Instance segmentation of buildings using keypoints
This work addresses the need for precise building segmentation in remote sensing interpretation, offering an incremental improvement over existing methods.
The paper tackles the problem of blurred boundaries in building instance segmentation from remote sensing images by proposing a network that detects keypoints and forms them into closed polygons, achieving better performance on the AIRS dataset compared to state-of-the-art methods.
Building segmentation is of great importance in the task of remote sensing imagery interpretation. However, the existing semantic segmentation and instance segmentation methods often lead to segmentation masks with blurred boundaries. In this paper, we propose a novel instance segmentation network for building segmentation in high-resolution remote sensing images. More specifically, we consider segmenting an individual building as detecting several keypoints. The detected keypoints are subsequently reformulated as a closed polygon, which is the semantic boundary of the building. By doing so, the sharp boundary of the building could be preserved. Experiments are conducted on selected Aerial Imagery for Roof Segmentation (AIRS) dataset, and our method achieves better performance in both quantitative and qualitative results with comparison to the state-of-the-art methods. Our network is a bottom-up instance segmentation method that could well preserve geometric details.