Vicarious calibration and validation techniques are important tools to ensure the long-term stability and inter-sensor consistency of satellite sensors making observations in the solar-reflective spectral domain. Automated test sites, which have continuous in situ monitoring of both ground reflectance and atmospheric conditions, can greatly increase the match-up possibilities for a wide range of space agency and commercial sensors. The Baotou calibration and validation test site in China provides operational high-accuracy and high-stability vicarious calibration and validation for high spatial resolution solar-reflective remote-sensing sensors. Two sites, given the abbreviations BTCN (an artificial site) and BSCN (a natural sandy site), have been selected as reference sites for the Committee on Earth Observation Satellites radiometric calibration network (RadCalNet). RadCalNet requires sites to provide data in a consistent format but does not specify the required operational conditions for a RadCalNet site. The two Baotou sites are the only sites to date that make spectral measurements for their continuous operation. One of the core principles of RadCalNet is that each site should have a metrologically rigorous uncertainty budget which also describes the site’s traceability to the international system of units, the SI. This paper shows a formalized metrological approach to determining and documenting the uncertainty budget and traceability of a RadCalNet site. This approach follows the Guide to the Expression of Uncertainty in Measurement. The paper describes the uncertainty analysis for bottom-of-atmosphere and top-of-atmosphere reflectance in the spectral region from 400 to 1000 nm for the Baotou sites and gives preliminary results for the uncertainty propagating this to top-of-atmosphere reflectance.

Land-cover information is significant for land-use planning, urban management, and environment monitoring. This paper presented a novel extended topology-preserving segmentation (ETPS)-based multi-scale and multi-feature method using the convolutional neural network (EMMCNN) for high spatial resolution (HSR) image land-cover classification. The EMMCNN first segmented the images into superpixels using the ETPS algorithm with false-color composition and enhancement and built parallel convolutional neural networks (CNNs) with dense connections for superpixel multi-scale deep feature learning. Then, the multi-resolution segmentation (MRS) object hand-delineated features were extracted and mapped to superpixels for complementary multi-segmentation and multi-type representation. Finally, a hybrid network was designed to consist of 1-dimension CNN and multi-layer perception (MLP) with channel-wise stacking and attention-based weighting for adaptive feature fusion and comprehensive classification. Experimental results on four real HSR GaoFen-2 datasets demonstrated the superiority of the proposed EMMCNN over several well-known classification methods in terms of accuracy and consistency, with overall accuracy averagely improved by 1.74% to 19.35% for testing images and 1.06% to 8.78% for validating images. It was found that the solution combining an appropriate number of larger scales and multi-type features is recommended for better performance. Efficient superpixel segmentation, networks with strong learning ability, optimized multi-scale and multi-feature solution, and adaptive attention-based feature fusion were key points for improving HSR image land-cover classification in this study.

Visible hyperspectral image (V-HSI) and thermal infrared hyperspectral image (TI-HSI) have been crucial data sources for land cover classification. V-HSI can directly provide the information of land surface, such as the shape, color, texture, and other features. TI-HSI contains rich long-wave spectral information, which can reflect the unique emission characteristics of ground objects in the thermal infrared spectral range. To fully leverage the advantages of V-HSI and TI-HSI while enhancing the classification accuracy, this paper proposes a self- and cross-attention enhanced transformer network (SCAET), integrated with convolutional neural network (CNN) for HSI classification. Initially, the proposed method employs a dual-branch spatial-spectral CNN (SS CNN) to extract spectral convolution features from V-HSI and TI-HSI respectively. Subsequently, a spectral feature mapping (SFM) module is proposed to perform feature transformation, extracting independent and interactive features of V-HSI and TI-HSI. Then, a self- and cross-attention interactive enhancement (SCIE) module is designed to extract deeper features and enhance the independent features by using the interactive features. Additionally, a self-projection mixing (SPM) module is formulated to promote feature interaction and improve the generalization capability of the model. To validate the effectiveness of the proposed network, extensive experiments are conducted on real-world datasets, and the results indicate that SCAET significantly outperforms current multi-source fusion networks.

