Research Of Moving Object Detection And Tracking Technology And DSP Implementation

With the rapid development of Modern Information Technology,ATR(Auto TargetRecognition)system plays an important role in the national economy and defence.A kind of high-speed digital image detection and tracking system is designed inthis paper. Study on the system is to improve processing speed and realize thedetection and tracking real-timely.This paper introduces the process of realization of the detection and trackingsystem in two aspects:the hardware design and the realization of software algorithm.Directed towards object detection, we study and analyse some segmentationalgorithms for Image segmentation. And then we propose a detection algorithm basedon background substraction and region growing. Simulation results are given to showthe object features could be extracted from image sequences using this algorithm.By comparing the object tracking algorithms, we propose a tracking algorithmbased on kalman prediction by setting up mathematical model. Tracking speed isimproved greatly because the matching process is taken place by the centroidposition prediction. Complicated moving situations could be solved effectively andeasily.We introduce hardware platform of development and Debugger based on DM642 inthis paper. And design method of function modules in image process system isillustrated .At last we realize the detection and tracking algorithms in hardwareand analyse the acquired data and parameters.