With the development of deep learning algorithms, more and more deep learning algorithms are being applied to remote sensing image classification, detection, and semantic segmentation. The landslide semantic segmentation of a remote sensing image based on deep learning mainly uses supervised learning, the accuracy of which depends on a large number of training data and high-quality data annotation. At this stage, high-quality data annotation often requires the investment of significant human effort. Therefore, the high cost of remote sensing landslide image data annotation greatly restricts the development of a landslide semantic segmentation algorithm. Aiming to resolve the problem of the high labeling cost of landslide semantic segmentation with a supervised learning method, we proposed a remote sensing landslide semantic segmentation with weakly supervised learning method combing class activation maps (CAMs) and cycle generative adversarial network (cycleGAN). In this method, we used the image level annotation data to replace pixel level annotation data as the training data. Firstly, the CAM method was used to determine the approximate position of the landslide area. Then, the cycleGAN method was used to generate the fake image without a landslide, and to make the difference with the real image to obtain the accurate segmentation of the landslide area. Finally, the pixel-level segmentation of the landslide area on remote sensing image was realized. We used mean intersection-over-union (mIOU) to evaluate the proposed method, and compared it with the method based on CAM, whose mIOU was 0.157, and we obtain better result with mIOU 0.237 on the same test dataset. Furthermore, we made a comparative experiment using the supervised learning method of a u-net network, and the mIOU result was 0.408. The experimental results show that it is feasible to realize landslide semantic segmentation in a remote sensing image by using weakly supervised learning. This method can greatly reduce the workload of data annotation.

Monitoring the spatio-temporal variation of the land-use/land cover change (LULC) and ecosystem service value (ESV) changes will help achieve regional sustainable development and management. Derong County is a part of the Hengduan Mountains area, the most crucial ecological functional area in China, and LULC has changed tremendously in the past 30 years. However, the effects of LULC changes on ecosystem services is not well understood. Based on 1992, 1995, 2005, 2013, and 2018 remote sensing images, we used visual interpretation to obtain LULC data and used global value coefficients and modified local value coefficients to assess the spatial-temporal changes of ESV and LULC from 1992 to 2018. The results showed that: (1) From 1992 to 2018, shrubland and grassland decreased, while built-up land, snow, forestland, water body, and cropland area increased. (2) The ESV with an overall decrease of 0.25 × 10⁸ yuan, ecological projects have played a positive role in improving ESV. In contrast, the main decrease factor of ESV was the increase in agricultural economic development and urban expansion from 1992 to 2018. (3) The ESV spatial distribution indicated the value density of ESV was on the decline, and with the greatest deterioration in Dianyagong. The highest density of ESV area is distributed in Waka, and the lowest density of ESV area is distributed in Bari. This research points out the important role of Derong County in the regional life support system and provides a scientific reference for the sustainable management of dry-hot valley regions’ land resources and ecosystem services.