基于TMS320DM6437的运动目标实时检测与跟踪  

Real-time checking and tracking of moving object based on TMS320DM6437

作　　者：蒋建国[1] 李明[1] 齐美彬[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2011年第7期1007-1010,1023,共5页Journal of Hefei University of Technology：Natural Science

基　　金：安徽省科技计划资助项目(080203095)

摘　　要：文章提出了一种基于TMS320DM6437图像处理平台上的运动目标检测与跟踪算法。为能从运动的背景中检测其中的运动目标并跟踪,该文给出一种基于块匹配的背景更新与帧差法相结合的算法,并在此基础上引入Canny算法中的边界连通思想;在图像分割中引用最大类间方差法进行自适应阈值的二值化,用4邻域交叉搜索法取代8邻域搜索法,采用分片NMI特征对目标进行特征匹配,并根据运动信息结合卡尔曼预测实现对目标的跟踪;实验结果成功地证明了此算法的有效性和可行性。A checking and tracking algorithm of moving objects based on TMS320DM6437 is proposed in this paper.In order to check and track the moving objects from the moving background,an algorithm which integrates background update based on block matching with frame difference method is proposed,and the boundary connectivity idea of Canny algorithm is quoted.Otsu threshold method is adopted for image segmentation to binarize the self-adaptive threshold value.The eight-neighbor searching method is replaced by the four-neighbor-alternating searching method,and segmental NMI feature is used to match the object.The tracking of the object is implemented by Kalman prediction method based on object motion information.The experimental results demonstrate the effectiveness and feasibility of this algorithm.

关 键 词：背景更新 4邻交叉搜索 CANNY算法 卡尔曼预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]