MISR algorithm utilizes pre-defined aerosol mixtures in lookup tables (LUT) to infer aerosol types and microphysical parameters, which performs well globally but remains subject to considerable uncertainties in regional scales. To make efficient use of MISR measurement, we developed a physics-informed Deep Learning (PDL) method to retrieve aerosol optical/microphysical parameters over land in eastern China. By combining the physical constraint of radiative transfer simulation and modeling ability of DL methods, each aerosol parameter can be modeled with the whole used MISR measurements separately with high computational efficiency. PDL Aerosol Optical Depth (AOD) and fine AOD(FAOD) have high correlation coefficients (R>0.95) with Aerosol Robotic Network (AERONET) observations, with 89% and 81% values falling into expected error (EE) envelope of ± (0.05+20%AODAERONET) respectively. Despite only a slightly higher accuracy than recent MISR Version 23 products, PDL retrievals have solved the underestimation problem of AOD and FAOD at moderate-high values (>0.4). Besides better constraint of abnormal values in coarse AOD(CAOD), PDL algorithm significantly improves retrieval accuracy of MISR Single Scattering Albedo (SSA). With reliable and robust performance, PDL algorithm provides a flexible and efficient aerosol retrieval framework for emerging multi-angle polarimetric measurements.

In weakly supervised semantic segmentation (WSSS) tasks on remote sensing images, it is a common practice to train a classification network from scratch using a large batch of images with a limited number of classes. Subsequently, class activation maps are extracted from the model based on predefined class indices, and these maps are then optimized to obtain pseudolabels. To make this strategy effective when introducing a new class, a substantial amount of data needs to be provided to the model. In this article, we present an innovative framework, RS-TextWS-Seg, designed to efficiently generate high-quality segmentation results for a wide range of remote sensing objects using concise descriptions. Our proposed framework comprises three sequential stages: initially, we undertake parameter fine-tuning of the contrastive language-image pretraining (CLIP) model to swiftly strengthen its capacity for zero-shot detection of a limited number of remote sensing features. Subsequently, we introduce a text-driven background suppression mechanism aimed at deriving class activation maps from the refined CLIP model based on textual cues, while concurrently mitigating background noises. Finally, we use the segment anything model (SAM) to refine the edges of the extracted class activation map. We widely researched the leading-edge methodologies in WSSS and conducted a range of comparative experiments and ablation studies to prove the efficacy of our proposed framework. The research findings underscore that RS-TextWS-Seg outperforms other state-of-the-art methods on renowned datasets such as DLRSD and Potsdam, as well as on bespoke datasets specifically curated for overground petroleum pipelines and oil well fields.

Semantic segmentation of high-resolution remote sensing images (HRSIs) is a challenging task because objects in HRSIs usually have great scale variance and appearance variance. Although deep convolutional neural networks (DCNNs) have been widely applied in the semantic segmentation of HRSIs, they have inherent limitations in capturing global context. Attention mechanisms and transformer can effectively model long-range dependencies, but they often result in high computational costs when being applied to process HRSIs. In this article, an encoder-decoder network (MSGCNet) is proposed to fully and efficiently model multiscale context and long-range dependencies of HRSIs. Specifically, the multiscale interaction (MSI) module employs an efficient cross-attention to facilitate interaction among multiscale features of the encoder, which bridges the semantic gap between high- and low-level features and introduces more scale information to the network. In order to efficiently model long-range dependencies in both spatial and channel dimensions, the transformer-based decoder block (TBDB) implements window-based efficient multihead self-attention (W-EMSA) and enables interactions cross windows. Furthermore, to further integrate the global context generated by TBDB, the scale-aware fusion (SAF) module is proposed to deeply supervise the decoder, which iteratively fuses hierarchical features through spatial attention. As demonstrated by both quantitative and qualitative experimental results on two publicly available datasets, the proposed MSGCNet exhibits superior performance compared to currently popular methods. The code will be available athttp://github.com/JingxiangZhou/MSGCNet.

Heavy-polluting enterprises burn fossil fuels to release large amounts of greenhouse gases, causing severe pollution worldwide. Heavy-polluting enterprises have a significant responsibility for carbon emissions, and more than 130 countries have set or are considering targets for achieving net-zero carbon emissions by 2050. Assessing these enterprises can provide data support for carbon emissions and aid in evaluating industry’s economic development. In view of the problem that the existing research data is not comprehensive and the generalisation ability is week. To address this issue, we construct a high-resolution remote sensing image dataset of global heavy-polluting enterprises and use the classic target detection network SSD, Faster R-CNN and YOLOv3 for training, testing and evaluation. The experimental results findings indicate that the SSD network is particularly well-suited for object detection of heavy-polluting enterprises in the remote sensing domain.

The number and geographical location of oil well sites can reflect the local oil production situation and there is a growing interest in automatically identifying oil well sites from remote sensing images. Traditionally, visual interpretation was employed to extract oil well sites locations from remotely sensing images. However, this approach is time-consuming and heavily dependent on domain experts. Advancements in remote sensing satellite technology and the widespread use of deep learning algorithms have enabled the automated extraction of oil well sites from remote sensing images. In this paper, we established the Northeast Petroleum University Oil Well Sites Dataset Version 1.0 (NEPU-OWS V1.0), and to evaluate its usability by comparing several different deep learning models based on semantic segmentation algorithms for optical remote sensing images. Experimental results show that current advanced deep learning models achieve high accuracy on this dataset, demonstrating great potential for remote sensing detection in oil well sites.

Most of the existing object detection methods have complicated hand-designed components, such as non-maximum suppression procedures and manual resizing of anchor boxes. Based on DETR, this paper not only eliminates the need for manual component adjustment, but also solves three problems of poor remote sensing image for directional object capture, slow DETR convergence, and the same attention allocated by different layers of Decoder. First, the D-Angle module is used to align the rotating object region while accelerating the convergence using the a priori angle. Then the overall computation of the model is reduced by using Adaptive Proposal Selection(APS) in the cascade structure. Finally, the Adaptive Query Selection(AQS) module is applied so that Decoder in different layers get different attention weights to optimize the layer-by-layer fine-tuning process. In this paper, the effectiveness of the proposed method is verified using two public datasets, DOTA and HRSC2016.

The multi-angle polarimetric (MAP) instruments have been a focus of recent satellite missions dedicated to enhanced detection of global aerosol microphysical properties. Considering that satellite observations can hardly infer all the unknowns of atmosphere and surface, it’s crucial to know how many and which aerosol parameters can be accurately retrieved from these different MAP measurements as well as their uncertainties. In this study, we present a comprehensive insight into the information content of POLDER-3 and 3MI observations for aerosol retrievals and estimate posterior errors of corresponding parameters based on Bayesian theory. The total degree of freedom for signal (DFS) of aerosol retrievals is around 6-8 from POLDER-3, and is raised by ~1.8-3.5 with 3MI. The retrieval accuracy of volume concentration and effective radius are high (<4%) in the fine-dominant case for both POLDER-3 and 3MI, but get much lower (~8% and ~15%) in coarse-dominant conditions. Furthermore, the advanced 3MI measurements can upgrade the retrieval uncertainties of POLDER-3 by ~50%. Though additional shortwave infrared bands of 3MI provide more information regarding coarse particles, the influence of aerosols on surface BRDF leads to a decrease of the total DFS. With a prior assumption that variations of refractive index depending on wavelength, satellite retrieval accuracy of the real (<0.03) and imaginary part (<0.003) reaches close levels with that of ground-based Sun photometers. Our results can provide a fundamental reference for MAP satellite retrieval of aerosol microphysical properties.

