Research And Implementation Of Visual Target Tracking Algorithm Based On FPGA

Target tracking technology is one of the hot topics in the field of machine vision,which is widely used in various fields,such as military,aviation,medicine and so on.After years of research,target tracking technology is becoming more and more mature.FPGA（the field programmable gate array）has the advantage of parallel processing,which is very suitable for hardware acceleration for complex tracking algorithms.So it is of great academic significance and engineering practice value to apply FPGA to target tracking technology.In this paper,a variety of target detection algorithm and target tracking algorithm are analyzed.The frame difference method is used to detect the moving target,and the template matching algorithm is chosen to track the detected target,and the system is designed on the DE2₁15 development platform of Altera company.The main contents of this paper are as follows:（1）Firstly,we study the three mainstream target detection algorithms: frame difference method,background difference method and optical flow method,analyze their principle of target detection and the applicable environment of the algorithm,and decide to adopt the frame difference method for target detection.The morphological operation is used to filter out the holes and noise points generated by the frame difference method.（2）The template matching algorithm is used to track the detected moving target.The template matching algorithm is mainly divided into the template matching algorithm based on the gray-scale and the template matching algorithm based on the feature.In order to avoid the gray-scale template matching algorithm is more sensitive to noise,and the feature-based template matching algorithm to match the problem of slower,a two levels template matching algorithm based on edge features is proposed.The algorithm changes the traditional global matching into two levels matching and improves the matching speed.The first level rough matching uses the mean of the template image and the window image to find the rough matching points.The second level accurate matching uses the edge and edge gradient information to calculate the SAD function,and the exact matching points are found from the rough matching points.（3）The TRDB-D5 M camera is used as the video acquisition system,and the DE2-115 development board of Altera company is used as the hardware design platform.The detection and tracking algorithm is designed by Verilog hardware description language,and compiled on the Quartus II13.0 development platform.Finally,the system is simulated by the logic analyzer.The simulation results prove that the system can meet the design requirements.In order to verify the algorithm’s target tracking effect,the algorithm is simulated on the MATLAB platform.The experimental results show that the algorithm has better anti-noise ability and higher tracking accuracy compared with the traditional algorithm.