Design And Implementation Of Target Tracking System Based On Fusion Detection Algorithm On FPGA

With the development of smart driving,intelligent transportation and smart cities,target tracking technology has attracted a lot of attention in academia and industry.Despite the growing maturity of target tracking technology,the speed,accuracy,and ability to handle large data streams are still challenging.In many areas,there is a high real-time requirement for target detection and tracking.Because FPGA have reconfigure high-speed data parallel processing capabilities,they are ideal for applications that require large amounts of data and require high real-time and volume requirements.The design adopts the method of detecting and tracking first to track the target.It can be said that the target detection is the cornerstone of the target tracking,and the quality of the detection directly affects the subsequent tracking accuracy and accuracy.A single three-frame difference method is used as the moving target detection algorithm,which will result in incomplete target edge information,which will affect the accuracy and accuracy of target tracking.The main research contents and innovations of this paper are as follows:(1)In order to obtain more complete edge information in order to obtain more complete edge information,this paper proposes a fusion target detection algorithm based on improved three-frame difference method and improved Canny.The fused algorithm not only overcomes the shortcomings of the three-frame difference method,but also combines the advantages of the Canny algorithm to make the edge information of the detected target more complete,accurate and continuous.Make a good foundation for follow-up target tracking.(2)Aiming at the problem of error detection caused by uneven illumination in the target detection process,this paper adopts the local adaptive segmentation algorithm instead of OSTU algorithm.As a threshold judgment of three-frame difference,compared with the OSTU algorithm,the algorithm can better overcome the influence of uneven illumination on edge segmentation.(3)Aiming at the occlusion and deformation problems in the target tracking process,this paper adopts an adaptive centroid tracking algorithm,which can adaptive track the change of the target size during the target tracking process,and can improve the occlusion and deformation problems.Moreover,this algorithm has strong anti-interference ability and good robustness.(4)Design and implementation of the above algorithms on FPGA.By comparing and analyzing the detection results through video frame sequences,the detection algorithm in this paper can obtain more complete edge information of the moving target,which can greatly improve the detection accuracy and accuracy.Testing on target tracking under conditions of occlusion,deformation,and lighting changes can achieve real-time and accurate target tracking results.