Clouds play an important role in the radiative energy balance of the Earth–atmosphere system. Compared with traditional optical satellite sensors, polarimetric sensors combine multi-angle, multi-polarization, and multispectral information, displaying the advantages of high spatial and temporal resolutions and global coverage. Such remote sensing measurements improve the accuracy of cloud properties retrieval. Due to the observation characteristics of passive satellites, even a tiny variation in position will result in a great change in the observation geometry. A large number of studies have shown that the scattering angle is very crucial for the polarization characteristics retrieval of reflected light. In this study, we analyze the dependence of the remote sensing retrieval implement of different cloud characteristics on the observed scattering angle coverage, considering both ice and water clouds. Three satellite sensors – POLarization and Directionality of the Earth’s Reflectance-3/Polarization and Anisotropy of Reflectance for Atmospheric Sciences coupled with Observations from a Lidar (POLDER-3/PARASOL), Directional Polarimetric Camera/ GaoFen-5 spacecraft (DPC/GF-5), and DPC/GF-5(02) – were selected to compare their scattering angle coverages and the number of angular measurements at equatorial, middle, and high latitudes. The requirements for angular polarized and nonpolarized observations varied depending on the retrieval of cloud properties. The impact of orbital characteristics and viewing settings was investigated for cloud detection, cloud phase classification, and cloud microphysical properties retrieval. Finally, an analytical model to comprehensively evaluate the effective angular measurements according to the orbital characteristics and viewing settings was developed to facilitate the future design of similar sensors for cloud remote sensing.

With the continuous growth of remote sensing data and the application of fast and effective atmosphere remote sensing inversion algorithm, this paper proposes a PM10 fast inversion approach based on Spark platform which uses Apache Spark as the analytics engine and integrates with the traditional atmospheric remote sensing inversion algorithm. We first store aerosol data which is MYD04_3K from NASA into HDFS. Then the inversion algorithm is combined with Spark via the function interface to realise rapid atmospheric remote sensing inversion. The experimental results based on Spark platform are compared with those obtained from the traditional physical hardware. The results prove that the proposed atmospheric remote sensing inversion method based on Spark has high efficiency.

The protection and monitoring of forest resources has drawn considerable national attention. Traditional deforestation monitoring requires a lot of manpower and material resources through manual visual interpretation and manual change patterns labelling, which has problems of low efficiency and high missed alarm rate. Therefore, this paper explores the detection for deforestation changes from remote sensing images based on deep learning framework, and aims to help forestry department manage and monitor forest resources. In this paper, an U-Net+LSTM framework is used to detect the changes of deforestation from remote sensing images. The evaluation data is Sentinel-2 dataset and the study area is Guangxi Sanjiang Dong Autonomous County in China. The results show that the F1 score of the framework is as high as 0.715, which proves the proposed model can effectively detect the change from forest to bare soil in remote sensing images.

Continuous spatial knowledge is required to control the regional ozone pollution. Measurements from ground-level sites are beneficial to this goal, but their number is limited due to the huge expenses of site establishment, operation, and maintenance. Remote sensing seems a promising data source, but its application is challenged by bad weather conditions. Always covered by thick clouds, Chongqing, a populated industrial city in west China, is facing serious ozone pollution, but relevant studies here are relatively insufficient. Another alternative is estimating ozone by models. Well-performed models degrade in Chongqing partially due to the very complex terrain. Modeled hourly ozone does not agree with ground-level measurements. Therefore, an optimization approach is proposed to improve model estimates for such regions. This approach integrates the ground-level information (e.g., measured ozone and meteorology) through the employment of ResNet (Residual Network). ResNet overcomes the notorious vanishing gradient issue in classic neural networks, and the ability of learning complex systems is largely boosted. Ozone distribution is like a gray image that varies every second, which is not the case usually learned by ResNet. A color-image alike data structure is raised to address this “nonstill image” problem; according to the Taylor Expansion, polynomials can describe a complex system, and the errors are acceptable. To facilitate the usage in business operations, this approach is designed to be robust, inexpensive, and easy to use. The scheme of control site selection is discussed in detail. In cross-validations, this approach performs well, averaged$R^{2}$is higher than 0.9 and the error is less than$5 ~\mu \text {g/m}^{3}$.

In this article, we developed a hybrid method to estimate surface shortwave radiation (SSR) for the new-generation Himawari-8 geostationary satellite. This hybrid method combines the advantages of a deep neural network (DNN) with high speed and radiative transfer model (RTM) to achieve high accuracy: the RTM provides training data for the DNN under various cloud and aerosol conditions (including heavy aerosol loadings). Moreover, our hybrid method can simultaneously output the byproducts of photosynthetically active radiation (PAR), ultraviolet A (UVA), and Ultraviolet B (UVB), the direct and diffuse components at the surface, and the upward solar radiation at the top-of-atmosphere (TOA). The trained DNN was applied to the Himawari-8 satellite atmospheric products for 2016 and comprehensively validated using a total of 118 stations from four networks located in the full-disk regions of Himawari-8. The results showed an RMSE of 125.9 Wm-2for instantaneous SSR, 105.4 Wm-2for hourly SSR, 31.9 Wm-2for daily SSR, and respective mean bias error (MBE) scores of 8.1, 27.6, and 12.3 Wm-2. The hybrid method developed in this study performed well, achieving high accuracy and high speed, and it is capable of providing near-real-time SSR estimates for many applied energy fields.

GF 5 is a hyperspectral imaging satellite which is configured with six forms of payloads. EMI is the first high-resolution imaging spectrometer used for the detection of atmospheric trace gases, the indicator of environmental pollution. The high spatial-temporal pollutants distribution information collected by EMI will assist government's decision-making and evaluation of air quality guarding in the future.

The prediction of PM2.5 concentration is a canonical predictive challenge due to the distribution of $PM_{2.5}$ appears serious spatiotemporal variability at multiple scales. Currently, using satellite-based remote sensing data to estimate ground-level PM2.5 is a promising method for providing spatiotemporal continuous information of PM2.5. In this letter, we proposed a deep neural network (DNN)-based PM2.5 prediction model to capture the spatiotemporal variability of ground-level PM2.5 using the remote sensing aerosol optical depth (AOD) data from the Himawari-8 satellite along with the conventional meteorological observation variables (denoted as PM25-DNN). The PM25-DNN model was trained and tested using the data from Beijing-Tianjin-Hebei region of China in 2017, and we compared the prediction performance between the PM25-DNN and the current state-of-the-art methods in this field. The results show that the PM25-DNN outperforms the other models with the cross-validated coefficient of determination ($\text{R}^{2}$ ), root-mean-square error (RMSE), mean prediction error (MPE), and relative prediction error (RPE) were 0.84, $19.9~\mu \text{g}/\text{m}^{3}$ , $11.89~\mu \text{g}/\text{m}^{3}$ , and 41.21%, respectively. Then, the trained PM25-DNN model was applied to estimate the hourly gridded PM2.5 with 1-km spatial resolution. Our results indicate that the DNN architecture can capture the essential spatiotemporal distribution associated with PM2.5 only using AOD data and conventional meteorological observational variables without more handcrafted features. The proposed PM25-DNN model can greatly improve the accuracy of PM2.5 estimation, and it provides a new perspective for PM2.5 monitoring using end-to-end deep learning method.

