基于图像处理技术的运动轨迹自动检测研究  

Research on motion trajectory automatic detection based on image processing technology

作　　者：陈晨[1,2] 张炜[2] CHEN Chen;ZHANG Wei(Inner Mongolia Normal University,Hohhot 010010,China;Baotou Railway Vocational&Technical College,Baotou 014060,China)

机构地区：[1]内蒙古师范大学,内蒙古呼和浩特010010 [2]包头铁道职业技术学院,内蒙古包头014060

出　　处：《现代电子技术》2021年第5期70-74,共5页Modern Electronics Technique

摘　　要：针对传统方法进行运动轨迹的自动检测速率低的问题,提出基于图像处理技术的运动轨迹自动检测方法。首先通过高斯背景混合模型对图像视频序列进行预处理,提取运动目标;然后采用平滑运动模型描述运动目标映射代价,计算运动模式,分割垂直区域目标平行切片,进行时空切片扫描从而对其进行层次聚类;最后对聚类后的运动目标数据进行降维,通过Hough变换程序实现运动轨迹自动检测。在相同条件下,与传统运动轨迹检测方法进行对比实验,实验结果证明,所设计方法提高了运动轨迹的检测速率,而且提高了运动轨迹检测精度,相对于传统方法,运动轨迹检测性能得到了提升,具有更高的实际应用价值。In view of the low speed of the traditional method in the motion trajectory automatic detection,a motion trajectory automatic detection method based on image processing technology is proposed.The Gaussian background mixture model is used to preprocess the image video sequence and extract the moving object.And then,the smooth motion model is used to describe the mapping cost of the moving object,calculate the motion mode,segment the paralleled object slices in the vertical area and scan the spatio⁃temporal slices,so as to carry out hierarchical clustering of the motion object.Finally,the data of the clustered moving objects are subjected to dimensionality reduction.The Hough transform program is used to realize the automatic detection of the motion trajectory.Comparative experiments were carried out to compare the proposed method with the traditional motion trajectory detection methods under the same conditions.Experimental results show that both the detection speed and detection accuracy of the proposed method are improved in comparison with the traditional methods.With an improved detection performance,it has a higher practical application value.

关 键 词：运动轨迹检测 图像处理技术 运动目标提取 映射代价描述 运动模式计算 数据降维 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP309.7[电子电信—信息与通信工程]