In the process of using digital canopy photography to measure Leaf Area Index (LAI), the estimation of canopy gap fraction (GF) is a critical factor that significantly influences LAI calculations. The accuracy of gap fraction estimates directly impacts the calculated LAI values. Typically, upward canopy images for gap fraction estimation are sensitive to exposure conditions. In this study, we conducted two experiments: the first experiment involved capturing canopy images with different lens focal lengths and relative exposures using a perforated test board with precisely known gap fractions; the second experiment employed a fixed-position fisheye lens to continuously capture a series of canopy images with varying shutter speeds. During the image processing, we utilized an adaptive 2D gamma adjustment algorithm based on luminance components to process the RAW images acquired by the camera during canopy photography. Our findings indicate that the algorithmically processed RAW images exhibit greater stablity to changes in exposure conditions when estimating canopy gap fraction compared to the JPEG images derived by the camera.This suggests the efficacy of the proposed algorithm in enhancing the reliability of gap fraction estimation in canopy photography under varying exposure conditions.

Nitrogen dioxide (NO2) is the main component of atmospheric pollution and mainly exists in the troposphere in the atmosphere. In this paper, NO2density in the atmospheric troposphere is monitored and analyzed based on the MAX-DOAS system, and the difference between the air mass factor (AMF) simulated by the geometrical approximation and the AMF simulated by the radiative transfer model SCIATRAN is analyzed in the inversion of the vertical column density (VCD) of NO2in the troposphere. The experiments show that at high elevation angles (e.g., 30°-60°), the AMF simulated by the geometric approximation is in good agreement with that simulated by SCIATRAN in the inversion of the tropospheric NO2VCD, but the geometric approximation ignores the influence of atmospheric complexity factors, and the relative deviation of the inverted tropospheric NO2VCD is larger.

Nitrogen dioxide(NO2), a common air pollutant, has received increasing attention for its harmful effects. At present, the monitoring of NO2by satellite remote sensing has insufficient poor real-time performance and low resolution. In this paper, a small atmospheric trace gas detection imaging spectrometer SWING(SWING, Small Whiskbroom Imager for atmospheric compositioN monitorinG) mounted on the UAV platform was used to carry out NO2emission monitoring research in thermal power plants, and the NO2differential slant column density was obtained based on the DOAS algorithm. Finally, the NO2regional distribution map was obtained after geocoding and statistical interpolation. The experiment successfully detected the main air pollution emission sources in the region, and established a method for airborne NO2imaging detector pollution source detection and verified its effectiveness. The comparison with Sentinel-5P shows that this method can fill the gap in the detection of small-scale pollution sources, and provides an effective means for the fine traceability of air pollution.

Leaf area index(LAI), defined as half of total leaf area per unit ground surface area, is a critical structural parameter. The objective of this paper is to validate MODIS LAI product(MOD15A2H) using decametric-scale sentinel-2 LAI and field measured LAI. The research was conducted in Yucheng station and the main crop types in Yucheng site are winter wheat and summer maize. After atmospheric correction and Snap software process, we obtain S2(Sentinel-2) 10m resolution LAI. The S2 10m LAI was validated by comparing with field measured LAI. The S2 10m LAI showed good agreement with field measured LAI(R2= 0.81, RMSE = 0.6). To convert the S2 10m LAI to the same 500m spacial resolution with MOD15A2H, we proposed a MODIS-Like upscaling method, the Sentinel-2 500m LAI was obtained. The MOD15A2H showed very good agreement with Sentinel-2 500m LAI(R2= 0.96, RMSE = 0.23). However, the comparison result of field measured LAI and MOD15A2H(R2= 0.68, RMSE = 0.58) was poor. Finally, we analyzed the time series of field measured LAI, sentinel-2 10m LAI, sentinel-2 500m LAI and MOD15A2H LAI.

According to reresearch, the corner detection-based threshold segmentation algorithm has been found to outperform the Otsu method in handling mixed pixels in canopy imaging. This paper aims to describe the principle of the corner detection-based threshold segmentation algorithm and compare its classification performance to that of the Otsu method through experiments. The results indicate that compared to the Otsu method, the corner detection-based threshold segmentation algorithm achieves higher accuracy in classifying mixed pixels, preserves more canopy information in overexposed areas of the image, effectively reduces the misclassification of mixed pixels, and greatly improves the accuracy of hemispherical photography leaf area index inversion results. Specifically, the correlation coefficient R2of the corner detection-based threshold segmentation algorithm is shown to increase from 0.8 to 0.9, demonstrating its superior performance.

In the indirect measurement of Leaf Area Index (LAI), previous studies commonly assumed that the actual LAI can be obtained by subtracting the Woody Area Index (WAI) from the Plant Area Index (PAI), which includes the entire plant. This can be calculated using the equation LAI = PAI - WAI. However, this study, conducted using 3D model simulations and the DHP (Digital Hemispherical Photography) method to measure LAI, discovered that this conclusion may not be correct. The LAI value obtained by subtracting WAI directly from PAI showed significant discrepancies from the true LAI value in the simulated environment. Through the re-derivation and verification of the LAI calculation equation, this study proposes a more reasonable explanation for this discrepancy and presents the following conclusions based on the 3D simulation experiment and hemispherical photography method: using PAI-WAI to calculate LAI leads to overestimation of the values, while the calculation method described in this study obtained more accurate LAI values in the simulation experiments (with higher R2, lower RMSE and MAE).

Studies have shown that camera auto-exposure underestimates the LAI (leaf area index) measured by DHP (digital hemispheric photography) to varying degrees. To address this problem, this paper proposes the use of multi-exposure fusion to reconstruct information from canopy images to compensate for the loss of information caused by overexposure or underexposure of canopy images acquired by the camera in auto-exposure mode. By fusing a series of canopy images with different exposure times from the same canopy layer, the LAI is then calculated using DHP on the fused image. Experimental results show that the method improves the R2from 0.698 to 0.837 and reduces the RMSE from 0.87 to 0.37 compared with the automatic exposure mode of the camera, with LAI-2200 measurements as a reference. This method contributes to resolving the problem of underestimating LAI in DHP caused by the automatic exposure mode, thereby improving the accuracy of DHP.

