on September 7, 2021, GaoFen5-02 (GF5-02) Satellite, of the new generation of Chinese hyperspectral remote sensing satellite was successfully launched. GF5-02, the successor to the GF5 satellite, was equipped with six advanced hyperspectral payloads. One of the most important payloads onboard the GF5-02 satellite, the Advanced Hyperspectral Imager (AHSI) has a spatial resolution of 30 m, 330 bands in a spectral range of 380-2500 nm. The spectral resolution of the visible-near infrared (VNIR) and shortwave infrared (SWIR) bands are better than 5 nm and 10 nm, respectively. In order to analyze the spectral performance of the GF5-02 AHSI, an on-orbit spectral calibration method that utilizes atmospheric limb observations with on-board calibration system was proposed in this paper. The on-orbit spectral calibration results were validated by atmospheric absorption features with synchronous measurements of surface reflectance and atmospheric parameters. For the GF5-02 AHSI, the shifts in the central wavelength of the Visible and Near-Infrared (VNIR) band is 0.117 nm, while the shifts in the Full Width at Half Maximum (FWHM) is 0.02 nm. In the Short-Wave Infrared (SWIR) band, these values are 0.25 nm for the central wavelength and 0.04 nm for the FWHM. The results demonstrate that the applied method is effective for on-orbit spectral calibration for GF5-02 AHSI.

Semantic segmentation of remote sensing images is crucial for resource exploration, precision agriculture, and environmental monitoring. However, conducting semantic segmentation on single-modality data for remote sensing images that contain various scenes, especially unique scenes, is highly challenging. To address this challenge, we propose SiMultiF, a Siamese architecture-based multimodal feature adaptive fusion semantic segmentation network. SiMultiF employs a dual-branch Siamese structure feature extractor. The adaptive feature weight adjustment module (AFWAM) and the multimodal fusion module (MFM) facilitate in-depth understanding and extraction of multimodal data. Specifically, the Siamese structure can extract features from multimodal data concurrently without adding to the number of parameters. The AFWAM module can adaptively identify the importance of different modal data and dynamically adjust the modal weight to enhance the network’s comprehension of complex scene data. Additionally, the cross-attention (CA)-based MFM module bridges modality gaps and achieves comprehensive multimodal feature fusion. Numerous experiments have demonstrated that the proposed SiMultiF outperforms other state-of-the-art semantic segmentation models (both multimodal and single modal) on the high-resolution ISPRS Potsdam dataset, ISPRS Vaihingen dataset, and special scene dataset (vegetation polarization dataset with extreme natural lighting contrast). Moreover, the robustness and generalizability of the network in multiscene and multimodal datasets are verified.

Interband radiometric calibration from the mid-infrared to visible bands in the ocean specular region is an effective way to calibrate on-orbit remote sensing sensors. It assumes that the referenced band has highly accurate radiance and that the interband radiometric relationship can be obtained in the ocean specular region. Most current research employs only the radiative transfer (RT) equation to derive interband radiometric relationships. However, two variables—water-leaving radiance and whitecaps—are challenging to obtain yet crucial for radiative transfer calculations. Typically, water-leaving radiance is assigned a fixed value since empirical data, whereas whitecaps are estimated via the wind speed alone. These assumptions make the uncertainties of the calibrated bands large and different from those of real satellite-measured data, reducing the reliability of the interband relationship between the reference and calibrated bands and limiting the application of the interband radiometric calibration method. To address this issue, this study proposed a novel interband radiometric calibration method called coupled deep neural networks and radiative transfer (CDR), which integrates radiative transfer and a deep neural network (DNN) to provide a reliable relationship between referenced and to be calibrated bands without accurate water-leaving radiance and whitecaps. For the four visible bands of FY-3D/MERSI-II, the relative errors were found to be 2.12%, 4.62%, 1.89%, and 4.02%, respectively. Uncertainty analysis identified the referenced band as the largest uncertainty source, followed by chlorophyll concentration, polarization effects, and aerosol loading. The CDR algorithm can be used to calibrate historical long-term satellite data without additional measurements.

