Combination of Motion and Edge Cues for Moving Object Detection
This paper proposes an automatic system to detect independent moving objects in image sequences taken from an airborne platform. The algorithm adopts an incremental refinement scheme consisting of motionbased candidate moving regions segmentation and edgeaided object contour refinement. The gray differences in a period of past time are registered and accumulated to the current frame to build the motion images and the candidate regions are extracted by hypothesis testing. Then, the edges are extracted within each region to locate the object. One advantage of this approach is that it exploits the efficiency of motion for moving regions extraction and the accuracy of the edges for localization. Experiment results of aerial videos demonstrate the performance and robustness of the proposed method.
Moving Object Detection Airborne Platform Motion Enhanced Image Hypothesis Testing
Jiang Guo Lei Ma Cheng-fei Zhu Hong-xing Chang
Integrated Information System Research Center Institute of Automation, Chinese Academy of Sciences B Integrated Information System Research Center Institute of Automation Chinese Academy of Sciences Be
国际会议
秦皇岛
英文
106-109
2010-11-05（万方平台首次上网日期，不代表论文的发表时间）