Realization Of Tunnel Traffic Detection System Based On FPGA

Intelligent transportation construction is the one of the key projects in the 13 th Five-Year Plan of China.In the construction of intelligent transportation,the intelligent management system of tunnel traffic is one of the most important research items.In order to realize the intelligence of the tunnel traffic management system,the real-time and accurate information about the tunnel traffic must be available.As FPGA has many advantages like of a large number of configurable logic resources,parallel processing and pipeline operation,which can realize the real-time processing of video images,it is of great practical value and significance to study the tunnel traffic detection and tracking system based on FPGA.The main purpose of this thesis is to realize the function of tunnel traffic detection and tracking based on The XC6SLX45 TM development board of Xilinx Spartan6.Our main work are as follows:(1)The common algorithm and domestic and international development status of Target Detection and Tracking Technology is introduced,compares the detection effects and speed of optical flow method,inter-frame difference method,GMM background difference method and ViBe algorithm based on a simulated.The function of tunnel traffic detection and tracking is realized in MATLAB based on GMM background difference method,Kalman filter and Hungarian matching algorithm.(2)This thesis analyzes the design method of tunnel traffic detection system based on FGPA,and adopts the top-down design method,video image acquisition module,DDR3 reading and writing control module,target detection and tracking module and VGA display module.Then the thesis completion the video image acquisition module,DDR3 reading and writing control module and VGA display module are introduced through Verilog hardware description language,a real-time processing hardware platform for tunnel traffic detection and tracking based on FPGA is built.(3)The target detection and tracking module is implemented based on inter-framedifference method and bounding box algorithm.The thesis completed the inter-frame difference module,median filter module,median filter module,the morphology processing module,the Sobel edge detection module and the bounding box module on the FPGA,Then the tunnel driving detection and tracking system based on the FPGA is implemented.(4)Test the tunnel driving detection and tracking system on FPGA.The test results shows that the detection and tracking system based on FPGA can realize the function of tunnel traffic detection and tracking,and its frame frequency is 30 fps,meeting the real-time requirements.But multi-objective detection can not be realized.