Abstract
Obtaining timely and reliable built-up area (BUA) information across extensive geographical zones holds crucial significance for understanding environmental change and human activities. Built-up areas often exhibit detailed textures and structures in high-resolution imagery but also present strong heterogeneity. Current methods for BUA extraction primarily relied on planar information from single-view imagery, struggling to effectively capture the three-dimensional (3D) attributes of urban landscapes. Therefore, to address this challenge, this paper proposes a Cross-Angle Propagation Network (CAPNet) based on multi-view remote sensing stereo observation imagery. Our contributions are threefold: 1) We propose the Cross-Angle Fusion Module (CAFM) to exploit BUA’s complementary spatial-spectral-angular context across different viewing angles. This module leverages attention mechanisms for the automated acquisition of multi-angle feature representation learning from diverse angle combinations. 2) We propose a Multi-Angular Propagation Decoder (MAPD) that pioneers the exploration of gradually propagating multi-angle disparity information through bidirectional-adjacent feature fusion across hierarchical levels. 3) We construct a large-scale, high-resolution multi-view Built-Up Area dataset (MVBA) over China’s 41 major cities based on the ZY-3 satellites. Extensive experiment results on MVBA and the public WV-3 multi-view semantic stereo datasets verify CAPNet’s superiority to existing state-of-the-art (SOTA) models, on preserving overall BUA shape, edge, and internal structures. The dataset and the source code of CAPNet will be publicly available at https://github.com/zuo-ux/Cross-Angle-Propagation-Network.
Funding Information
  • National Key Research and Development Program of China (2022YFB3903701)
  • National Natural Science Foundation of China (42271328, 42071311)

This publication has 58 references indexed in Scilit: