Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6938086 | Journal of Visual Communication and Image Representation | 2018 | 38 Pages |
Abstract
Optical flow techniques have been applied to motion object detection under a moving camera over the years. In this paper, we propose an effective motion object detection method based on optical flow estimation, characterized in that the complete boundary of the motion object can be extracted from the combination of the optimized horizontal flow with the optimized vertical flow. The optimized horizontal flow and the optimized vertical flow can be obtained as follows: introducing a third frame into the two frames to obtain the horizontal flow and the vertical flow, and then optimizing the horizontal flow and the vertical flow by applying gradient function and threshold method, respectively. The complete boundary of the motion object, after being subjected to region filling, could be optimized by Gaussian filtering technique to obtain the final detection results. Our proposed method is tested on the following datasets consisting of Cdnet2014, KITTI2015, MPI Sintel datasets, natural YouTube sequences and the collected data, respectively. The results show that our proposed method outperforms state-of-the-art significantly.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Vision and Pattern Recognition
Authors
Yugui Zhang, Jin Zheng, Chi Zhang, Bo Li,