The Design Of A Motion Tracking System Based On FPGA

Video moving target tracking in computer vision is one of the challenging issues in the rich, has many potential applications in the field of military guidance, navigation, security surveillance, intelligent transportation, video coding, medical diagnostics and meteorological analysis, rather processing speed is the main factor affecting the structure of a video processing system, with more high speed of operation and faster, more and more low prices, FPGA chip is becoming the core video processing system integration of FPGA chip devices.This paper designs and implements a moving target tracking system based on FPGA. Papers on the subject in the background and current status of the development of domestic and international moving target tracking system summary, based on the analysis of the two core algorithms Target Tracking:moving target detection algorithm and moving target tracking algorithm. In moving target detection algorithm based mainly discussed the value of the time series and pre-filtering motion detection algorithm proposed by Antoine Manzanera and other post-processing problems; in moving object tracking algorithms mainly discuss various feature point extraction algorithm and based on the minimum absolute value and the SAD matching tracking algorithm.Depending on system requirements, a detailed description of the overall structure of the paper FPGA chip, the video image acquisition module within the FPGA chip, SDRAM controller module, FIFO data buffer module, VGA video output module control, PLL clock PLL circuit design module.On the basis of the on-chip hardware circuit, the paper details the specific design and implementation of moving target detection and motion tracking algorithm. During the moving target detection algorithm design, in order to avoid errors in the video noise and jitter caused by light, using a pretreatment smoothing filter, motion analysis, binarization, expansion, corrosion and other post-processing; moving target tracking algorithm in achieving When using an algorithm based on feature points matching SAD track in achieving algorithm process, ensuring high throughput on the basis of the system, using a path allocation of resources based on extended compatibility and integration (ECPB) hardware integration algorithm, the resource consumption of the system is improved.The author uses Modesim, Quartusii, SignalTapii and other EDA tools on the DE2-115evaluation board for the circuit design of each module of the code verification, timing simulation and synthesis verification. The simulation and test results show that, under different backgrounds, different light and a different rate of movement, this article is designed to track the accuracy of the moving object tracking system reaches the technical requirements of the system.