Quantifying the accuracy of burned area (BA) estimations is crucial in wildfire monitoring and loss assessment based on remote sensing technology. In this letter, a novel approach to quantitatively evaluate the accuracy of BA estimations, called the vector distance algorithm (VDA), was proposed based on boundary sampling and$t$tests. To validate the effectiveness of this approach, Sentinel-2 images were used to estimate the BA, and a field survey and a GaoFen-6 (GF-6) image were then utilized to evaluate the accuracy based on the VDA. The results were as follows: 1) the proposed algorithm could provide not only the percent accuracy of the evaluation but also the confidence interval of the BA; 2) the accuracy validation of the BA extracted by the normalized burn ratio (NBR) index and normalized difference vegetation index (NDVI) was verified by the VDA; 3) based on the field survey, the VDA confirmed that the NBR index had high accuracy, while the NDVI index had a large error, which is consistent with the results of using ground truth observations as a reference; and 4) the analysis based on the GF-6 image showed similar results. This study indicated that the VDA is effective and has the potential for widespread use in evaluating the accuracy of a BA.

This paper studies the reliable shortest path (RSP) problem in stochastic transportation networks. State-of-the-art RSP solutions usually target one specific RSP problem; moreover, the corresponding algorithm’s computational complexity scales at least linearly with the size of the underlying transportation network. While in this paper, we propose a graph embedding and deep distributional reinforcement learning (GE-DDRL) method, which serves as a universal and scale-free solution to the RSP problem. GE-DDRL uses deep distributional reinforcement learning (DDRL) to estimate the full travel-time distribution of a given routing policy, and improves the given routing policy with the generalized policy iteration (GPI) scheme. Further, in order to achieve the generalization ability to new destination nodes, we employ one of the canonical graph embedding techniques (Skip-Gram) to compress the nodes’ representation into$d$-dimensional real-valued vectors. With the properly compressed node features, GE-DDRL is able to generalize its estimation of the routing policy’s travel-time distribution to untrained destination nodes, and hence achieve the ‘all-to-all’ navigation functionality. To the best of our knowledge, GE-DDRL serves as the first RSP planner, which applies simultaneously to almost all RSP objectives and in the meanwhile, is scale free with the size of the transportation network in terms of the online decision-making time and memory complexity. Experimental results and comparisons with state of the arts show the efficacy and efficiency of GE-DDRL in a range of transportation networks.

Hemispherical photography is an important method of leaf area index (LAI) measurement, but the intermediate processes such as image segmentation and clumping index estimation will introduce errors. In this paper, an end-to-end model was proposed to directly estimate LAI from a hemispheric image based on GhostNet, which avoids errors introduced by the intermediate processing. Hemispherical images of crops and forest vegetation obtained from Shihezi, Xinjiang and Xiong'an, Hebei, and the measured values using LAI-2200 were utilized as the dataset. Compared with the LAI-2200 measurement results, the analysis results show that the correlation between them is extremely significant, with$\mathrm{R}^{2}=0.80,\ \text{RMSE}=0.65,\ \text{MAE}=0.46$. The experimental results show that the estimation model based on GhostNet can accurately estimate the LAI of the hemispheric image and is suitable for timely and accurate estimation of the LAI value of various types of vegetation through edge devices.

The fine classification of vegetation is a prerequisite for relevant remote sensing applications. Based on the phenological differences of different vegetation and the temporal changes of vegetation index, a crop fine classification algorithm is proposed, which combines spectral angle mapper algorithm and spectral information divergence algorithm. Taking Shihezi City, Xinjiang Province as the research area, 15 Sentinel-2 high-resolution images are selected as the data source, combined with the local phenological information, the NDVI time series curves of different crops are constructed, the refined classification of crops in this area is carried out, and compared with the traditional SVM algorithm and maximum likelihood method, The results show that the overall classification accuracy of this algorithm is 89.25% and 86.45% respectively, the kappa coefficient reaches 0.79 and 0.78. The experimental results fully show the advantages of the proposed algorithm compared with the traditional single temporal data source algorithm, and have a certain reference significance for the application of remote sensing images in ground feature classification.

Optimal exposure time is essential for accurate measurement of LAI (leaf area index) by DHP (digital hemispheric photography) method, The results showed that LAI was underestimated in different degrees under automatic exposure mode. To address this problem, in this paper we constructed a model of light intensity and optimal exposure time by studying the quantitative relationship between light intensity under the canopy and exposure time. The result shows that using the exposure based on this model rather than the automatic exposure, the comparison of LAI from the LAI-2200 and digital photographs is greatly improved, with R2 increasing from 0.398 to 0.845, and RMSE decreasing from 1.298 to 0.293. The method helps to solve the uncertainty of optimal exposure time in DHP method and improve the accuracy of DHP method in LAI measurement.

In this paper, an unsupervised classification method is proposed for polarimetric synthetic aperture radar (PolSAR) images. The proposed method combines the Freeman decomposition and Cloude decomposition to improve the initial classification take advantage of the scattering mechanism. Decomposition coefficients of three scattering mechanism components are extracted through Freeman decomposition first; then the scattering entropy and angle are calculated by Cloude decomposition. Employing the aforementioned five parameters, the initial classification plane is redefined and divided based on the dominant scattering mechanism. Finally, a clustering algorithm is adopted to the iteration of the initial classification, which optimizes the boundaries between different terrains. The experiment results demonstrate the superiority and effectiveness of the proposed method.