Semantic segmentation is an essential technique in remote sensing. Until recently, most related research has focused primarily on advancing semantic segmentation models based on monomodal imagery, and less attention has been given to models that utilize multimodal remote sensing data. Moreover, most current multimodal approaches consider only limited bimodal situations and cannot simultaneously utilize three or more modalities. The increase in expensive computational costs associated with previous feature fusion paradigms hinders their application in broader cases. How to design a unified method to cover a wide variety of quantity-agnostic modalities for multimodal semantic segmentation remains unsolved issues. To address the aforementioned challenges, this study explores a feasible way and proposes a cost-effective multimodal sensing semantic segmentation model (MultiSenseSeg). MultiSenseSeg employs multiple lightweight modality-specific experts (MSEs), an adaptive multimodal matching (AMM) module, and a single feature extraction pipeline to efficiently model intra- and inter-modal relationships. Benefiting from these designs, the proposed MultiSenseSeg can serve as a unified multimodal model capable of addressing both monomodal and bimodal cases and readily extrapolating to scenarios with more modalities, thereby achieving semantic segmentation of arbitrary quantities of multimodal data. To evaluate the performance of our method, we select several state-of-the-art (SOTA) semantic segmentation models from the past three years and conduct extensive experiments on two public multimodal datasets. The results show that MultiSenseSeg can not only achieve higher accuracy but also exhibits user-friendly modality extrapolation, allowing end-to-end training for consumer-grade users based on limited hardware resources. The model’s code will be available at https://github.com/W-qp/MultiSenseSeg.

Identifying land cover changes based on hyperspectral images (HSIs) has been a research hotspot in the field of remote sensing. In recent years, deep learning-based change detection (CD) methods have advanced the development of this subject due to their powerful feature representation capabilities. However, it is difficult for these methods to mine changed information between bi-temporal HSIs with limited labels. To overcome this limitation, we propose a graph-based hyperspectral CD framework using difference augmentation and progressive reconstruction (ARCD), which enhance the recognition ability of changes of HSIs with limited labels. This framework consists of three components: 1) a dual-brach multi-scale dynamic GCN (DMGCN) sub-network, which is developed to emphasize the changed information and learn global features of HSIs at various scales; 2) a difference augmentation feature fusion (DAFF) module, which is designed to fuse spectral-spatial augmentation information and the difference information to accurately capture discriminative features for the changes between bi-temporal HSIs; 3) a progressive contextual information attention reconstruction (PCAR) module, which is proposed to focus on key information in the context, and progressively reconstruct multi-level features to reduce semantic gaps between different scale features. ARCD not only enhances the representation ability of changed features, but also alleviates the demand for HSI labels. We test the performance of ARCD on four hyperspectral datasets. Quantitative and qualitative results reveal that it outperforms some state-of-the-art methods with limited labels.

This study proposes a multiband synthetic method to calculate the BC DRF. It is compared with the empirical formula method and broadband method. The multiband synthetic method inputs the 6 narrowband black sky albedo of MODIS and other data into the 6S model to calculate the sky scattered light ratio of each narrowband and obtains the narrowband blue sky albedo. And input the BC AOD and the synthesized broadband shortwave blue sky albedo into the 6S model to obtain the shortwave BC DRF. This study found that the BC DRF calculated by the three methods at the TOA is significantly different, while the difference between the three methods at the SFC is small. The values of BC DRF calculated by these three methods are multiband synthetic method, broadband method, and empirical formula method from high to low. Therefore, when calculating shortwave BC DRF, the difference in atmospheric scattering effects on narrow bands should be considered.