In the autumn and winter, aerosol is the important atmospheric pollutant over the Beijing-Tianjin-Hebei region. For monitoring aerosol in the autumn and winter, the lack of vegetation and the aging of MODIS sensor are two problems that needed to be solved. In this paper, after analyzing the characteristics of aerosol radiance in the red and shortwave infrared (2.12 μm) bands of MODIS, we develop a new algorithm for terrestrial aerosol with the assumption that the reflectance ratio between the red and 2.12 μm bands is invariant. With MODIS data over the Beijing-Tianjin-Hebei region from September 2016 to February 2017, the algorithm is applied to aerosol retrieval. The retrieved aerosol optical depth images show that our algorithm can retrieve aerosol over sparse vegetation, and the validation with the AERONET/PHOTONS Beijing site shows that the correlation is greater than 0.9% and 77% of the retrievals fall within the expected error. An error analysis shows that a 2% error in the proportion of the soot component can lead to 15% retrieval error, and over more than 60% of the surface area, the error from the changes in the ratio between the red and 2.12 μm bands can lead to retrieved errors less than 0.1.

As the demand for enabling high-level autonomous driving has increased in recent years and visual perception is one of the critical features to enable fully autonomous driving, in this paper, we introduce an efficient approach for simultaneous object detection, depth estimation and pixel-level semantic segmentation using a shared convolutional architecture. The proposed network model, which we named Driving Scene Perception Network (DSPNet), uses multi-level feature maps and multi-task learning to improve the accuracy and efficiency of object detection, depth estimation and image segmentation tasks from a single input image. Hence, the resulting network model uses less than 850 MiB of GPU memory and achieves 14.0 fps on NVIDIA GeForce GTX 1080 with a 1024 512 input image, and both precision and efficiency have been improved over combination of single tasks.

During the last decades, large increase in anthropogenic emissions has led to severe air pollution problems in China, with high concentration of fine particles and widespread haze layers in many areas. The complex sources and high emissions of atmospheric pollutants has exerted great challenge on air quality in China. Compared with regular measurements in ground sites, satellite observations can provide a unique view of the amounts of atmospheric components and formation processes of haze pollution from regional to global scales. Considering the special atmospheric conditions of high aerosol loading and large spatial and temporal variations in China, we made several improvements such as identification of haze areas in the retrieval of aerosol loading. In particular, we conducted comprehensive investigations in optical properties, spatial variation, and formation processes of the regional haze pollution of China using integrated satellite observations, ground measurements, and meteorological data.

Global climate change is one of the most challenging issues facing the world today. Atmospheric aerosols and carbon dioxide (CO2), as two key factors driving the global climate change, have earned enormous attention from scientist around the world [1]. One challenge for the satellite measurements of CO2using this SWIR wavelength range (∼1.6µm) is the impact of multiple scattering by aerosols and cirrus [2]. Since the rapid economic growth and associated increase in fossil fuel consumption have caused serious particulate pollution in many regions of China [3], remote sensing of CO2using SWIR band in China needs to pay more attention to the scattering properties of aerosol particles and the multiple scattering. Considering the complexity of morphological and chemical properties, aerosol particles are grouped based on a large number of TEM/SEM images, and then their scattering properties at 1.6µm band are calculated by the T-matrix method [4] and GMM method [5]. In this study, the Monte Carlo method is used to solve the multiple scattering problem by simulating photons transport in the scattering media. We combined this multiple scattering model with the LBLRTM [6] as a forward radiative transfer model for studying the impact of aerosol scattering on the satellite observations of CO2using SWIR band. Finally, based on the GOCART aerosol component products, AERONET aerosol size distribution products, CALIPSO aerosol profile products, and MODIS aerosol optical depth and surface albedo products, the monthly variability of errors in CO2concentrations over China were calculated and analyzed. The results indicate that AOD and surface albedo are two of most important factors for the satellite observations of CO2. For low surface albedo, the retrieved CO2columns are undervalued when aerosol scattering is neglected. While for moderate and high surface albedos, the retrieved CO2columns are overvalued. As shown in Fighre 1, CO2concentrations are overestimated in western regions of China, especially in desert areas (a maximum of ∼7.08% in September), and those are underestimated in eastern regions (a minimum of ∼−6.9% in June).

The cloudbow structure is directly related to the retrieval of cloud droplet size distribution (droplet effective radius and effective variance). This study investigated the effect of the cloud optical thickness, ground surface albedo and the above-cloud absorbing dust layer on the cloudbow structure based on the modeled airborne directional polarimetric camera (DPC) measurements, which are simulated in 670 nm using Mie scattering theory and the vector radiative transfer mode. It is found that the polarized reflectance increase as the increase of the cloud optical thickness (COT) and saturate when COT=10. The absorbing dust layer's signal would cover the signal from the cloud layer as the aerosol optical thickness increased to 1. Additionally, the surface albedo has negligible effect on the cloudbow structure.

Launching in October 2011, the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument has been stable operating for more than 5 years on board the Suomi National Polar-orbiting Partnership (S-NPP) spacecraft. Some researchers indicated that its aerosol optical thickness (AOT) product had larger biases than MODIS over land. Although the VIIRS-AOT algorithm is based on MODIS Dark-Target algorithm, some differs exist, including cloud mask. In considering the algorithm independence and identification complexity, we develop a new quick cloud mask algorithm for aerosol retrieval. Based on the spatial variability test inherent from MODIS, we add a expand test to remove the pixels mixed with cloud or effected by neighboring pixels. The results illustrate that this new test can screen out the confident cloudy pixels that VIIRS algorithm regard as clear sky. Throughout hundreds test in different weather conditions, the algorithm perform well.

Nitrogen dioxide (NO2), sulfur dioxide (SO2), and smoke are major pollutants, which are used to evaluate the air quality. This study developed a dual-phase air quality monitoring system to monitor the air quality, which based on the Shuffled Complex Evolution algorithm (SCE-UA), ground-based AQI data and satellite observations of NO2, SO2, and Aerosol Optical Depth (AOD). The system is implemented in two phases: the optimization of model coefficients and the air quality index (AQI) simulation. A comprehensive evaluation system of the air quality was then established. The model coefficients of the AQI regression model are optimized by the SCE-UA algorithm in the optimization phase, and the optimized coefficients are used as the final model coefficients in the AQI simulation phase. The experimental results indicate that the SCE-UA algorithm can effectively optimize the coefficients of the AQI regression model. It provides a promising solution to monitor the air quality through using the satellite observations and optimizing model coefficients.