Operational aerosol optical depth (AOD) products are limited to coarse resolution (kilometers or hundreds of meters). In this study, the Sentinel-2 images were used to generate high-resolution (60 m) AODs over urban surfaces. Compared with traditional aerosol retrieval algorithm, the proposed algorithm has three major improvements including: 1) taking advantage of the aerosol-sensitive coastal band in aerosol retrieval; 2) no estimation of surface reflectance; and 3) not using of the shortwave infrared (SWIR) band. For validation, measurements from four Aerosol Robotic Network (AERONET) sites located in Beijing covering 2018 to 2021 were collected. The validation results show that the retrieved Sentinel-2 AODs highly correlate with AERONET measurements, with overall correlation coefficient for all four sites of 0.927, expected error (EE) of 68.75%, mean absolute error (MAE) of 0.082, and root-mean-square error (RMSE) of 0.108. The proposed algorithm can provide reliable AODs at 60 m resolution over urban surfaces.

Existing image matching is mostly limited to homologous images, but there are few researches on heterogeneous image matching such as UAV images and satellite remote sensing images. Aiming at the slow speed and low accuracy of traditional image matching methods when matching small-scale UAV images with satellite remote sensing images, a secondary matching method based on image edge features is proposed. The gradient information calculates the normalized cross-correlation formula to obtain the first matching result, and then obtains the optimal matching result through the shape context algorithm. Experimental results show that this method can well overcome the influence of factors such as illumination, low resolution, rotation, etc., and match UAV images and remote sensing images in a short time. The average precision and recall of the algorithm can reach 67.75% and 64.66%. Compared with the traditional SURF algorithm, it has faster matching speed and better matching accuracy.

The fuzzy c-means (FCM) algorithm and many of its variations have been widely adopted for image segmentation tasks. However, these methods are unable to present satisfactory segmentation results when dealing with synthetic aperture radar (SAR) images owing to the intrinsic speckle noise. In order to achieve the effective segmentation of SAR images, a robust FCM algorithm, namely NCBS_FCM, is proposed. The nonlocal spatial information is utilized to reduce the effect of speckle noise. Furthermore, NCBS_FCM takes advantage of the comentropy based on local gray histogram to acquire the adaptive weighting parameter for nonlocal spatial information term, which can achieve a better balance between speckle suppression and edge detail preservation. In addition, this paper incorporates the between-cluster scatter term into the objective function to adjust the distance between the cluster centers accordingly. Therefore, NCBS_FCM is more robust to various images and achieves satisfactory segmentation accuracy. Experiments on simulated and real SAR images show that NCBS_FCM outperforms other proposed variations of FCM by a significant margin.

This paper discusses the temperature factors leading to permafrost instability, and explains the causes of permafrost instability from the perspective of land surface temperature (LST). First, the land surface deformation of the typical permafrost area in some period is extracted by the using the dual-track differential interferometry (D-InSAR) technique with Sentinel-1A single look complex(SLC) data. Next, the annual average LST and LST range are calculated using the spatio-temporal interpolation algorithm for MYD11A2 products. Comparing the distribution of land surface deformation and LST, findings show that permafrost regions with lower average annual LST and larger annual temperature range are more unstable and more prone to surface deformation. This research can contribute to unveiling the process where LST affects the freezing and thawing of permafrost.

In this paper, Persistent scatterer interferometric synthetic aperture radar (PS-InSAR) is employed to extract the deformation of the transmission tower foundations. The extracted results are compared with the measured results to verify the correctness of the extracted results. 26 SAR images from January to December 2021 are selected for the deformation extraction of the high voltage transmission tower foundations near the landslide area in Guangyuan City, Sichuan, China. Time sequence analysis was performed using the StaMPS software, and GACOS products were used for atmospheric correction. The final deformation results are compared with the monitoring results of the sensors installed on the foundations of the transmission tower. The comparison shows the effectiveness of the method. The conclusions obtained from the experiment provide scientific guidance and suggestions for the power maintenance personnel.

This paper proposes a landslide detection method combining gray level co-occurrence matrix (GLCM) and polarization decomposition technique. The proposed method makes full use of the characteristics of the SAR image and combines the polarization information with the texture feature. Landslide candidate areas were identified using GLCM and final landslide areas were determined using polarization decomposition techniques. The data used for evaluation is taken from Guangyuan City, Sichuan, China. The proposed method successfully detects the changes of two SLC sentinel-1A C-band VV and VH dual polarization SAR images before and after the landslide, which proves the effectiveness of this method.

High-resolution aerosol retrieval is of great significance for understanding the impact of aerosols on air pollution and climate change. In this study, an algorithm for aerosol retrieval at a spatial resolution of 60 m over complex urban areas is developed using Sentinel-2 images. The proposed algorithm has two assumptions: 1) the blue–red surface reflectance ratio does not change temporally in a single season and 2) surface reflectance over bright areas is invariant over three months. Then, the aerosol optical depth (AOD) is retrieved from the surface reflectance correlations with a combination of temporal signatures over the vegetated areas and bright areas. The aerosol robotic network (AERONET) measurements in Beijing and its surrounding from 2017 to 2019 are collected and used to validate the retrieved 60-m Sentinel-2 AODs; 77% of the retrieved Sentinel-2 AODs fall within the expected error (EE), and the correlation coefficient is of 0.927. MODerate resolution Imaging Spectroradiometer (MODIS) AOD products at 1- and 10-km resolutions (MCD19A2 and MOD04_L2, respectively) are acquired to compare with the retrieved Sentinel-2 AODs. Comparison results show that the retrieved Sentinel-2 AODs are superior to the MOD04_L2 dark target (DT) and MCD19A2 AODs, and slightly better than the MOD04_L2 deep blue (DB), and DT and DB combined (DTBC) AODs. The validation and comparison results indicate that the proposed algorithm is able to describe aerosol distributions at high resolution continuously. However, further work is needed to apply the proposed algorithm on a global scale.

