非连续轨迹下的公路车辆智能跟踪技术研究

Research on intelligent tracking technology of highway vehicle under discontinuous trajectory

摘要:

针对运动车辆轨迹跟踪控制中的遮挡问题,研究创新地提出以SVM分类检测器优化STAPLE跟踪算法,以保证在目标退出遮挡时可以重新搜索并定位目标；同时对颜色特征直方图、HOG算法进行了改进,以提高算法特征提取效率；最后选择VOT2016国际标准序列集对优化后的STAPLE跟踪算法进行验证。研究结果显示,改进STAPLE算法能在目标退出遮挡后更为快速地重新捕捉目标位置；改进算法对目标中心的跟踪精度达到了0.81,对目标框的跟踪精度达到了0.89；在目标进入遮挡状态时,改进算法的跟踪精度最高。这次研究提出的STAPLE优化算法表现出较好的跟踪效果,其面对长时遮挡的跟踪能力具有较高的应用价值。

Abstract:

Aiming at the occlusion problem in the trajectory tracking control of moving vehicles, this paper innovatively proposes a SVM classification detector optimized stable tracking algorithm to ensure that the target can be re searched and located when the target exits the occlusion; at the same time, the color histogram and hog algorithm are improved to improve the efficiency of feature extraction; finally, vot2016 international standard sequence set is selected to optimize the sta Ple tracking algorithm is verified. The results show that the improved cascade algorithm can capture the target position more quickly after the target exits the occlusion; the tracking accuracy of the improved algorithm for the target center reaches 0.81, and the tracking accuracy of the target frame reaches 0.89; when the target enters the occlusion state, the improved algorithm has the highest tracking accuracy. The stable optimization algorithm proposed in this study shows good tracking effect, and its tracking ability in the face of long-term occlusion has high application value.