An improved Optimal Estimation (OE) method is presented for methane (CH4) column density retrieval from satellite observations in short-wave infrared band (SWIR), to avoid non-convergence of iteration process for CH4retrieval caused by the singularity or non-positivity of the Hessian matrix. We add a constraining factor γ and a step factor α to the OE iteration algorithm. Then, total column averaged CH4dry air mole fraction, XCH4is retrieved using GOSAT Level 1b data. Retrievals are validated by comparisons with ground-based FTIR measurements from TCCON stations. Comparison shows good agreement and the correlation coefficient is more than 0.55. Preliminary validations approve the utility of proposed retrieval algorithm.

This paper presents an algorithm for the retrieval of ozone profiles from the Cross-track Infrared Sounder (CrIS) aboard the Suomi National Polar-orbiting Partnership satellite using a nonlinear optimal estimation method. The issue of channel selection is discussed. Based on a sensitivity analysis, we selected a spectral range of 990-1070 cm-1for the ozone profile retrievals. Compared with ERA-Interim ozone profiles and eigenvector regression method profiles, the ozone climatology profile is better able to construct the a priori state. The retrieved CrIS profiles are in good agreement with smoothed high-vertical-resolution ozonesonde profiles. An analysis of the information content of the CrIS retrievals demonstrates that the CrIS measurements can provide useful information for capturing the spatial and temporal variations in ozone and are insensitive to ozone below 400 hPa. An error analysis revealed that smoothing error represents the main error source for the retrieved CrIS ozone profiles.

China's industrialized regions have seen increasing occurrence of heavy haze caused by severe particle pollution. However, aerosol retrieval under these circumstances is often excluded from NASA's Moderate Resolution Imaging Spectrometer (MODIS) aerosol products due to cloud mask and suspected high surface reflectance. An algorithm to retrieve the haze aerosol optical thickness (HAOT) is developed using MODIS data to supplement the current MODIS retrieval algorithm. This method includes 1) haze identification, 2) the generation of a surface reflectance database using MODIS data in hazy conditions, and 3) the development of haze aerosol models with four aerosol components simulated by a global 3-D atmospheric chemical transport model (GEOS-Chem). This algorithm was used in combination with the MODIS dense dark vegetation algorithm to retrieve 1 km HAOT over North China Plain from March to September of 2008. The values of the retrieved HAOT values are mostly between 0.7–3, with a correlation coefficient of 0.82 with the Aerosol Robotic NETwork observations and a 19% mean relative difference. Retrieval uncertainties associated with the errors in haze detection, surface reflectance, and haze models were analyzed using ground measurements.

There are four CCD cameras with spatial resolution of 30 m in Environment Satellite 1 (HJ-1), the new satellite developed by China. In the paper, deep blue algorithm for CCD/HJ-1 is applied to Ning-bo. Based on the database of surface reflectance from MODerate-resolution Imaging Spectroradiometer (MODIS) spectral reflectance product and look-up table (LUT), aerosol optical depth (AOD) over cloud-free land pixel is retrieved from apparent reflectance in the first band of CCD/HJ-1. AODs over Ning-bo area are retrieved from January to September in 2011, and the results are validated by ground-based measurements of CE318 in the center of Ning-bo city. The validation shows that the retrieved AODs are larger than that of ground-based measurements, but correlation coefficient (R) is greater than 0.7. Further improvements of overestimated AODs is the focus of our future research.

In this paper, we explored the changes of air quality over Jing-Jin-Tang (Beijing-Tianjin-Tangshan) district during the period from 2002 to 2009. Based on Moderate Resolution Imaging Spectroradiometer (MODIS) data, Dense Dark Vegetation (DDV) algorithm is employed to retrieve the aerosol optical thickness (AOT) with 1-km resolution. Comparison of the satellite inferred AOT and the values from ground-based Aerosol Robotic Network (AERONET) sun/sky radiometer measurements indicates a good agreement (R2=0.786) in Beijing site. We compared the spatial, monthly and annual variation over Jing-Jin-Tang district and analyzed the main factors of these changes. Our study indicates that there is a decreasing trend in the annual variation of AOT since 2004. The averages of AOT were commonly higher in spring and summer than those in autumn and winter, and the retrieved AOT over cities and southern areas is obviously larger than that over rural and northern areas respectively.

The Ring effect refers to the filling in of Fraunhofer lines, which is known as solar absorption lines, caused almost entirely by Rotational Raman scattering. The Rotational Raman scattering by N2and O2in the atmosphere is the main factor that leads to Ring effect. Basically, the Ring effect is considered as a pseudo-absorption process in retrieval of trace gas constituents in atmosphere. The solar spectrum measured by OMI/AURA is convolved with rotational Raman cross sections of N2and O2, divided by the original solar spectrum, with a cubic polynomial subtracted off, to create a differential Ring spectrum. This method has been suggested in order to obtain an effective differential Ring cross-section for the DOAS fitting process. The differential Ring spectrum could be used to improve the accuracy of the retrieval of the trace gases concentration. The results have been in basic agreement with the corresponding results calculated with RTM, and the R2statistic is 0.9663.Next, the differential Ring spectrum calculated with rotational Raman cross sections of atmosphere in the fixed wavelength of 410nm and 488nm are derived. The results with the fixed wavelength have been also in basic agreement with the corresponding results calculated with RTM, and the R2statistics are 0.9624 and 0.9639 respectively. At last but not the least, the computational complexity calculated at fixed wavelength of 410nm or of 488nm is 0.128% of that calculated with wavelengths from 410nm to 488nm.

AOTs(Aerosol optical thicknesses) and aerosol size distribution functions in Pearl River Delta area are derived from PARASOL (Polarization and Anisotropy of Reflectances for Atmospheric Science coupled with Observations from a LIDAR)multi-directional, multi-spectral polarized signals. Based on analyzing the products of AERONET(Aerosol Robotic Network), aerosol size distribution function and complex refractive index over Pearl River Delta area are received. After that, the particular aerosol model is abstracted. The land surface polarized contribution is calculated using semi empirical model as a function of surface type and NDVI, and the pure atmospheric contribution is computed with a radiative transfer code. Compared with the products of the ground-based AERONET, the derived AOTs are underestimated against AERONET measurement. The retrieved size distribution for the radii bigger than 0.2 micron is also underestimated due to that polarization is insensitive to coarse model aerosol.

Nitrogen dioxide (NO2) plays a very important role among the anthropogenic trace gases. The tropospheric NO2vertical column density (VCD) maps derived have been used to study many scientific applications, pollution emissions and pollutant distribution. During the 2008 Summer Olympic Games in Beijing, NO2is one main air pollutant which should be monitored. This paper presents the NO2inverse algorithm, the Differential Optical Absorption Spectroscopy (DOAS), from satellite measurements and the results using this method. The results show 1) the tropospheric NO2VCD in Beijing is about the same as that in other cities nearby in June, 2008; 2) from July 1, the tropospheric NO2VCD in Beijing decreases significantly, however, it changes little in other cities nearby; 3) the tropospheric NO2VCD in Beijing increases a little in August, 2008, which is much lower than that in other cities around, such as Tianjin, Tangshan.

