基于DM6437的运动目标跟踪系统研究  

Research on moving target tracking system based on DM6437

作　　者：袁晓东[1] 牛晰[1] 

机构地区：[1]郑州航空工业管理学院,河南郑州450000

出　　处：《现代电子技术》2017年第7期52-54,58,共4页Modern Electronics Technique

基　　金：国家自然科学基金(41001235);河南省教育厅科学技术研究重点项目(12B520061)

摘　　要：为了克服智能视频监控系统可视范围有限,单纯硬件改进增加系统成本的问题,提出一种运动目标跟踪系统,运用混合高斯背景模型的方法实现运动目标的提取,采用卡尔曼滤波器和均值漂移算法相结合实现运动物体的跟踪算法。使用TI公司达芬奇芯片中的DM6437对该算法进一步优化改进,并实现了运动目标跟踪系统,通过对运动小车跟踪测试可知,当小车和背景在RGB色彩空间中反差很大时,跟踪的效果理想;反差很小时,跟踪的失败率会有所提高,整体跟踪效果良好。Since the visible range of the intelligent video monitoring system is limited, and the improvement of the pure hardware can increase the cost of the system, a moving target tracking system is put forward. The method of using the mixed Gaussian background model is used to extract the moving target. The Kalman filter and mean-shift algorithm are combined to re- alize the moving target tracking algorithm. The DM6437 of Leonardo Da Vinci chip made by TI Company is used to optimize and improve the algorithm further, and implement the moving target tracking system, The moving car was performed with tracking test. The conclusion shows that the tracking effect is satisfied when the car and background have strong contrast in RGB color space, and the tracking failure rate will be improved and the overall tracking effect is good when the car and background have poor contrast.

关 键 词：智能跟踪 KALMAN滤波器 均值漂移算法 达芬奇芯片 

分 类 号：TN953[电子电信—信号与信息处理]