Application Research of GPU in Vehicle Detection and Tracking System
|Keywords||GUP CUDA VDTS corner detection|
Intelligent Transport System is an accurate, real-time and efficient comprehensive management system of traffic and transport which puts advanced electronic sensing technology, computer processing, data communications transmission technology, electronic control technology and information technology effectively in application of the entire traffic management. Vehicle detection and tracking system is the core of intelligent transport system which based on machine vision technology. In recent years, the system has gradually become a hot topic among experts at home and abroad.With the increasing demand for practical applications,VDTS calls a higher need for real-time and accuracy. Since more and more data need to be handled, larger and larger accounting amounts are involved in the system, personal computer can not meet the needs of practical application any more. Therefore, how to improve the calculating efficiency of VDTS furtherly to meet the requirement of practical applications is quite a rush problem which need to be addressed in this filed.Currently, thanks to the great development of GPU, wen can get a quantatively improvement of speed. From this point, powerful computing and parallel processing abilities of GPU could break the bottleneck of speed-handle in VDTS, and upgrade the execute efficiency of algorithm.This essay firstly give a brief introduction about development process、working principle、features and trend of GPU general computing; then we discuss the CUDA platform which can exert GPU powerful calculating and parallel handleing ability by using a concrete example, including overview of CUDA, the requirement environment of soft ware and hardware, programming model, characters, applied filed and developed trend; lastly, we analyze the whole process of GPU application in VDTS of video-views, and make the image grey scale,image smoothing and corner detection come true.