Light use efficiency model is one of the method to retrieval regional scale Gross Prime Productivity (GPP). Absorbed Photosynthetic Active Radiation (APAR) and Light use efficiency are the main parameters of this kind of model. At the same time, light use efficiency is affected by air temperature and precipitation. In this article, one of the Light use efficiency model is used to retrieval daily GPP of the Chinese five typical ecosystem experimental station in 2003. The inversion results are compared with MOIDS NPP product and station measurement data. Based on the different air temperature and precipitation condition of the different station, it also analyses the sensitivity of parameters.

Operational global quantitative retrievals of aerosol have been made from Moderate Resolution Imaging Spectrometer (MODIS) data for several years by NASA EOS teams. Dark Dense Vegetation (DDV) algorithm has shown excellent competence at aerosol distribution and properties over land. In Sep. 2008, China successfully launched environmental satellite and received Charge Coupled Device (CCD) sensor data, it will provide a new way to monitor aerosol optical thickness (AOT) and distribution at a higher resolution (30 m * 30 m). According to DDV algorithm and HJ-1-CCD camera characters, we measured different surface reflectance spectra in Beijing and Pearl River Delta areas, then ascertained NDVI value and the surface reflectance radio between HJ-1-CCD red and blue band. This paper introduces the aerosol retrieval process including lookup table establishing, cloud detection and so on; Finally, the retrieved AOTs were validated by ground measurement and compared by MODIS aerosol products.

On the eve of Beijing Olympic Games (BOG), frequent haze days had been extensively concerned by home and abroad. Ground-based and satellite remote sensing project was carried out to monitor haze distribution and intension by Chinese Academy of Sciences. Based on the assumption that surface reflectance vary slowly in a relative short period, the Haze Optical Thickness (HOT) retrieval model using MODIS data is established. Aerosol type is selected according to the ground experiment of haze particle composing in Central North China Plain. This model avoid that NASA Dense Dark Vegetation (DDV) algorithm couldn't determine the radio of surface reflectance between middle-IR and visible channel on haze day. From Jun. 1st to Sep. 30th, 2008, AOT observed by sun photometer (CE318) on the ground was used to validate HOT, and it had shown good coherence.

In this paper, the retrieval of aerosol over land surfaces from CCD data of China Brazil Earth Resources Satellite (CBERS) 02B was studied. The method is dark dense vegetation (DDV) algorithm: 1) dense vegetation (dark pixel) was recognized by the NDVI threshold; 2) the look up table (LUT) of atmosphere was computed from the Satellite Signal in the Solar Spectrum (6S) model; 3) the aerosol was retrieved through interpolating the LUT by CCD data. The method was applied to Beijing area, and the retrieved aerosol was validated by ground-based measurements of CE318. The result shows that from CBERS02B data, the aerosol can be retrieved well.

Cirrus clouds are known to play a key role in the Earth's radiation budget and global climate change, the radiative effects of cirrus clouds depend critically on cloud properties such as optical thickness and particle shape and size. The studies of the optical, microphysical, and physical properties of cirrus have become the popular issue. This paper simulated the bidirectional reflectance distribution function (BRDF) and bidirectional polarization reflectance distribution function (BPDF) at 1380 nm in cirrus cloudy conditions on the basis of an adding-doubling radiative transfer program. Based on the sensitivity of 1380 nm spectral reflectance and polarization reflectance on cirrus optical thickness and aspect ratio, a conceptual approach has been developed to simultaneous retrieve the particle shape and optical thickness of cirrus clouds using the remote sensing data of multi-angular total and polarized at 1380 nm.

Polarization space-borne sensor, just like POLDER (Polarization and Directionality of the Earth's Reflectances), is a new instrument devoted to the global observation of solar radiation reflected by the Earth surface-atmosphere system. It is necessary to acquire polarized information in retrieval of aerosol properties over land surface. Often the aerosol contribution is small compared to the surface, especially by covered vegetation. Atmospheric scattering is much more polarized than the surface reflectance. Using polarized information could solve the inverse problem of separating the surface and atmospheric scattering contributions. This paper presents retrieval of aerosols properties from multi-wavelength polarized measurements. The results suggest that it is feasible and possibility for discriminating the aerosol contribution from the surface in the aerosol retrieval procedure using multidirectional and multiwavelegth polarization measurements.

It is difficult to determinate the aerosol model when retrieving the aerosol over land surfaces. In this paper, from the products of AErosol RObotic NETwork (AERONET), the aerosol model in city area of Beijing was analyzed. Then, the aerosol properties was retrieved from polarized data of the Polarization and Anisotropy of Reflectances for Atmospheric Science coupled with Observations from a LIDAR (PARASOL), and validated by the AOD products of AERONET. The results show: (1) the aerosol model based on ground based observations in Beijing is different from standard aerosol model of POLDER; (2) the particular aerosol model over Beijing can improve the retrieval accuracy over Beijing.

Because of a series of complex environmental and instrumental effects, there are strip-pattern noises in images of many bands of MODIS, especially in that of Band 5. Several destriping algorithms have been used to remove the strips in MODIS images, but since they could not properly take account of physical or spectral relevance maintained in the data, the destriping results are hardly ideal. A new algorithm is proposed for the removal of strips in Band 5 images. It can locate the strips accurately so as to retain as much original information of the image as possible. Additionally, the new algorithm makes full use of the spectral correlation of multi-band remote sensed data, and can reconstruct the Band 5 value of strip-pixels based on appropriate interpolation. The validation and comparison with other algorithms indicate that, this new algorithm could remove strips perfectly and the reconstructed values have a small deviation of about 3% from the "true values".

Tropospheric aerosols play an important role in the Earth radiation budget directly through scattering and absorption of solar and infrared radiation, and indirectly by modifying cloud microphysical and radiative properties. Satellite remote sensing provides a means to derive aerosol distribution at global scales. The common operational algorithm to retrieve aerosol optical depth (AOD) is dark pixels method based on the atmospheric effect on the path radiance. However, when the land surface has a high reflectance, this method always becomes invalid. Another way was contrast reduction method (or structure function method) using the change in contrast for several scenes to determine the optical thickness between the scenes. China Brazil Earth Resources Satellite (CBERS) series carry high resolution CCD camera with a small scan degree, which is fit to retrieve AOD using contrast reduction method. CBERS02B was launched on 19thSeptember, 2007, and the CCD camera has been calibrated on board. This paper retrieves AOD from CBERS02B using contrast reduction method in Beijing urban area. The experiment result overestimates the AOD by about 0.2 while has the right tend.

This paper, we detect clouds in China regions from combination of POLarization and Directionality of the Earth's Reflectances(POLDER) data and Moderate Resolution Imaging Spectroradiometer(MODIS) data, based on the spectral, multi- angular, and polarized characteristics of cloud. Four tests are applied to the measurements. The first one is blue channel reflectance test. The second one is the test on polarization at 865 nm. The third one is the test on reflectance at 1380 nm. The fourth one is the test on reflectance at 645 nm and 1640 nm. At last, the performance of method is evaluated using a large dataset of surface face observations of cloud cover. The result demonstrate the methods of cloud detection are feasible and believable..