The indirect estimation of leaf area index (LAI) is important to agricultural science and ecological science. In this paper, we reconstructed the three-dimensional (3D) structure of the plant based on 3D reconstruct technology, such as the Structure from Motion (SFM) and the multiview stereo (MVS). A LAI measurement method was also proposed based on the ratio of the leaf point cloud area to the ground area. The experiment results showed that the R2between the true value of the leaf area (LI-3000) and the measured value is 0.9657; compared with the true LAI value calculated based on the leaf area, the relative error of the LAI value extracted from the 3D model is about 44.4%. After removing the stems and smoothing the leaves, the R2between the true value of the leaf area and the measured value reached 0.9741; the relative error of the measured value of LAI is 9.13%. The experimental results also implied that the deviation of the LAI measurement value mainly comes from the incomplete reconstruction of the covered part of the plant, the influence of the stem, and the smoothness of the leaf surfaces after reconstruction.

To solve the problem of the unstable feature point extraction and the difficult feature matching in SAR image processing, we proposed a detection algorithm based on hybrid feature extraction. First, the repeated gray information is removed to reduce the complexity and import the real-time performance of the algorithm. Then MSER is utilized to extract regional extremes as candidates. And Harris-Laplace algorithm, which ensures uniform and stable feature-point-distribution, is employed to detect the feature points. Next, an improved SAR-SIFT algorithm is used for the feature matching. Finally, FSC and VFC are devised to refine the feature matching results. The dataset we used is the GaoFen-3 series image as it retains fine edge and structural texture information. The experiments have demonstrated the effectiveness of the proposed algorithm with insert performance metric here.

This paper proposed a novel river detection algorithm for SAR images with an improved version of OTSU algorithm. The width of the river can also be estimated using the detected results. The gray threshold, which is critical in the detection of the river, can be obtained adaptively by the proposed improved OTSU algorithm. Thus, it is more accurate than the conventional OTSU algorithm. In addition, DEM is employed in the river region extraction process to mitigate the effect of mountain shadows. The target refinement, the skeleton extraction and the endpoint removal algorithms are all devised to acquire the river centerline image. And the distance transformation algorithm is applied to generate the river distance map. Henceforth, one can estimate the river width by combining the previously acquired images. In order to create a more constant river centerline image, we employed a bridge removal procedure before the river skeletonization. The data used for the evaluation is from Sentinel 1A single-polarization data of the Yangtze River near Nanjing City, Jiangsu Province. The experiment results have shown that the proposed algorithm is more effective in the river extraction.

Road extraction from remote sensing images based on deep learning has always been a hot research topic. However, due to the complexity of road conditions, many deep-learning models cannot obtain satisfactory results of road extraction. To improve the accuracy of road extraction, this paper proposes an improved deep-learning model based on the structure of Deeplabv3+. The proposed model uses four blocks with ResNeSt and ASPP to extract the feature maps, which can improve the integrity of the extracted roads. And each of the extracted low-level features is transmitted to the decoder. The decoder focuses on the effective fusion of low-level and high-level feature maps and gradually restores this information layer by layer, which can produce a more accurate result of segmentation. And because the road results extracted by neural networks often have broken road lines, this paper also proposes a data post-processing method to effectively solve this problem. The final experimental results show that the proposed model has further improvements in accuracy, mean Intersection over Union, and Fl-score compared to some other state-of-the-art(SOTA) models.

Monitoring aerosol distributions at the city level is of great importance. However, because the differences in spectral response functions and spectral channels among different sensors, the widely used second-generation operational aerosol retrieval (V5.2) algorithm cannot be directly applied to high-resolution Landsat 8 OLI images. In this study, we proposed a new high-resolution aerosol retrieval algorithm for Landsat 8 OLI images over urban areas. This new algorithm retrieved aerosol optical depth (AOD) directly from the red-blue surface reflectance relationship, without estimation surface reflectance. Therefore, satellite measurements even without V5.2-required$1.24\mu \mathrm{m}$or$2.1\mu \mathrm{m}$channel are able to retrieve AODs using this algorithm. For validation, the AERONET measurements located in Beijing were collected. The validation results show that our 30m retrievals agree well with the AERONET measurements, with correlation coefficient of 0.944, EE of 85.19%, MAE of 0.051, and RMSE of 0.065, and are able to provide detailed aerosol distributions over urban areas.

With the further improvement of industrialization, the problem of atmospheric pollution has become a major issue closely related to our human existence that is of common concern to scientists worldwide. The rapid development of remote sensing technology makes it possible to observe the atmosphere over a long period of time and on a large scale. In this study, we compared the Moderate-resolution Imaging Spectroradiometer (MODIS) aerosol product MCD19A2 with the NO2 retrieved by Sentinel-5 Precursor (Sentinel-5P). It is found that the data based on Sentinel-5P can clearly detect the local steel smelting enterprises while MODIS cannot, and the column density distribution trend has a better correlation with the steel output of the current month. If the spatial resolution of related products is improved in the future, this method of atmospheric observation can be better used for pollution monitoring.

SAR image enhancement plays an important role in the process of SAR image processing and information interpretation. Super-resolution reconstruction is a widely adopted enhancement method. However, it is difficult to achieve a decent tradeoff between reconstruction effectiveness and the convergence speed for existing methods. To combat such problem, this paper proposed a novel solution to it, we started from the modeling of SAR image degradation principle, applying the adaptive line search strategy to the SAR image super-resolution reconstruction process, and redefined the step size selection in the reconstruction process, made it possible to achieve both reconstruction effectiveness and convergence speed. Compared with the existing empirical setting or iterative selection, the proposed method can reduce the number of iterations while guarantee the reconstruction results.

