Intelligent Vehicle Tracking System Based On DSP Platform

Intelligent vehicle tracking system based on DSP which combines many fields of technology including digital image processing, artificial intelligence, automatic control etc. has broad application prospects. With the rapid economic and social development and people's urgent need for traffic safety, people have become increasingly demanding for safety and reliability of transportation vehicles.. Since the realistic traffic environment is often complicated, traditional passive technology can't solve safety problem fundamentally, thus the active safety technology which based on computer vision and pattern recognition is much considering. The intelligent detection and tracking technology is one of the research hotspot.The function of the designed intelligent vehicle tracking system is to achieve a car(the smart car) to follow another car(the guide car) initiatively without human intervention. The whole system includes navigation system, self-guided vehicles and the smart car. The front car is a guide vehicle, on which is equipped with a CCD camera, the DSP video processing board basing on DM642, serial communication and data conversion board, the wireless communication module and power supply system. The main task of the navigation system is to realize target detection and tracking as well as control the smart car, which is based on the DSP video processing board. The own car is made of car models, motor, servo, control board, motor drive, speed detection module, power supply systems and other modules. It receives control instructions from navigation system, and imitates the action of guided vehicle. Detection algorithm of the navigation system is based on the YUV space of a single Gaussian background modeling algorithm, and tracking algorithm is CAMShift algorithm based on the HOG feature. Considering the carrying capacity and drive performance of the own car, the navigation is installed on guided vehicle which is mainly for bearing the navigation system.The paper mainly includes three parts: (1) the detecting and tracking algorithm have a detailed analysis and description, and improved the detection and tracking algorithm simulation; (2) the two improved algorithm realized on the DSP, and from DSP software and hardware design development angle described DSP program development process; (3) the design of the intelligent vehicle control system hardware and software.This topic mainly completed the following operate: (1) Discusses the movement based on machine vision common target detection and tracking algorithm basic principle. According to the characteristics of video processing DSP platform of classic algorithm, made improvements, and proved the feasibility of the algorithm.(2) YUV color space based on single Gaussian background modeling moving targets detection algorithm based on gray texture characteristics and CAMShift algorithm ,which all are been realized on the PC platforms.(3) Put the two algorithm transplanted into a based on DM642 video processing plate, from the Angle of engineering application to optimized algorithm computational cost, adopt various methods to reduce computational complexity and optimization algorithm to save precious computational time raise the overall system of real-time. Finally give the design test report.(4) Completed independent car control system design in hardware and software systems, and realized between two vehicles the wireless communication, has constructed a navigation system for independent car wireless closed-loop control system of intelligent vehicle (5) Complete the system of joint debugging between subsystems, and verified system function under laboratory environment. Experimental results show that, in low-speed cases vehicle tracking system is accurate and valid.