Remotely sensed land surface temperature (LST) with spatial continuity and high spatiotemporal resolution (hereafter referred to as high resolution ) is a crucial parameter for studying the thermal environment and has important applications in many fields. However, adverse atmospheric conditions, sensor malfunctioning, and scanning gaps between orbits frequently introduce spatial discontinuities into satellite-retrieved LST products. For a single sensor, a tradeoff occurs between temporal and spatial resolutions; therefore, it is almost impossible to obtain images in high resolution.

Urban water is important for the urban ecosystem. Accurate and efficient detection of urban water with remote sensing data is of great significance for urban management and planning. In this paper, we proposed a new method by combining Google Earth Engine (GEE) with a multiscale convolutional neural network (MSCNN) to extract urban water from Landsat images, which can be summarized as “offline training and online prediction” (OTOP). That is, the training of MSCNN is completed offline, and the process of urban water extraction is implemented on GEE with the trained parameters of MSCNN. The OTOP can give full play to the respective advantages of GEE and the convolutional neural network (CNN), and can make the use of deep learning method in GEE more flexible. The proposed method can process the available satellite images with high performance, without data download and storage, and the overall performance of urban water extraction in the test areas is also higher than that of the modified normalized difference water index (MNDWI) and random forest classifier. The results of the extended validation in the other major cities of China also showed that OTOP is robust and can be used to extract different types of urban water, which benefits from the structural design and training of MSCNN. Therefore, OTOP is especially suitable for the study of large-scale and long-term urban water change detection in the background of urbanization.

Satellite remote sensing has been increasingly employed for the estimation of ground-level atmospheric PM $_2.5$ . There have been several cross-validation (CV) approaches applied for the validation of satellite-based PM $_2.5$ estimation models. However, these validation approaches often lead to confusion, due to the unclear applicable conditions. For this, we fully analyze and assess the existing validation approaches, and provide suggestions on applicable conditions for them. Furthermore, the existing validation approaches still have limitations to disregard the uneven distribution of ground stations, and tend to overestimate the performance of the PM $_2.5$ estimation models. To this end, a CV-based validation approach considering the uneven spatial distribution of monitoring stations (denoted as SDCV) is proposed. SDCV introduces the spatial distance between validation station and modeling station into the CV process, and evaluates the spatial performance through a strategy of excluding modeling stations within a specific distance. Meanwhile, this approach has designed reasonable evaluation indices for the model validation. Taking China as a case study, the results indicate that SDCV can yield a more complete and effective evaluation for the popular PM $_2.5$ estimation models than the traditional validation approaches.

After the Wenchuan earthquake in China, debris flows have been more frequent. Multiple debris flows commonly occur in earthquake-affected areas during heavy rainfall, often causing losses of lives and property. We analyzed the dynamic movement of debris flows and proposed a quantitative method of debris flow hazard assessment, based on kinetic energy. Validated using field study of an actual debris flow disaster, these analyzes help determine the type, quantity, distribution, economic worth, and susceptibility of elements at risk. We established a method to determine vulnerability of elements at risk and we propose a systematic and quantitative method for local risk analysis of debris flows. We applied the proposed method to a debris flow in Qipan gully, which caused serious damages for Duwen Highway and Qipan settlements of Sichuan Province in 2013. With the support of a debris-flow movement numerical simulation, remote sensing, and GIS techniques, the proposed method analyzed disaster effects and divided the hazardous areas into three risk zones. Calculated risk zones coincided with the actual distribution and severity of damage of the event, suggesting that the risk assessment generated by the proposed method is consistent with results from the actual disaster.

Solid waste is a widespread problem that is having a negative effect on the global environment. Owing to the ability of macroscopic observation, it is reasonable to believe that remote sensing could be an effective way to realize the detection and monitoring of solid waste. Solid waste is usually a mixture of various materials, with a randomly scattered distribution, which brings great difficulty to precise detection. In this article, we propose a deep learning network for solid waste detection in urban areas, aiming to realize the fast and automatic extraction of solid waste from the complicated and large-scale urban background. A novel dataset for solid waste detection was constructed by collecting 3192 images from Google Earth (with a resolution from 0.13 to 0.52 m), and then a location-guided key point network with multiple enhancements (LKN-ME) is proposed to perform the urban solid waste detection task. The LKN-ME method uses corner pooling and central convolution to capture the key points of an object. The location guidance is realized through constraining the key point locations situated of the annotated bounding box of an object. Multiple enhancements, including data mosaicing, an attention enhancement, and path aggregation, are integrated to improve the detection accuracy. The results show that the LKN-ME method can achieve a state-of-the-art AR ₁₀₀ (the average recall computed over 100 detections per image) of 71.8% and an average precision of 44.0% for the DSWD dataset, outperforming the classic object detection methods in solving the solid waste detection problem.

Recently, deep learning has been reported to be an effective method for improving hyperspectral image classification and convolutional neural networks (CNNs) are, in particular, gaining more and more attention in this field. CNNs provide automatic approaches that can learn more abstract features of hyperspectral images from spectral, spatial, or spectral-spatial domains. However, CNN applications are focused on learning features directly from image data—while the intrinsic relations between original features, which may provide more information for classification, are not fully considered. In order to make full use of the relations between hyperspectral features and to explore more objective features for improving classification accuracy, we proposed feature relations map learning (FRML) in this paper. FRML can automatically enhance the separability of different objects in an image, using a segmented feature relations map (SFRM) that reflects the relations between spectral features through a normalized difference index (NDI), and it can then learn new features from SFRM using a CNN-based feature extractor. Finally, based on these features, a classifier was designed for the classification. With FRML, our experimental results from four popular hyperspectral datasets indicate that the proposed method can achieve more representative and objective features to improve classification accuracy, outperforming classifications using the comparative methods.