The polarimetric synthetic aperture radar (PolSAR) image filtering is an essential step before the quantitative inversion. However, the existing filtering algorithms often change the scattering features of the original PolSAR images. This leads to a performance decrease for the subsequent quantitative inversion. In order to combat such drawback, we adopted a novel filtering algorithm preceding the quantitative inversion. This algorithm adopted the idea of “classify first, filter later”. By utilizing the hybrid four-component polarization decomposition (HPD) method to pre-classify the pixels, and implementing non-local means lee (NLM-Lee) filtering only between the same ground object points, the scattering features are well-protected. The experiment results show that the proposed algorithm can not only effectively reduce the speckle and preserve the structural features of the image, but is far superior in preserving the scattering features of the images. Hence, when the PolSAR image filtering is carried out by the proposed algorithm, the performance of the quantitative inversion will be improved.

In this work, an effective 3-D reconstruction method for the transmission tower in mountain area, which combines the Persistent Scatterer Interferometry (PSI) and SAR Tomography (TomoSAR) based on generalized Orthogonal Matching Pursuit (gOMP) algorithm, is proposed. This paper is focused on reconstruction performance of gOMP algorithm on SAR signal and the 3-D reconstruction of transmission tower, using the results obtained in mountain area on simulated and real TerraSAR-X data. The vertical signal of slant distance from SAR images can be regarded as sparse signal, so the gOMP algorithm can be used to obtain extraction of elevation and deformation rate of transmission towers in mountain area. When only primary and secondary scattering points is present in each range-azimuth resolution cell, this method can be exploited to estimate the elevation and the deformation rate of the transmission tower. The results have demonstrated the advantage of this method.

SAR image enhancement is of great significance to SAR image processing. Aiming at the defects of traditional partial differential algorithms sensitive to noise, this paper proposes a SAR image enhancement method using both P-M nonlinear equations and coherent enhanced partial differential equations. Combining the physical model of the SAR image, the generalized diffusion process synthesizes the properties of both forward and backward diffusion, and has a good smooth enhancement effect at edges as well as smooth areas of the image. This hybrid model not only avoids noise enhancement, but also enhances image edges. Experiment results show that the proposed hybrid model has significantly improved the sharpness of enhanced SAR images with minimal distortion in image details, which is proven to be superior than applying only single one of the models.

Aerosol distributions may change at fine spatial scales in urban areas due to building and transport infrastructure and human population density variations. The recent availability of Sentinel-2 satellite data provides the opportunity for aerosol optical depth (AOD) estimation at higher spatial resolution than provided by other satellites. In this study, a novel high-spatial-resolution AOD retrieval algorithm for bright unban surfaces was developed based on the Sentinel-2 images. The surface reflectance for AOD retrieval was estimated from the image that has the minimal aerosol contamination in a temporal window. Validation of the Sentinel-2 AOD retrievals was conducted against four Aerosol Robotic Network (AERONET) sites located in Beijing. The results show that the Sentinel-2 AOD retrievals are highly consistent with the AERONET AOD measurements ( R=0.9424), with 85.56% of them falling within the Expected Error (EE). The mean absolute error (MAE) and the root mean square error (RMSE) are 0.0688 and 0.0882, respectively. These results suggest that our high-resolution AOD retrieval algorithm is robust and useful to retrieve high-resolution AOD over bright urban surfaces based on Sentinel-2 images.

In this paper, an improved change detection method (ICDM) is proposed to obtain the soil surface moisture (SSM) with a resolution of 20 meters, which greatly improves the resolution of SMAP soil moisture products. First, NDWI was calculated using a time series data of Sentinel-2 and then converted to vegetation water content (mveg). With the help of the water cloud model, the influence of vegetation on the backscattering coefficient is eliminated. Then, in order to make the change detection method (CDM) more reasonable, a method for removing influence of roughness change on backscattering coefficient is proposed. It uses the difference between co-polarization and cross-polarization ( σ0vv- σ0vh) for roughness normalization. SSM was obtained by combining the soil relative moisture retrieved from the time series data of Sentinel-1 with the maximum and minimum water content of the study area within one year obtained from the SMAP data. The results showed that the MRE of SSM extracted by ICDM was 8.34%, which was less than 17.49% of CDM.

In this paper, a landslide detection method using Small Baseline Subset Differential Interferometry (SBAS-DInSAR) combined with gray level co-occurrence matrix (GLCM) that combines the surface deformation rate with the texture feature information of the ground object to make full use of the ground object information. Compared with the traditional technique using only texture features, taking Qinglong County, a mountainous region in Guizhou Province in the southwestern region as the research area, can improve the accuracy of landslide detection. The final experimental results demonstrate the effectiveness of the proposed method, which can be used for landslide warning, the general survey of landslide bodies, and provide auxiliary information support for the detailed investigation.

Detecting the change of texture information is very important for landslide detection. In polarimetric SAR imagery, the landslides not only have typical polarimetric features but also have rich texture features. In this study, the texture feature extraction technique based on gray scale co-occurrence matrix was applied to process two SLC Sentinel-1A C-band SAR images and two COSMO-SkyMed X-band SAR images acquired before and after landslip for the landslide detection leading to the Jomda landslide event in 2018, Tibet of China and the Qinglong landslide in 2018, Guizhou of China. Several important characteristic parameters of the gray scale co-occurrence matrix were studied, and the optimal sliding window size and the optimal step size for landslide detection were obtained through rigorous analysis and comparison. The experimental results show that the analysis of the complex texture information can effectively detect the landslide in the mountain area with high accuracy.