Since 1999, the ecological restoration project has a great impact on the ecosystem service value in China. However, it is still unclear how the temporal and spatial characteristics of ecosystem service value (ESV) in karst areas in southwestern China have changed before and after the ecological restoration project. Land-use data from the five phases from 1980 to 2018 were used in combination with the actual, physical geography and socioeconomic situation of the karst areas in southwestern China. The equivalent factor method and spatial, autocorrelation-analysis method were used to study the temporal and spatial distribution of ESV. The results show that the change trend of ESV in the study area decreased at first and then increased from 1980 to 2018, ESV increased by 19.62 billion yuan, with a growth rate of 0.35%; the ESV in the southwest direction is higher, while in other areas, from west to east, the ESV generally shows a spatial distribution pattern of “high-low-high-Iow”; and the ESV and its changes in the study area from 1980 to 2018 have significant, positive autocorrelations in spatial distribution. The spatial aggregation of ESV among cities is mainly the aggregation of spatially similar values. The results of this study provide reference data for ecological infrastructure construction and ecological economic development in karst areas in southwestern China.

Accurate estimation of the fine-resolution fraction of absorbed photosynthetically active radiation (FPAR) across broad spatial extents and long time periods requires efficient and applicable methods. The existing methods can hardly provide a balance between accuracy, simplicity, and transferability through space and time. Within the remote-sensing trend-surface conceptual framework, this article proposes a scaling-based method to efficiently retrieve FPAR from fine-resolution satellite data using coarse-resolution FPAR products as a reference. The method was particularly developed and applied to Moderate Resolution Imaging Spectroradiometer (MODIS) FPAR product and Landsat imagery. First, necessary prior knowledge related to FPAR retrieval and scaling theories was used to explicitly linearize the complex relationship between MODIS FPAR and Landsat surface reflectance. Second, the explicit linear model for FPAR estimation was trained through one-pair image learning for each date to estimate FPAR from Landsat imagery in real time. Both homogeneous and heterogeneous cases were considered. The method was validated at ten selected worldwide sites from the Validation of Land European Remote Sensing Instruments (VALERI) program and derived an overall root mean squared error (RMSE) of 0.133. A long time series of FPAR data set at the 30-m resolution was generated at the regional scale (approximately 2000 km2) for 13 years (2000–2012). The results were accurate (RMSE = 0.072) and MODIS-consistent, which were significantly better than those of the normalized difference vegetation index (NDVI) downscaling-based and regression tree methods. The scaling-based method provides accurate, MODIS-consistent and spatially consistent FPAR estimates in real time, is highly transferrable through space and time, and allows for future extension of FPAR estimates to the era of the Landsat series satellites.

The absolute radiometric calibration of remote sensing sensors is crucial to the accurate retrieval of biogeophysical parameters through remote sensing. The radiometric calibration uncertainty is the index that describes the reliability of a calibration result and is usually empirically determined by assuming that all of the factors involved are independent of each other. Through a field campaign carried out in Inner Mongolia, China, which aimed to accurately calibrate remote sensing sensors, we developed a Monte Carlo method that statistically evaluates the radiometric calibration uncertainty. From Monte Carlo simulations, it was revealed that the overall uncertainty is much smaller than the root sum of squares of each factor, suggesting that there is some negative correlation among some of the factors. For a surface with a low reflectance (∼5%), the radiometric calibration uncertainty was ∼7.0%, whereas for a surface with a reflectance larger than 20%, the uncertainty was stable at ∼3.0%. This result suggests that the quality of remote sensing data should be carefully examined for surfaces with a low reflectance.

The Airborne Hyperspectral Imager (AHI) is a civilian remote sensor made in China, which uses a multichip charge-coupled device (CCD) butting technology to enlarge the frame. With the aging of the sensor, overall nonuniformity and stripe noises are observed. To solve these problems, a new idea of using the outdoor calibration and the image statistics together was developed to correct AHI images. In this method, outdoor calibration was conducted to calculate the relative radiometric calibration coefficient of each CCD detector for correcting the overall nonuniformity of an image. Then, image statistics were used to remove the residual stripes. This paper presents the principle, the experiment, and the results of this method. To validate the effectiveness of the proposed method, it was compared with some other methods and was evaluated by three quality indices. Moreover, the effects of each step of this method were analyzed separately. The experimental results suggest that the proposed method is convenient and practical for the relative radiometric correction of an airborne image.