In this paper, based on the normalized polarized radiances at top of atmosphere (TOA) have been simulated with different viewing geometry (along-track viewing angle and viewing angle number), the aerosol optical depths (AOD) have been retrieved. The retrieved AODs have been used to analyze the best viewing geometry for the new multi-angle polarized camera design. The results show that the accuracy of AOD is better as the viewing angle number is getting more, but for alongtrack viewing angle, the accuracy of AOD does not increase while the along-track viewing angle range changes from ±40° to ±60°.

An atmospheric correction algorithm for ocean color data with multiple viewing and polarization information is proposed. The correction is based on using the directional and polarized data for 865 nm and 665 nm to estimate the properties of aerosols over the ocean. The aerosol models used in atmospheric correction consist of bimodal size distributions. And, a best-fit optical thickness is grossly determined using the generated look up table of the upwelling radiation. Moreover, Validation of the improved algorithm with the standard POLDER atmospheric correction algorithm is given.

This Paper described that in-flight radiometric calibration for thermal channels of TERRA-MODIS sensors using a new calibration test site-Dali-lake, China. The radiance of water surface was measured by CE312, and the spectral transmittance and upward radiance of the atmosphere was calculated using radiance transfer model MODTRAN4. At the same time the spectral response of Satellite sensor and that of ground-based sensor are coupled. At last the apparent radiance of sensor spectral channels is compared to the digital count of satellite's output to give the calibration coefficient. The calibration result in May 31 showed the difference between inflight and on-board calibration was equivalent to a brightness temperature of 1.44 k for TERRA-MODIS channel 31 and 0.35 K for channel 32 respectively.

The ear is the top layer of mature wheat and has very different geometric and thermal characteristics from that of leaves. Compared to the directional brightness temperature (DBT) of wheat canopy without ears, the DBT at the ear stage has specific features, and the ear effects could not be explained by previous models. This paper proposes a hybrid geometric optical and radiative transfer model to reveal the combined influences of the geometric structure of ears and leaf; the temperature distribution of ear, leaf, and soil; and the Sun-target-sensor geometry on the canopy DBT. The soil, leaf, and ear layers are taken into account in the model so it is named as the Soil Leaf Ear Combined (SLEC) DBT model. We compare the model prediction with the field measurement data. The results show that the new SLEC DBT model can simulate the DBT of wheat at the ear stage with an accuracy of 0.78 K.

A high-precision method for mapping the fraction of wheat Area in Luancheng County, North China Plain is presented and validated. The method is based on a spectral mixture analysis model and a optimal temporal TM image election where six main endmembers (greenhouse , soil, wheat, roof, nursery garden, shadow), constitute the observed pixel reflectance from the satellite. Given the reflectance observation, the fractional wheat area(FWA) is solved from the spectral mixture analysis(SMA) model. The high-precision estimation for six endmember SMA can be expected at an optional temporal range, which is the last ten days of March month every year. The simple landcover in March month specifies the effect of best simulation on the FWA of TM image pixels while no disturbed vegetations have grown green leaves. This approach enables operational wheat planting area mapping for extensive areas with an almost 95% classification precision to sub-pixel level. Our study area covers 346 km2. comprising whole Luancheng County of Hebei Province. Applying SMA to Landsat/TM cloud-free data acquired on March 21th, 2004, we estimated the areal fraction of wheat cover for the whole County field. The validation against statistical data from the Luancheng County Statistic Bureau indicates that with finite SMA, 92.3% accuracy is gained. Better results were also obtained from the validation against statistical information, for example, 95.8% of wheat-covered area were recognized when heavy haze area-free TM image is applied. A general formula for deriving the fractional wheat planting area mapping provided by SMA is presented, too. It can be concluded that wheat planting area to sub-pixel level in NCP can been operatively monitored with limited endmember SMA in a good precision.

Modeling the soil component temperature distribution is useful to study multi-angular thermal remote sensing. SVAT (soil-plant-atmosphere transfer) model could be a good choice because it can predict canopy temperature distribution. However, most of them, including CUPID model 111. were unable to separate shade soil and sunlit soil. They only gave a single temperature for the soil surface. In this paper, based on the difference of net radiance and evaporation rate between the shade and sunlit soil, an extended model from CUPID was proposed to simultaneously retrieve the shaded temperature and sunlit temperature of soil surface. The comparison showed good agreement between simulated soil temperatures and measured ones.

Dust storms greatly affect the environment and the resources in arid and semi-arid areas. In this paper, we study the process of an Asian dust storm in the north of China in 2005 by means of Terra and Aqua MODIS data, which could monitor dust storms both in daytime and nighttime. We use three thermal infrared (TIR), 8.5 mum, 11 mum, and 12 mum, and a tri-spectral technique presented by Ackerman (1997). We find that combining brightness temperature differences BT8.5- BT11 with negative BT11- BT12 can detect dust storms in China. Furthermore, the pre-processing of images, such as cloud screening, is important for dust storm monitoring. Compared to daytime visible and near infrared RGB images, the TIR method predicted dust storm very well and could also obtain useful information in nighttime. We also analyzed dust storm motion by four MODIS images within 24 hours. Our results showed that the dust storm motions from inter land to ocean and relies on the main wind direction. This information is very important for studying the processes of dust storm.

In this paper, a synthetic strategy has been employed to model 3D canopy's radiation transfer in the whole optical spectral domains. 3D plant architecture model (the Clumped Architecture Model of Plants: CLAMP) (1) is used to generate the realistic vegetation scene. In the visible and NIR region, the canopy BRDF was decomposed into three parts: single scattering contribution from leaves, single scattering contribution from the soil, and multiple scattering part of the canopy. The single scattering contributions come from illuminated leaves and soil components which are computed by the reverse ray-tracing procedure (2) with their corresponding reflectance. The multiple scattering contribution is approximated by the four-stream theory. As a result, the modeling of VNIR region is more efficient and fairly accurately describes the anisotropically scattering features of vegetation. In the TIR region, the directional brightness temperature of canopy is calculated as the linear combination of four component's (illuminated leaves, illuminated ground, shadowed leaves, and shadowed ground) brightness temperature multiplied by its fractional cover computed by the reverse ray-tracing procedure. Initial modeling results show typical features of vegetation's anisotropic scattering and directional temperature distributions, for example, hot spot, bowl shape and reach a good agreement with theoretical results in those three domains. This strategy shows potential of exploring the impact of canopy structure on the radiometric response measured by remote sensors.

Although the development of a spectral library has been a hot topic in China and abroad since the '90s, it has defects and cannot meet the demands of theoretical research and application of remote sensing nowadays. The gap is further extended for the featured agricultural remote sensing application in South China. Aiming at establishing a practical spectral library of South China's featured crops (including lichee, longan and sugarcane etc.), crop spectra and their environment parameters, application models are integrated based on Web techniques. The paper concentrates on the spectral data measurement method, store and organization, realization of querying and presentation and Web-interface design. It offers the details of regional featured crop spectra and an application demonstration for remote sensing application of spectral library of featured crops in South China.

