基于DM642的嵌入式实时车辆跟踪系统

作者：程金汉,杜爱民 日期：2007-12-28/span> 浏览：3158 查看PDF文档

基于DM642的嵌入式实时车辆跟踪系统
程金汉,杜爱民
（同济大学 汽车学院，上海 200092）

摘 要：提出了一种基于DM642的实时车载车辆跟踪系统,该系统采用实时的Mean Shift作为跟踪器。为了提高系统跟踪的准确性，利用车辆的颜色分布和外型特征同时作为跟踪器的输入特征；为了提高跟踪的鲁棒性，利用卡尔曼（Kalman）滤波器对Mean Shift的跟踪系统进行预测和滤波。实验显示，该实时车载车辆跟踪系统具有良好的鲁棒性和实时性。
关键词：车辆辅助驾驶；DM642; Mean Shift;卡尔曼滤波器

A realtime vehicle tracking system based on DM642
CHENG Jinhan, DU Aimin
(College of Vehicle, Tongji University, Shanghai 200092, China)

Abstract: A realtime vehicle tracking system based on DM642 was proposed. System applied the Mean Shift as realtime tracker. In order to improve the system accuracy, both color feature set and shape feature set were used as the tracking input; In order to improve the system robustness, Kalman filter and Mean Shift tracker were made work together. Experiment results show that, the realtime vehicle tracking system has good filter robustness and great efficiency.
Key words: driver assistance system; DM642; Mean Shift; Kalman

参考文献(Reference)：
［1］Texas Instruments. TMS320DM642 Video/Imaging FixedPoint Digital Signal Processor: Data Manul［M］. Texas instruments,1999.
［2］李方慧，王 飞,何佩琨.TMS320C6000系列DSPs原理与应用:第2版［M］.北京:电子工业出版社,2003.
［3］FUKUNAGE K, HOSTETLER L D. The estimation of the gradient of a density function with application in pattern recognition［J］.IEEE Trans. on Information Theory,1975,21(1):32-40.
［4］CHENG Y. Mean Shift, mode seeking and clustering［J］. IEEE Trans. on Pattern Analysis and Machine Intelligence,1995,17(8):790-799.
［5］邓自立.最优滤波理论及其应用——现代时间序列分析方法［M］.哈尔滨：哈尔滨工业大学出版社,2000.
［6］JJLIER S, UHLMANN J K. A New Approach for Filtering Nonlinear Systems［C］//The Proceedings of the American Control Conference, IEEE Press,1995:1628-1632.