When the buildings themselves or the environments they are located in change, it poses a cross-domain problem for building extraction. Since the semantics of the same category should be consistent across domains, a memory mechanism can be used to drive models to capture it. A key prerequisite for a satisfactory memory mechanism is that the memory content is highly task-relevant, i.e., the memory mechanism represents domain-invariant semantic features of buildings as much as possible. It is worth noting that images from different domains have diverse styles, which can interfere with the representation of domain-invariant semantic features of buildings. To maximize the effectiveness of memory mechanisms in cross-domain building extraction tasks, it is necessary to reduce the interference of style information. Therefore, we propose a cross-domain extraction method for buildings based on decoupling style-semantic features. It expresses the image style features and building semantic features separately and optimizes the domain-invariant features of buildings by constraining their relationship through an orthogonal loss. Specifically, on the one hand, style features within the same domain as well as semantic features within and across domains are narrowed down through contrastive learning to ensure the consistency of style features extracted within the same domain and the consistency of semantics across domains. On the other hand, the style features and domain-invariant semantic features in the memory mechanism can be stored and reused to fully exploit self-supervised information. Results of the cross-domain experiments show that the proposed method can achieve optimal building extraction.

Lakes are an important component of global water resources. Lake water bodies extraction based on satellite remote sensing mainly utilizes optical or radar data. However, due to the influence of water quality, ground features with low reflectivity, and smooth surface features, it is still challenging to accurately extract water bodies in complex geographic environments. In this work, we proposed a lake water bodies extraction method by fusing Sentinel-1/2 data. Firstly, the proposed method analyzed the difference of the spectral polarization features between water and non-water in complex geographical environment. Then, the spectral polarization and water index were fused to multidimensional features by feature stacking. Finally, support vector machines are used to classify. Six typical lakes (including urban, mountains, and polluted and clean lakes) in China were used to verify the mapping accuracy. The results showed that extracting lake water bodies by fusing Sentinel-1/2 data had a better performance than using optical or radar data solely, all types of lakes achieved better extraction results, the overall accuracy of lake water extraction is improved by 3%, and the error of commission and omission is controlled within 6%. Comparative experiments indicate that combine radar polarization information with spectral information is helpful to improve the accuracy of different types of lakes extraction in complex geographical environment.

The Yunnan–Guizhou Plateau (YGP) is a typical ecologically fragile region in southwest China. Water-erosion desertification (WED) is one of the most significant environmental and socio-economic issues on the YGP and has seriously restricted the socio-economic development of this region. However, the research on monitoring of the desertification trends in this region has been limited to long time-series Landsat imagery. The objectives of this research were to monitor the WED trends on the YGP using time-series Landsat imagery data from 1989 to 2016. In this paper, we present a multi-indicator rule-based method, which was used to map the WED on the YGP during this period. The results show that the addition of multiple indicators improved the WED classification accuracy to 90.61%. Overall, the following results were obtained by using the proposed method. (1) The slight desertification area on the YGP increased from 89,617.09 km² in 1989 to 100,976.47 km² in 2016 with an annual growth ratio (AGR) of 0.48%, the moderate desertification area increased from 80,276.65 km² in 1989 to 90,768.39 km² in 2016 with an AGR of 0.50%, and the severe desertification area increased from 8149.3 km² in 1989 to 13,220.16 km² in 2016 with an AGR of 2.39%. (2) The WED expansion on the YGP can be divided into three stages. Firstly, the total WED area increased slowly from 17.80×10⁴ km² in 1989 to 17.98×10⁴ km² in 2010 with an AGR of 0.05%. Then, the WED rapidly expanded from 17.98×10⁴ km² in 2010 to 20.28×10⁴ km² in 2013 with an AGR of 4.26%. Finally, the WED increased slightly from 20.28×10⁴ km² in 2013 to 20.50×10⁴ km² in 2016 with an AGR of 0.36%. The total areas of the different degrees of WED decreased in 1992, 1998, 2001, and 2004. (3) The driving factors of WED were analyzed based on the Geographically Weighted Regression (GWR) model. We found that precipitation, vegetation area, and gross domestic product have key roles in the processes of desertification reversion and development. However, the regression coefficients between WED and these factors exhibited considerable spatial variations. The regression coefficients of the key driving factors showed different spatial distributions based on the GWR model in the YGP. The research results can provide scientific reference information for the prevention and control of WED in the YGP.

