Research on Robots in Competition Systems Based on Monocular Vision Localization and Moving Object Detection Algorithms
The robot competition system has strong real-time performance,which not only requires accurate positioning ability,but also the ability to detect and recognize mobile robots.Research focuses on the positioning and detection issues in student robot competitions.Based on the Visual Background Extractor(ViBe)algorithm for feature matching,combined with directional FAST fea-ture point detection and rotating BRIEF descriptors,the competition system is improved to achieve accurate positioning of the student competition system.The results show that the improved ViBe algorithm has the highest accuracy,with an average accuracy of 87.34％;The improved ViBe algorithm has the lowest false detection rate,with an average false detection rate of 9.12％.The im-proved ViBe algorithm can effectively recognize ghosts and perform elimination processing,while also eliminating the impact of sudden changes in lighting,effectively locating and recognizing robots.The results indicate that the research method has superiority and can improve the matching speed,which is of great significance for robot localization and motion target detection in student competition sys-tems.
万方数据
维普