Abstract -The spectra of maize field are of great uncertainty duo to the difference in planting date, irrigation condition, fertilizing and soil etc. The spectral library is a quick means acquiring the crop spectrum in a growth stage. The variance of image and in-situ maize spectra is analysized for extracting pure pixels of maize crop. It presents a good result for removing disturbed greensward, residential area, and nursery garden from variance based pure maize pixel classification. Keywords- Image endmember, spectral library, variance, maize, pure pixel, recognition Ⅰ. INTRODUCTION Remote sensing of the extent and distribution of individual crop types has proven useful to a wide range of end-users, including governments, farmers, and scientists (Qi-Jing Liu et al,2005). Maps of cropland distributions are usually generated by supervised classification of multiple Landsat images throughout the growing season. These approaches require amounts of manual interpretation and cloud-free high spatial resolution imagery that are prohibitive for operational implementation over large areas and in multiple years (David B. Lobell,2004). The NASA Moderate Resolution Imaging Spectroradiometer MODIS, including the daily global coverage, moderate spatial resolution (0.25 to 1 km), has rapid availability of various products, and cost-free status may allow for operational mapping of croplands. However, the large size of even the 250-m MODIS data relative to most fields results in MODIS pixels containing mixtures of different fields, crop types, and non-crop surfaces. As a result, approaches that assign a single hard classification to each pixel may be prone to significant errors when mapping crop types ( Fisher, 1997). If the pure spectral pixels of different crops can be gained in finer resolution, a big problem will be solved for mixture pixel unmixing and scaling reversion in quantitative remote sensing application by coarser resolution satellite , for example Landsat TM to Terra MODIS sensor. Maize is one of the most important crops in China. During maize growth with rain season of May to September month every year in North China Plain, it is low available for Landsat TM images. During maize growth of June to September, 2003, there is only a scene of TM image captured while the maize spectrum experiment of ground all-growth period is performing. It is necessary to research the maize spectrum characteristic of each month all over the growth period. Another, the every-day gained MODIS image is becoming an important data source of crop monitoring. The ground orchard, greensward and woodland (using in virescence of urban and road) usually disturb the maize classification by TM image. Therefore, it is of great meaning to extract pure maize pixels for improving the classification of the pure maize pixel recognition for maize condition monitoring and yield estimation by scaling method. In this study, we investigated the impact of greensward

It is feasible to retrieve LAI over large area from remote sensing data with physical models;however,it is quite difficult to get accurate LAI and thus limit the remote sensing application without enough prior knowledge due to the underdetermined parameters in the physical inversion models.A spectrum database system of typical objects in China(SpecLib) has been set up recently,which may provide a priori knowledge of typical land cover for LAI inversion.MODIS data is used to retrieve LAI after atmosphere correction,geometrical correction and cloud identification.The SAIL(Scattering by Arbitrarily Inclined Layers) model is applied for the inversion of LAI for MODIS data.The vegetation coverage of the mixed pixels of the MODIS data are calculated based on the TM data sets.The LAIs of pure pixels(computed from the retrieved LAIs and vegetation coverage) are compared with the field measurement data in Luancheng,Heibei Province,China.Meanwhile,the LAIs of pure pixels are also compared with the MODIS LAI data products.The inversion results show that the!SpecLib effectively improved the accuracy of leaf area index inversion.

Due to the failure of Landstat 7 ETM+, Landstat 5 images are widely used in many fields again since May 2003. In this letter, a method, based MODTRAN+TM+topographical maps, was used to solve the preprocessing problems of TM images towards retrieving of the endmembers. It proved effective to retrieve the reflectance values of surface substances and confirm the pure surface endmembers before an in-site spectral experiment. The result was also consistent with actual reflectance of surface substances. A case of retrieving endmembers in Luancheng, China was too demonstrated.

This work presents an application in which field data, TM data and MODIS data are used for mapping LAI of Qianyanzhou in Jiangxi province, south China. The field LAI data are collected by Tracing Radiation and Architecture of Canopies (TRAC). We get the linear relationship between measured LAI and simple ratio vegetation index (SR: the ratio of NIR reflectance and Red reflectance) from the corresponding pixels, then the LAI image from TM data is produced and is referred as LAI ground true for further studies. After comparing and analyzing the different LAI images, it shows that the uncertainties really existed among the different kinds of the LAI images with same resolution, it also reveals that the spatial up-scaling effects are closely related to sub-pixel complicacy.

Usually we use multi-channel image data, such as TM, and empirical relationship, such as NDVI-LAI relation or SR-LAI relation, to estimate LAI. Multi-angular remote sensing data provide more information for canopy structure. This paper presents a method to estimate LAI using multi-angular data and model inversion method. The airborne multi-angular data were acquired by AMTIS (Airborne Multi-angle TIR/VNIR Imaging System), which was a prototype sensor designed by the Institute of Remote Sensing Applications of Chinese Academy of Science. Our study is based on two datasets: one was acquired in Beijing Shunyi in April 11, and the major crop is sparse winter wheat; another was acquired in Haerbin in August 24, and major crops are dense corn and soybean. Both datasets have been geometrically atmospherically corrected. Ground based measurements were carried out during the flight experiment. SAIL model is chosen to predict reflected radiance of a presumed LAI. Various view angles relate to the different components ratio in view field, and the reflected radiance is different accordingly. Hence, a certain LAI value was given, SAIL model predicts a set of reflected radiances of various angles. We compare the model predict radiance with the radiance viewed by an multi-angular sensor, to find the optimized LAI which can make the radiance predicted by the model be closest to the viewed radiance, then take this LAI value as the right value

In this paper, a generalized split-window method to derive land surface temperature (LST) from MODIS (Moderate Resolution Imaging Spectroradiometer) data is applied. A major problem in land surface temperature inversion is that there are too many unknown variables, especially for MODIS data which is in low resolution, one pixel is a mixture of several cover types. To analysis the uncertainties of the LST retrieval algorithm based on MODIS images, the field measurements, together with fine resolution images, AMTIS (the airborne multi-angle TIR/VNIR imaging system) data and ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) data have been used

Vegetation indices are subjected to many external perturbations such as soil background variations, atmospheric conditions, geometric registration, and especially sensor viewing geometry. Subsequent use of these indices to estimate crop yield and monitor crops growth would result in substantial uncertainties. To reduce the uncertainties due to sun-view angle variations, some methods mere generated by use the reflectance or albedo generated from the BRDF models. MODIS vegetation composition algorithm uses the empirical BRDF model (developed by Walthall et al. to normalize the sun/view angles to certain angle, and then composite the VI by several day's data. In this paper, we present a new method based on prior knowledge to normalise vegetation index on pure pixels of crops, which can be recognized from MODIS image by high resolution land cover map. We simulated different BRDFs of winter wheat in different grow stages by radiative transfer models, using the plant canopy parameters obtained from prior knowledge. Then, we use this BRDF to normalize vegetation indices. The method was tested by the ground based measurements and MODIS Data. It shows our results are good consistent with the ground based measurements. We compare our methods with the algorithm of MODIS vegetation composition, it proved that the result calculated by our method is in better agreement with the surface reflectance characterizations and our method is more effective to monitor the crop growth in regional scale

The evapotranspiration of forestland surface was estimated using MODIS data. The study area is located in Jiangxi province, south China, dominated by tall indeciduous conifers. The parameters of land surface energy balance, i.e., leaf area index (LAI), land surface temperature (LST), surface albedo etc. were inversed from VIS/NIR and TIR band data of MODIS sensor. Some of the assistant data, such as meteorological factors and canopy structure parameters were obtained in-situ at Qianyanzhou ecological experiment station, which is one of the sites in the ecology observation system of Chinese Academy of Sciences. The radiation components and net radiation of surface were estimated with these parameters and ancillary data. Soil heat flux was estimated as a fraction of net radiation from the area coverage of the canopy. Surface sensible heat flux was decided according to surface temperature gradient and aerodynamic resistance. The excess-resistance for scalar flux transfer was also considered. Eventually, the latent heat flux (instantaneous evaporation rate) was obtained as the residual term of the surface energy balance equation. Daily evapotranspiration was then derived from the one-time-of-day estimation.

