Moving Vehicle Detection Algorithm Based On FPGA

Moving Vehicle Detection System(MVDS) is an important part of the Intelligent Transportation System, and plays an important role in urban traffic surveillance system(UTSS), urban traffic management system(UTMS).ln recent years, MVDS has become one of important reseach subjects in intelligent transportation field that apply computer vision and video image processing technology.With the popularity of automobile, systems based on PC and card can’t meet the needs of the present. FPGA as today’s mainstream programmable logic devices, it has large-scale, high integration,and parallel processing characteristics, and can able to meet the requirements of large volumes of data and processing speed in video image processing.Therefore, the research of FPGA-based moving vehicle detection has very important meaning and value in the modern intelligent transport systems.Firstly, the paper describes the meaning and status quo of the research,and introduces the basics of video processing, digital systems modeling and simulation tools and design process.Then it analyzes the basic principles of optical flow method, background subtraction and frame difference what commonly used detection algorithm,and compares their advantages and disadvantages, selectes inter-frame difference that is simple, fast processing speed, easy hardware implementation as a detection algorithm, building the model and simulation of frame difference algorithm with Xilinx System Generator. In view of the shortcomings of inter-frame difference algorithm for the detection of moving vehicle, and its improvements, uses the improved three frames difference to detect moving vehicles and makes model and simulation demonstrate that it has a strong dynamic environment adaptability and anti-interference ability,whitch can achieve ideal detection.Finally, the overall structure and function modules of a moving vehicle detection system are described, and video image acquisition, image processing algorithm and VGA display outputing modules are designed in detail.In this paper, through the analysis and research of moving vehicle detection algorithms, it selected the fit detection algorithm based on three frames difference, combined with System Generator software as modeling and simulating tool, to achieve the processing of the moving vehicle detection algorithm in XUP Virtex-Ⅱ Pro development board. The design implemented by combining System Generator with FPGA, it has strong flexibility and short development cycle, and can improve real-time performance.