Abstract: To better apply remote sensing technology in urban planning mapping, high-resolution remote sensing image has become the most important data source in urban mapping. Faced with the urban area with complex feature information and a large of buildings, seamline extraction between two adjacent remote sensing images is of great practical significance when practicing the mosaics for urban mapping. GF-2 remote sensing imagery has the highest spatial resolution and the smallest breadth among domestic satellites. To develop the applications in urban research with GF-2 remote sensing imagery, an automatic seamline detection method is needed to avoid buildings in urban region. Based on GF-2 remote sensing imagery and building footprints, this paper converts seamline detection into shortest path search from start point to goal point. The result shows that fusing building footprint data and A* searching algorithm can automatically generate seamline to avoid building in complex urban context. This method can be widely used in urban mapping with high-resolution remote sensing imagery.

Key words: high-resolution, remote sensing images, seamline extraction, A* algorithm, building avoidance