Semantic segmentation of high-resolution remote sensing imagery (HRSI) suffers from the domain shift, resulting in poor performance of the model in another unseen domain. Unsupervised domain adaptive (UDA) semantic segmentation aims to adapt the semantic segmentation model trained on the labeled source domain to an unlabeled target domain. However, the existing UDA semantic segmentation models tend to align pixels or features based on statistical information related to labels in source and target domain data, and make predictions accordingly, which leads to uncertainty and fragility of prediction results. In this paper, we propose a causal prototype-inspired contrast adaptation (CPCA) method to explore the invariant causal mechanisms between different HRSIs domains and their semantic labels. It firstly disentangles causal features and bias features from the source and target domain images through a causal feature disentanglement module. Then, a causal prototypical contrast module is used to learn domain invariant causal features. To further de-correlate causal and bias features, a causal intervention module is introduced to intervene on the bias features to generate counterfactual unbiased samples. By forcing the causal features to meet the principles of separability, invariance and intervention, CPCA can simulate the causal factors of source and target domains, and make decisions on the target domain based on the causal features, which can observe improved generalization ability. Extensive experiments under three cross-domain tasks indicate that CPCA is remarkably superior to the state-of-the-art methods.

Forest disturbance mapping plays an important role in furthering our understanding of forest dynamics. The Landsat-based detection of Trends in Disturbance and Recovery (LandTrendr) algorithm is widely used in forest disturbance mapping. However, it neglects the quality of the temporal trajectory and its change trends for forest disturbance mapping. Therefore, the aim of this paper is to improve LandTrendr (iLandTrendr) for forest disturbance mapping by optimizing its temporal trajectories and the post-processing of detection results. Specifically, the temporal trajectory of complex forest disturbance types was optimized using the Savitzky–Golay (SG) filter with constraints. That is, the smooth value generated from the SG filter for the disturbance year was replaced by the satellite observations when the nonlinear abrupt signals were included in the multi-temporal data. The forest disturbance detected by LandTrendr was further modified using the consistency of spectral variation trends. A case study using iLandTrendr to detect forest disturbance in Yunnan Province was conducted. Compared to the LandTrendr method, which has an overall accuracy (OA) of 35.88%, iLandTrendr generated forest disturbance mapping with an OA of 89.32%, which was significantly higher. The total mapped area of disturbance was 1,985,820.9 km², accounting for 49.69% of the total area. The disturbances were predominately caused by natural factors, such as wildfires, pests and diseases, and forest degradation, accounting for 85.31% of the total disturbed area. iLandTrendr can quickly and accurately detect the occurrence year of complex forest disturbance types and can be extended for the forest disturbance mapping of a large area.

The frequent occurrence of landslides poses a serious threat to people’s lives and property. In order to evaluate disaster hazards based on remote sensing images via machine learning means, it is essential to establish an image database with manually labeled boundaries of landslides. However, the existing datasets do not cover diverse types of mountainous landslides. To address this issue, we propose a high-resolution (1 m) diverse mountainous landslide remote sensing dataset (DMLD), including 990 landslide instances across different terrain in southwestern China. To evaluate the performance of the DMLD, seven state-of-the-art deep learning models with different loss functions were implemented on it. The experiment results demonstrate not only that all of these deep learning methods with different characteristics can adapt well to the DMLD, but also that the DMLD has potential adaptability to other geographical regions.