Remote Sensing Image Semantic Change Detection Boosted by Semi-supervised Contrastive Learning of Semantic Segmentation

Semantic change detection (SCD) is a challenging task in remote sensing image interpretation, which adopts multi-temporal images to detect, locate, and analyze pixel-level land-cover “from-to” changes. In SCD, the severe class imbalance problem and the occurrence of confusing categories are very typical, making it challenging to accurately distinguish the easily-confused categories with limited semantic context information. However, previous works didn’t address these issues in depth. This paper proposes a novel semantic change detection method named SSCLNet, in which a simple and effective semantic change detection network is designed as a strong baseline, and a semi-supervised contrastive learning module of semantic segmentation is presented to enhance the distinguishability of categories. Our baseline extracts semantic context through HRNet, gets change information simply through an absolute difference, and then directly performs semantic change detection based on the fusion of semantic context and change information. To utilize the semantic context information of the unlabeled non-changed regions, we employ a self-training method for semi-supervised semantic segmentation. To learn distinguishable feature representations for easily-confused categories, we present contrastive learning with an adaptive sampling strategy for semantic segmentation. It selects challenging negative samples for each category from the other categories that exhibit similar features or attributes. The sampling space includes both the labeled changed samples and the non-changed samples predicted by self-training. The comprehensive experiments on the SECOND and the Landsat-SCD dataset demonstrate that the proposed SSCLNet achieves the state-of-the-art performance, with a significant improvement of 2.07% and 4.15% in the Score value, respectively.