In this paper, to address the problem in the registration of synthetic aperture radar (SAR), the speeded up robust features (SURF) approach based on multi-level FAST and improved random sampling consistency is proposed. This approach is used to address the high mismatch rate in SAR image registration, and it is based on the characteristics of SAR images. We construct a new corner detection by combining with the hierarchical theory of the system, which is referred to as multi-level FAST (MFAST). The contribution of MFAST algorithm mainly lies in the efficiency of SURF algorithm, which determines the main direction and feature descriptor. And improving the random sampling consistency (RANSAC) to remove mismatched feature points and ameliorate the registration effectiveness. The proposed algorithm is suitable for multi-sensor images with large gray differences and significant edge features. The experimental results show that the registration efficiency of the proposed algorithm is more than three times of the SURF algorithm, and the registration accuracy also is better than traditional SURF algorithm, which can reach 0.39 pixels.

For the measurement of LAI (Leaf Area Index) by DHP (Digital Hemispherical Photography) method, imprecise segmentation is the key error source. In this paper, to our knowledge, a deep learning algorithm is used for the first time to segment upward hemispherical image of vegetation. Pix2pix, a general mapping learning model, was improved in our study to make it more suitable for processing segmentation problem. Thousands of images collected in the field were labeled to train the model, and the conventional methods based on pattern recognition, such as the Otsu and HSV, were compared. The result shows that the improved pix2pix algorithm significantly improved the accuracy of the segmentation, which reached to 0.9834. Furthermore, this model has a good performance in processing pictures of complex environments, and the segmentation of edge details has also been optimized. Those results show that the method has great potential to improve the LAI measurement accuracy.

The popular enhanced deep blue (DB) algorithm, though performs well in aerosol retrieval over the entire land surfaces, limited by the low temporal resolution and insufficient satellite-derived products of most high resolution satellite sensors, is hardly to be applied to urban areas using high spatial resolution satellite imageries. In this paper, we developed a simplified deep blue algorithm to retrieve 30m spatial resolution aerosol optical thickness (AOT) over urban surfaces using Landsat 8 OLI measurements. With a few atmospheric correction surface reflectance-apparent reflectance pairs, robust relationships between visible (0.65μm and 0.48μm) and 2.2μm reflectance can be constructed by this algorithm. Difficulties described above are overcame, and high consistency with ground-based AERONET measurements are achieved over Beijing, with correlation coefficient (R2) ~0.951, root mean square error (RMSE) ~0.005, mean absolute error (MAE) ~0.05 and 82.61% retrievals fall within the expected error (EE) envelop. This study demonstrates that the simplified deep blue algorithm has advantages in retrieval high spatial resolution AOTs over urban areas.

Differential SAR tomography (D-TomoSAR) extends the SAR tomography to the time dimension for four-dimension imaging. Hence, it is a powerful tool for retrieving the four-dimensional structure of buildings in urban area. D-TomoSAR can also be applied to electric power transmission tower evaluation. However, among the previous D-TomoSAR estimators, some reduces the azimuth-range resolution and has large sidelobe while the others cost more computation to retrieve. To overcome aforementioned drawbacks, this paper proposed an improved Stagewise Weak Orthogonal Matching Pursuit (SWOMP) based on twice iterations and scatter number estimators to calculate the height and deformation velocity of the electric power transmission tower. Through simulations and real data (TerraSAR-X) experiments, a comparison between the SWOMP, adaptive beamforming (Capon) and compressive sensing (CS) was carried out. The results indicate that the proposed method obtains less sidelobe than SWOMP and Capon. In addition, this method obtains similar result to CS in less computation time. the results have demonstrated the advantages of this method.

An improved deramping method which is an essential part of differential tomography SAR (D-TomoSAR), is proposed in this paper. Based on the time series of SAR data, first the coherence coefficient method is employed for image registration. Next, the image center is chosen as the reference point. The reference point's exact coordinates can be calculated by Gradient-descent algorithm to solve the Range-Doppler (R-D) equations, and it converges when the difference between the calculated elevation and the one from DEM database meets the threshold requirements, The experiment results have demonstrated that the elevation estimation of the target tower using deramping method proposed in this paper is closer to the real height of the tower by Yunnan electric power bureau compared to least squares method. Furthermore, the accurate deformation rate of the transmission tower can be extracted by this method.

This paper proposed a deep learning road extraction model based on similarity mapping relationship for extracting roads from high resolution remote sensing images. The model adopted a full convolutional network structure, and used the similarity between each pixel in the input image and the typical label sample to perform mapping relationship search, thereby, the classification of each pixel in the image to be classified is marked to the classification category of the typical sample. The model breaks through the traditional deep learning mindset and directly stores knowledge in the network, instead of just learning a set of feature extraction and integrated network parameters. Under the training data of 10,000 small samples, the corresponding image results are obtained after 100,000 iterations. The experimental results show that the accuracy of the method is 0.87, which proves that this method is more accurate and efficient than traditional algorithms.

Conventional measurement methods with low spatial resolution, differential interferometric synthetic aperture radar (D-InSAR) is susceptible to signal decorrelation and atmospheric delay, so InSAR time series analysis is widely used for detection and monitoring of slow surface deformation. Persistent scatterer interferometric synthetic aperture radar (PS-InSAR) is based on a large number of SAR images, but small baseline subset interferometric synthetic aperture radar (SBAS-InSAR) only needs a small number of images and performs better than PS-InSAR for obtaining nonlinear deformation information. In this paper, 21 COSMO-SkyMed radar images acquired between 2016 and 2017 are processed for land subsidence monitoring with the Small Baseline Subset (SBAS) approach in Yongshan County transmission corridors in China.

In order to accurately extract landslide disaster information from SAR images, we propose a clustering algorithm for landslide detection using multi-temporal fully polarimetric SAR image. By combining Freeman-Durden decomposition with H /α / A decomposition method, this algorithm makes full use of scattering power and scattering entropy information. Furthermore, a hierarchical method is proposed to overcome the drawback that fuzzy membership function family of multi-region classification cannot always satisfy the constraint conditions to guarantee functional convexity. Such method ensures that the constraint conditions can always be satisfied and improves the accuracy of classification. Finally, the landslide disaster area is accurately extracted from the two classification maps by change detection. The results of the C-band RadarSat-2 data are provided to demonstrate its competitive general performance of image classification.