Land-cover classification represents one of the most fundamental applications of remote sensing, and is widely used to estimate carton stocks and parameters hydrological and biogeochemical models. Several studies reveal that changing the spatial resolution of land-cover maps has important effects on the proportion of a landscape occupied by a particular land cover type. We study the proportions of vegetations based on multi-scale land-cover classifications in the area of Qianyanzhou in the province of Jiangxi in the South China on the base of ground investigations. The viability of coarse spatial resolution data for land-cover classification is evaluated using degraded Landsat Thematic Mapper (TM). The uncertainties of multi-scale land cover classifications are finally analyzed based on the different aggregated TM land-cover maps.

A revised sunlit time computation model based on the DEM data in rough terrain has been developed in This work. In order to improve the accuracy of the calculation, an earth curvature revising factor Q was carried out in the model base on the former researchers. In this study, 6 sample areas of representative terrain types within P. R. China were selected and their spatial distributions of sunlit time on the vernal and autumnal equinoxes, the Summer Solstice and midwinter day in two different resolutions (500 m and 1000 m) were calculated using the model developed in the study. In This work, the calculated results of sunlit time using revised and unrevised models were finely compared and the topographic and spatial-scaling effects on the sunlit time were deeply analyzed through two aspects: the different terrain types and different spatial scales of the original DEM data. The changing rules of sunlit time according to the geomorphology and spatial scale were found out and very significative in applications.

Vegetation Index (VI) and Leaf Area Index (LAI) are very important parameters for crop growth situation monitoring and crop yield estimation. However, it is not easy to get accurate VI or LAI. One of the reasons is because of the difficulty in the inversion of aerosol optical depth (AOD), which is the key factor in atmospheric correction. This study addresses an algorithm of AOD inversion in Agriculture region based on the support of Spectrum database. It is usually supposed that we can get the surface reflectance of blue band for most of the algorithm of AOD inversion. As the Dense Dark Vegetation (DDV) method, the surface reflectance in blue and red bands is calculated from the reflectance at 2.1 or 3.8 mum band. However, it is not easy to retrieve the AOD value of each pixel for a whole satellite image because of the unknown surface reflectance on some regions such as the sparse vegetation area. For agriculture area, the surface reflectance varies from bare soil, sparse vegetation, and then Dense Dark Vegetation, during the whole crop growth period. We have carried out a series of field spectrum measurement during different crop growth period and set up a crop spectrum database. By analyzing the soil, the leaf and the canopy spectra, we selected a set of models to calculate the surface reflectance of agriculture region during different crop growth period, which include bare soil model, sparse vegetation model and continuous vegetation model. Then, the surface condition is put to the atmospheric radiation transfer model to calculate Look-up Table (LUT) for MODIS bands, which is used to retrieve the AOD value of MODIS image. North China Plain is selected as the experiment area, the AOD value measured by sun-photometer is taken as true value to evaluate the inversion algorithm's accuracy and the results show good agreement

A series of environmental and resource problems have merged due to the rapid urban development, including the encroachment on agricultural land, land reclamation and silt deposition in rivers. The paper has demonstrated that remote sensing can be used to monitor dynamic changes of the coastal areas, such as coastline move and urban expansion, land use, shoals and deep channels. Remote sensing data from 1978 to 1998 are used to reveal the rapid changes that have taken place in the study area. Geographic information systems are also used to assist planners in the analysis of such changes.

The purpose of this paper is to validate the BEPS model in crops for net primary productivity (NPP) estimation and to study the spatial scaling effects of NPP using both airborne and field data. The results show that the highest differences between modeled NPP at resolution 15 m and 30 m are greater than those re-sampled from modeled NPP at 3 m resolution, especially at the boundary of winter wheat.

Data experiment of Mont-Carlo directional radiation transfer model for continuous vegetation was made to decide the optimum view angle of thermal temperature for reliable estimation of sensible heat flux. The true heat fluxes were simulated with classical two-layer model. The conclusions of this study are: 1) the optimum view angle varies within a large range according to the change of leaf area index, leaf angle distribution, soil moisture and other parameters, so it's difficult to define a universal optimum angle; 2) however, the fractional coverage of vegetation in FOV (field of view) under optimum angle is relatively stable and could be used in a new corrective method.

In this paper, temperature/vegetation index relation is used to evaluate soil moisture conditions by 8 km, 10-day composite AVHRR data set. To define the T/NDVI slope, window size and automatic linear fit are studied. Then T/NDVI slope is analyzed spatially and temporally. Results show that soil moisture is closely correlated to the T/NDVI slope. A new approach is proposed to estimate soil moisture availability by T/NDVI slope and T/NDVI space. Finally, soil moisture and drought distribution maps are produced in China.

A couple-inversion algorithm that retrieves geophysical parameters from MODIS measurements was developed. The retrieved geophysical parameters include atmospheric temperature-humidity profiles, pixel-averaged surface temperature and emissivity within the thermal infrared regions (6/spl sim/16 /spl mu/m). The Modtran atmospheric radiative transfer code was used to simulate the measured radiances and atmospheric transmittance. Then the genetic algorithm was employed to generate a regularization solution that updates the first-guesses of atmospheric temperature, water-vapor profiles, surface skin temperature and emissivity. The algorithm proposed in this paper was first tested with simulated data.

In order to determine surface temperature at large scale from space, a new definition of effective emissivity has been proposed for the whole pixel area for heterogeneous and non-isothermal surfaces. This new effective emissivity depends on the structure of the whole pixel, the optical features of components in the pixel. In order to further illustrate the new effective emissivity, the continuous canopy vegetation is taken as an example, its approximate expression of effective emissivity has been studied by the aid of the Monte Carlo algorithm.

In order to evaluate and improve the remotely sensed land surface parameters' accuracy and assimilate accumulating remote sensing data with land surface models, an integrate field campaign was carried out during the winter wheat growth season in 2001. The field campaign was funded by the Chinese Special Funds for Major State Basic Research Project: Quantitative remote sensing theory and application of land surface parameters (QRSLSP). Research on scale effect and angular characteristics of remote sensing data is the most important objectivity. Three different scales of research area are responding to different spatial resolution of remote sensing sensors. The experimental database include space-borne remote sensing data, airborne remote sensing data and field observation data. Some significant research advances have been achieved based the experiment data set, while most research proposals are carrying on.