Graph-Cut RANSAC for Outlier Rejection
Graph-Cut RANSAC is an innovative method for outlier rejection in Visual Odometry field.Research aboutoutlier rejectionhas been gradually developedin visual odometer field.These technologies have been successfully applied in the fields of visual odometer, ORB-SLAM, point set registration, and 3D reconstruction.However, current algorithmsthemselves have the disadvantages of inefficiency, poor robustness and insufficient precision.In this context, the Graph-Cut RANSAC algorithm has become a hot topic of interest to researchers in recent years.The critical part of GC-RANSAC is the theorem of Graph cuts.Such method converts the outlier point rejection problem to the minimum cut problem of the graph which equals minimizing the energy function.In order to test the performance of GC-RANSAC,we run this algorithm to estimate fundamental matrixof real image pairs.With a comparison of Plain RANSAC and LO-RANSAC, the experimental results show that GC-RANSAChas a significant improvement in accuracy with no notably deterioration inprocessing time.
Visual Odometry(VO) Energy Minimization GC-RANSAC Algorithm
Li Ao Wang Jikai Chen Zonghai
Department of Automation, University of Science and Technology of China, Anhui, Hefei, 230026
国内会议
第20届中国系统仿真技术及其应用学术年会(20th CCSSTA 2019)
合肥
英文
476-480
2019-08-01（万方平台首次上网日期，不代表论文的发表时间）