From High-resolution Remote Sensing to Tower-mounted Video Surveillance System: Dual Feature Image Matching-Based Precise Positioning Method in Complex Suburban Areas

Abstract. With the continuous advancement of urbanization, the demand for comprehensive monitoring of diverse suburban areas is increasing, which simultaneously raises the need for high-precision positioning technologies especially for the automatic positioning for video surveillance imagery. The current challenges in high-precision positioning for surveillance data include: (1) The difficulty of matching wide-angle images; (2) The significant scale differences between Tower-mounted Video Surveillance (TVS) and Remote Sensing (RS) ortho-images complicate multi-scale automatic focusing; (3) The positioning accuracy decreases due to the physical factors in suburban scene areas. To address these issues, this study proposes a Dual-Feature Image Matching-based Method for High-Resolution RS and TVS Collaboration method for precise positioning in complex suburban areas, referred to as DFMC. Through the processes of geo-referencing, dual-feature image matching, homography transformation and geographic coordinate computation, the proposed DFMC overcomes the challenges of high-precision matching between wide-angle video images and RS images, facilitating the projection of any point coordinates from the pixel coordinate system of TVS frames to the projected Cartesian coordinate system of RS ortho-images. Experiments are conducted using over 28,000 video frames of tower-mounted surveillance system across 210,000 km² in Hunan Province of China. The results indicate the proposed DFMC achieves a positioning accuracy error of less than 1.5 meters in flat areas and less than 3 meters in complex suburban areas. Therefore, DFMC enables rapid monitoring and positioning in complex suburban areas, providing valuable informational support for relevant authorities.