The Sustainable Development Goals Science Satellite-1 (SDGSAT-1) is equipped with both low-light and thermal infrared imagers, which can detect infrared radiation and weak light information emitted by vessels. Compared to similar products, its spatial resolution has undergone a significant improvement. Existing remote sensing vessel detection methods only consider the use of single-source data for vessel detection, fail to effectively utilize the complementary information in multisource data, and have difficulty processing the complex vessel shapes of high-resolution satellite data, resulting in unsatisfactory detection results. In view of the sparse distribution of vessels at sea, this article proposes a new vessel target detection method using SDGSAT-1 high-resolution thermal infrared and low-light satellite data. Specifically, guided filtering is used to fuse thermal infrared and low-light data, the background and foreground are separated by partial sum of tensor core norms (PSTNN) model, and then the ordering points to identify the clustering structure (OPTICS) clustering algorithm and intercluster merging are used for detection. This article established a vessel dataset by choosing Shanghai Port, Hong Kong Port, and the Gulf of Mexico and then applied the algorithm. The detection accuracy and recall rate were found to be 97.67% and 97.80% respectively, which were significantly superior to other algorithms. This algorithm overcomes the complex background noise in the dataset and achieves good detection results.

The urban land surface temperature (LST) is very important in urban development, changes and local climate in the city, etc. Many methods have been proposed to inverse LST from satellite remotely sensed data. In this study, an inversion method of combining deep learning and physical model was proposed to estimate the urban surface temperature from CHINESE SDGSAT-1 satellite, i.e., an improved temperature and emissivity separation (TES) algorithm based on sky view factor (SVF). Finally, two data sets were selected to evaluate the accuracy of the proposed algorithm. Results show that the root mean squared errors (RMSEs) using the proposed algorithm are approximately 0.39K for LST and 0.023 for emissivity, respectively. The proposed algorithm is applied to inverse the LST/LSE of Beijing and Wuhan, China. Compared with Landsat-8 satellite products, the proposed algorithm has consistent accuracy. The cross-validated RMSEs with the Landsat-8 surface temperature product in Wuhan and Beijing were 2.18 K and 1.11 K, respectively.

Urban land surface temperature (LST) plays a crucial role in observing and comprehending energy exchange within urban environments. Urban geometric structure and material composition are critical parameters to characterize urban thermal radiation accurately. In this article, an urban thermal radiation analytical model based on the sky view factor (UTRAM-SVF) was developed by considering the multiple scattering within the urban canopy and the radiation composition of urban components. The cross-comparison of the proposed method was conducted in four ways: the discrete anisotropic radiation transfer (DART) model, urban effective emissivity model based on SVF (UEM-SVF), Landsat 8 Thermal Infrared Sensor (TIRS) data, and Airborne Hyperspectral Scanner (AHS) TIR data. Compared with DART, the results revealed that the root-mean-square error (RMSE) of radiance by the UTRAM-SVF model is 0.04 W/(m$^{2}\cdot $sr$\cdot \mu $m). Furthermore, two field applications were conducted using Landsat 8 TIRS data and AHS TIR data. The differences of at-sensor radiance from UTRAM-SVF model and Landsat 8 TIRS data vary from 0.05 to 0.3 W/(m$^{2}\cdot $sr$\cdot \mu $m), and the RMSE is 0.17 W/(m$^{2}\cdot $sr$\cdot \mu $m). The results of AHS TIR images on the UTRAM-SVF model present that the average biases of at-sensor radiance are 0.07 and 0.19 W/(m$^{2}\cdot $sr$\cdot \mu $m) for two TIR channels. The cross-comparison results show that the proposed method outperformed the UEM-SVF model in evaluating radiance over the complex and heterogenous urban areas (SVF <0.4). The UTRAM-SVF model can be used to monitor urban thermal radiation based on high-resolution TIR data and can further be helpful to retrieve high-resolution urban LST.

The separation of land surface temperature (LST) and land surface emissivity (LSE) is an ill-posed problem in thermal infrared (TIR) remote sensing. By building a new observation matrix to compress the LSE unknows and a dictionary training method to reconstruct complete LSE spectra, a new dictionary-based sparse representation for emissivity (DSRE) method has been proposed to retrieve LST and LSE from the atmospherically corrected hyperspectral TIR data. The proposed method fully utilizes the sparsity of compressed sensing and the empirical knowledge of the trained emissivity dictionary. The sensitivity analysis shows that the modeling accuracies of the proposed method are 0.215Kand 0.0060 for LST and LSE, respectively. Even with the instrument noise of 0.3Kand the uncertainties in atmospheric transmittance, atmospheric upwelling, and downwelling radiance of 10 %, the retrieval accuracies are 0.811Kfor LST and 0.0241 for LSE, respectively. Then a field experiment was conducted to validate the proposed method, and a comparison was executed to three published methods, including ASTER temperature-emissivity separation (ASTERTES), linear spectral emissivity constraint TES (LSECTES), and iterative spectrally smooth TES (ISSTES). The accuracies of retrieved LST and spectral LSE are 1.41K/ 0.009, 2.57K/ 0.071, 1.59K/ 0.038, and 2.00K/ 0.077 for DSRE, ASTERTES, LSECTES, and ISSTES. In contrast to the three published methods, our proposed method is more accurate and effective than other published methods. Especially in the atmospheric absorption band, the proposed method has a strong anti-noise capability to the residuals of environmental downwelling radiance.

Remotely sensed hyperspectral data can support more effective water quality monitoring. Nevertheless, the variability and complexity of urban river water make it hard to retrieve comprehensive water quality characteristics directly, so that most current water quality assessments rely on semiempirical, semianalytical, or bio-optical approaches. In this study, we carried out simultaneousin situhyperspectral data and water quality measurements. We used the 382 hyperspectral data from urban rivers of Zhongshan City in the Pearl River Delta to test how well the random forest (RF) and one-dimensional convolutional neural networks (1D-CNNs) algorithms retrieved the newly established water quality index (WQI). The RF algorithm also identified essential wavelengths for retrieving the WQI. Our results demonstrate that the RF and 1D-CNN algorithms performed well in WQI estimations. The 1D-CNN model performed significantly better than the RF model, especially on high WQI samples. Both models were insensitive to smoothing of the hyperspectral data, showing that the noise of the original hyperspectral reflectance data has a limited impact on the algorithms. In addition, when we used the essential wavelength data (mainly located between 580–590 nm and near 722, 751, 821, and 830 nm) as input data, we achieved better retrieval results. The 1D-CNN model performed the best with an$R^{2}$of 0.87, RMSE of 0.574, and RPIQ of 3.082 when we used the top tenth percentile of the essential wavelength data. This study demonstrates the potential of the 1D-CNN algorithm for hyperspectral data analysis to retrieve comprehensive water quality.