Resolution is one of the important indices to measure the quality of SAR images. Super-resolution reconstruction is a widely adopted resolution enhancement method. Many algorithms have been developed for the super-resolution reconstruction. Among these algorithms, this paper applies projections onto convex sets algorithm to SAR image reconstruction processing. The POCS can efficiently obtain high-resolution SAR images with enhanced details. However, the POCS requires many low-resolution SAR images of the same area to gain a better result, usually 10 to 20 images. Such requirement is very difficult to achieve when only single-polarization mode is included. In this paper, we propose a novel method that utilizes all the polarimetric images of the same original SAR data for the algorithm. Thus, the number of the available images is increased exponentially. The experiment results have demonstrated the effectiveness of our proposed method: The reconstructed high-resolution SAR image based on multi-polarimetric information is more detailed and clearer than that based on single-polarization information.

In order to accelerate the calculation of the similarity metric between two patches in non-local means filter of polarimetric synthetic aperture radar (PolSAR) images, this paper proposes a weighted acceleration algorithm to eliminate redundant operations. It can greatly reduce the computational complexity by using the accumulated covariance data to avoid repeated computations. Furthermore, the algorithm introduces the polarization similarity parameter to preserve the details of the image. It also processes the covariance matrix of each pixel. Thus, the polarimetric properties are maintained. The results of the C-band RadarSat-2 data are provided to demonstrate its rapidity and competitive general performance of speckle reduction and edge preservation.

Synthetic aperture radar (SAR) interferometry is a technique that provides high-resolution measurements of the ground displacement associated with many geophysical processes. Currently there are two broad categories of algorithms for processing multiple acquisitions, persistent scatterers interferometry (PSI) and small baseline methods(SB), which are optimized for different models of scattering. However, the scattering characteristics of transmission corridors in mountain areas usually lay between these two end-member models. In the experiment, we selected a set of SAR images of TerraSAR-X satellite, covered Mao County area, to detect the subsidence during the fifteen months, using stanford method for persistent scatterers-multi-temporal InSAR technology (StaMPS-MTI).

The complex terrain, frequent movement of the earth's crust, dense vegetation exists in the western sichuan plateau where has frequent disasters. PSInSAR can extract the target with stable, strong scattering characteristics to monitor deformation effectively. The experimental area is in maoxian, erlang- mountain, chengdu. Paper uses, Pseudo-3D Phase Unwrapping and LAMBDA Method respectively related to GPS integer ambiguity resolution to unwrap phase and extract the surface deformation of the research area, and comprehensively analyze the effects of different surface and different solutions to the results. Experimental results show that the Network Adjustment Method in elevation correction and linear deformation rate has a comparative advantage, can be more applicable to maoxian, mountain complex mountainous area surrounding the transmission channel of deformation monitoring.

The inclination direction monitor of the power transmission tower can issue an early-warning for the risk of the tower collapse, as well as the collapse direction. In this paper, we proposed a new method to retrieve the inclination direction of the power transmission tower using geocoding. First, the backscattering coefficient was utilized to separate the power transmission tower from the background, and figure up those coefficients along the tower direction to extract the main body of the tower. Second, the whole image was traversed by a$3^{\ast}3$template window to refine and mark the most possible learning tower points. Finally, a distance-Doppler model was built to retrieve geographical coordinates of those marked points. According to the geographical coordinate of the tower point at different time, we got tower inclination direction. Compared with the actual measurement results of the study area, the retrieval result has a 0.0025° error, and the direction is north-east, consistent with the actual direction.

High temporal and spactial resolution aerosol retrieval is a difficult task because of the absence of corresponding satellite data. Due to the lack of a shortwave infrared band near 2.1 um aboard on Gaofen-4 instrument, which is critical for determining surface reflectance. In this paper, a new algorithm resolving the problem based on Gaofen-4 that was placed in Geosynchronous (GEO) orbit and Landsat-8 data was proposed. Also, Gaofen-4 sensor band mean solar irradiance (BMSI) was calculated which was not open to public until now. In the algorithm, normalized difference vegetation index (NDVI) was used to identify dark target pixels that have certain linear relationship over blue and red bands' surface reflectance. In order to remove Gaofen-4's geometric deformation, the data were processed with RPC Landsat-8 panchromatic data. The algorithm was applied to two cities, Beijing and Chengdu. The result, aerosol optical thickness (AOT) with a 50m × 50m resolution, indicated the algorithm can be effective for vegetation area or low surface reflectance area. The algorithm is very useful and significant for environmental protection, air quality monitoring and atmospheric pollutants sources tracing.

With the development of technology of digital camera, the performance of image sensor is continuously improving, which makes it possible to measure leaf area index by digital photos in a simpler and cheaper way compared with the other ways. This article proposed a new way to measure LAI by using digital images, and the formulation of measuring LAI is deduced from both the principle of LAI2000 and theory of digital camera. Considering the system error, the modified formula is also deduced. In this article, there are about 35 measurement locations which include kinds of different density of canopy, where the LAI were measured by both proposed method and professional instrument LAI2000 to validate the accuracy. The experiment results show that the correlation coefficient was 0.9794 between the proposed method and LAI2000, which proved the accuracy and feasibility of measurement of LAI based on digital photos. The new method has a bright application prospect.

In this paper, a novel method was developed to orientate and quantify the air pollution sources based on remotely sensed aerosol data and Glowworm Swarm Optimization (GSO). In practice, based on source apportionment technique, the air pollution sources could just be identified to certain industries, such as transportation, power plants, biomass burning, and et.al. To our knowledge, the problem of orientating and quantifying the pollution to the individual factories is faced for the first time. In this study, the aerosol retrieved from remotely sensed image (MODIS) and GIS were used to locate and quantify the pollution to each enterprise in the study area based on an improved Glowworm Swarm Optimization and meteorological condition. As a result, the polluting contribution of each factory were be listed, and the most polluting factories were be found. Some experiments were carried out to validate the method, and the Key monitoring factories by authority was ferreted out accurately.