Thermal infrared hyperspectral imagery presents a superior capability for capturing intricate spectral details of atmospheres and ground objects compared to multispectral images, thus offering a more nuanced dataset for land surface temperature (LST) retrieval. However, extensive interband correlations pose computational challenges and undesirable “dimension disaster” problems. To address this issue, this article proposes a purpose-built framework of thermal infrared hyperspectral band selection (BS) using a graph neural network for LST retrieval. Specifically, the thermal infrared hyperspectral data is first mapped onto a graph topology, followed by feeding it into a graph attention module with brightness temperature constraints to extract band features. Following this, the extracted band features undergo a comprehensive analysis through a multiscale convolution module consisting of convolution kernels with multiple sizes, which have more variety and larger receptive fields for calculating the correlation between different band features, assigning different weights to each band. Finally, a weight selection module is designed to filter the bands based on their assigned weights, creating a subset of bands with greater significance for LST retrieval. Training the designed model, 65100 observations are simulated utilizing MODTRAN, with 80% allocated for training and 20% for testing. The experimental results validate the effectiveness of the proposed model, with a root mean square error (RMSE) of 1.85 K in practical applications on IASI imagery. This accomplishment substantiates the model’s capacity to reliably employ a judiciously selected subset of thermal infrared hyperspectral bands for LST retrieval applications, thus offering a promising contribution to the advancement of thermal infrared hyperspectral image processing methodologies.

The Chinese second-generationHuanjing disaster monitoring satellite (HJ-2A) was launched on September 27, 2020, and underutilized due to the lack of accurate operational methodologies for land surface temperature (LST) retrieval.In this article, an operational LST retrieval method is proposed to retrieve LSTs from HJ-2A thermal infrared observations. The LST retrieval methodology involves two main steps. The land surface emissivities (LSEs) over all land cover types are obtained with the improved normalized difference vegetation index-based threshold method, and then the LST is retrieved operationally from the adjacent infrared bands.The algorithm coefficients for LST retrieval are from regression analysis of radiative transfer simulations, and LSTs could be retrieved based on thermal images without any additional auxiliary data. The simulation results demonstrated that the root-mean-square errors (RMSEs) of LST retrieval were less than 2.4 K in all subranges, and the minimum RMSE for the two emissivity groups (high- (low-) emissivity group) was 0.16 K (0.20 K) and appeared in the tractable subrange with water vapor content (WVC) varying from 0 to 1.5 g/cm ² and view zenith angle (VZA) being 0°. Furthermore, an error analysis was performed, the results showed that the LSE, NEΔT, and atmospheric water vapor uncertainty of 1%, 0.2 K, and 20% caused the LST retrieval errors with 0.88–1.21 K (0.84–1.19 K), 0.1 K (0.09 K), and 0.006 K (0.008 K) for the high- (low-) emissivity group, respectively, with WVC∊[0–1.5] g/cm ² and VZA = 0°. Finally, the retrieved LSTs were applied to seven images of the Wuhai, Geermu, Dunhuang, and Baotou sites from January to March and cross validated by the moderate resolution imaging spectroradiometer (MODIS) LST products. From the cross-validated results, it can be found that the RMSEs of the retrieved LSTs and the MODIS LST products were between 2.3 and 3.7 K, and the mean RMSE value was 2.89 K.

Potatoes, a crucial staple crop, face significant threats from late blight, which poses serious risks to food security. Despite extensive research using ground and unmanned aerial vehicle (UAV) hyperspectral data for crop disease monitoring, satellitescale identification of diseases like Potato Late Blight (PLB) remains limited. This study employs a multi-scale analysis approach, integrating high-resolution Sentinel-2 multispectral satellite data with UAV and ground spectral data, to monitor and identify PLB. A key finding of this study is the general similarity in spectral patterns across different scales, with consistent valley values in bands of Blue and Red and peak values in bands of Near Infrared and Narrow near Infrared, accompanied by a consistent decrease in reflectance correlating with increasing disease severity. Furthermore, the study highlights scaledependent spectral variations, with changes in bands of Vegetation Red Edge2, Vegetation Red Edge3, Near Infrared and Narrow Near Infrared being more pronounced at the ground scale compared to UAV and satellite scales. Based on the developed Red Edge Index and Disease Stress Index with a suite of machine learning algorithms, we proposed a XGBoost-based model integrating spectral indices for PLB monitoring (PLB-SI-XGBoost). Notably, the proposed model demonstrated the highest average evaluation score of 0.88 and the lowest root mean square error (RMSE) of 13.50 during ground scale validation, outperforming other algorithms. At the UAV scale, the proposed model achieved a robust Rsquared value of 0.74 and an RMSE of 18.27. Moreover, the application of Sentinel-2 data for disease detection at satellite scale yielded an accuracy of 70% in the model. The results of the study emphasize the importance of scale in disease monitoring models and illuminate the potential for satellite-scale surveillance of PLB. The exceptional performance of PLB-SI-XGBoost model in detecting PLB suggests its utility in enhancing agricultural decisionmaking with more accurate and reliable data support.