Learning from Noisy Pseudo-labels for All-Weather Land Cover Mapping
This work addresses the challenge of accurate land cover mapping in remote sensing, particularly under adverse weather conditions, but it is incremental as it builds on existing pseudo-label generation techniques.
The paper tackles the problem of noisy pseudo-labels in semantic segmentation of SAR images for all-weather land cover mapping by introducing a method that incorporates semi-supervised learning, image resolution alignment augmentation, and a symmetric cross-entropy loss, achieving first place in the GRSS data fusion contest.
Semantic segmentation of SAR images has garnered significant attention in remote sensing due to the immunity of SAR sensors to cloudy weather and light conditions. Nevertheless, SAR imagery lacks detailed information and is plagued by significant speckle noise, rendering the annotation or segmentation of SAR images a formidable task. Recent efforts have resorted to annotating paired optical-SAR images to generate pseudo-labels through the utilization of an optical image segmentation network. However, these pseudo-labels are laden with noise, leading to suboptimal performance in SAR image segmentation. In this study, we introduce a more precise method for generating pseudo-labels by incorporating semi-supervised learning alongside a novel image resolution alignment augmentation. Furthermore, we introduce a symmetric cross-entropy loss to mitigate the impact of noisy pseudo-labels. Additionally, a bag of training and testing tricks is utilized to generate better land-cover mapping results. Our experiments on the GRSS data fusion contest indicate the effectiveness of the proposed method, which achieves first place. The code is available at https://github.com/StuLiu/DFC2025Track1.git.