Aerosol retrieval over urban areas is a difficult task because of the high reflectance of the underlying surface. In this paper, a new aerosol retrieval algorithm based on the spectral analysis of soil and vegetation from spectral library was proposed, the simulated correlation between the normalized difference vegetation index (NDVI) and the surface reflectance of red, blue bands was established. And also to solve scale problem, a conversion method based on maximizing mutual information (MI) was used. The algorithm was applied to Beijing city using the China HJ-1A/1B of the Environment and Disaster Monitoring Microsatellite Constellation Charge-Coupled Device (CCD) and MODIS NDVI data. The result, aerosol optical thickness (AOT) with a 100m×100m resolution, was compared to the ground measurement data from Aerosol Robotic Network (AERONET), which shows a high consistency with observation data, and the overall correlation coefficient of approximately 0.935 and a root-mean-square error (RMSE) of about 0.34. The algorithm is very useful and significant for environmental protection and air quality monitoring over urban areas.

In this paper, a new aerosol retrieval algorithm based on a method, named statistical segmentation, was proposed. Firstly, the image of Landsat 8 OLI was divided into many segments by the statistical segmentation method based on band 6 and band 7. Then, according to the characteristics of the segmentation, two ways based on the segmented results were used to get the surface reflectance. And then combined with the apparent reflectance equation and a lookup table built by 6S model, aerosol retrieval could be performed. In principle, this algorithm is based on clean pixels (almost no aerosol) at band 1 to retrieve contaminated pixels in the same segment. The retrieved results show that, compared with DDV (Dense Dark Vegetation) algorithm, this algorithm is more suitable for bright surfaces, such as urban areas.

This paper developed a method of validation test research of domestic satellites LST products. The experimental data include high resolution landsat8 image, low resolution FY-3B and MOD11A1 LST product. The main steps of the method are as follows: (1) Three different algorithms for retrieving land surface temperature on landsat8 data. Through the comparative analysis of the results, we select an algorithm for subsequent verification. (2) Use the inversion result and MOD11A1 product to compare with FY-3B product respectively.(3) Use matlab to compute some parameters and curve data fit. The study results show that the FY-3B temperature product is similar with the others. It is available in a certain extent.

Fine registration is a fundamental step for further application of remote sensing images. Focused on deficiencies in traditional manual registration, this paper presents a new method for automatic fine registration of multi-spectral images. To make the most of image information, the algorithm detects and matches feature points in the selected bands. Then pick up the common control points which contain more reliability relative to others after eliminating wrong matching points. The last registration model can be built based on common control points and the points selected by common ones. Experimental results with Landsat TM5 images demonstrate that the method is more accurate and suitable for automatic batch processing.

This paper proposes a new method of spatial distribution of PM2.5concentration, which is based on Aerosol Optical Thickness (AOT) inverted by Landsat7 ETM+ Enhanced Thematic Mapper Plus (ETM+) data and PM2.5ground-based instruments. The main steps of the method are as follows:(1) Determination of dark pixels, search for the dark surface targets with the 2.2- μm channel; (2) Determination of the surface reflectance in the blue and red channels;(3) AOT retrieval, inverted from Look-up Tables (LUT) established from a set of atmospheric geometrical conditions;(4)Establish the model of PM2.5predicting with predictors and estimate the PM2.5value. The study results show that the method is an effective means for predicting the PM2.5concentration, and a beneficial supplement to the conventional ground-based measurement.

Radar image has now becoming more and more widely used in the ecological environment monitoring. The soil moisture inversion for SAR(Synthetic Aperture Radar) image has important significance in the field of agriculture and environment. However, as the SAR image data is large scale and the algorithm of soil moisture inversion is complicated, it's a time-consuming work for SAR image processing, so how to get fast processing for SAR images has now become an important problem. Therefore, the this paper use MPI(Message Passing Interface) cluster system accelerate the processing speed of algorithm, thus speeding up the SAR image processing and improving the efficiency of the use of computers. This paper designed and implemented two cluster computing modes that are master-slave mode and peer-to-peer mode. At last, we takes an experimental testing, results show that relative to master-slave mode, peer-to-peer mode can get a better acceleration effect, greatly improves the processing speed.

In today's society, as town roads and construction infrastructure has been improved gradually, the noise pollution has effected the environment more and more seriously. The noising map which has unique way to show the distribution of noise in the real-time and can monitor the situation of the noise effectively, it can let user master the situation of noise pollution clearly, and the application of noising map has become a main research direction of denoising work in recent years. This paper mainly analyzed the domestic and foreign research of the noising map and the main noise prediction method, selected the RLS90 model to develop the software. In the development, we used the ArcGIS Engine setups and Visual Studio programming environment, using the object-oriented development method and the GIS component second development function to develop a relatively perfect function noise prediction and noising map generation system based on the geographic information system.

This paper proposes a new method of spatial scale conversion which is combining the MODIS product geometry information with the decomposition method basing on NDVI pixel. The method makes use of the LAI ground data, high-resolution images (TM) data and MODIS pixel geometry information, which make the 30m resolution image of TM-NDVI weight up to 1km by space response function, and then we can get the value of LAI inversion after pixel decomposition. The result of the scale conversion is compared with the MODIS LAI product to achieve the validation of the low resolution (MODIS) LAI product data. The experiment results show that the new method can be fully consider of the heterogeneity of the surface and pixel correspondence to the effect of results validation, which can effectively improve the accuracy of product certification, and is much better than existing space scale method.