Research And Implementation Of Real-time Moving Targets Detection And Tracking Based On FPGA

In recent years,intelligent video surveillance technology has been applied in more and more areas,and tends to be high-definition and real-time.As a key part of the intelligent video surveillance system,real-time detection and tracking of moving targets has gradually been the attention of many scientific research workers.Due to the limitations of software-based methods in terms of processing speed and algorithm efficiency,real-time detection and tracking of moving objects based on hardware has became a hot topic in current research.FPGA(Field Programmable Gate Array)chips have a lot of logic resources,parallel processing,programmable character and so on,making it very suitable as a real-time video image processing platform.The main task of this paper is to use the Lattice's ECP3 FPGA chip as the core processor to complete the research and design of real-time objects detection and tracking system based on parallel architecture and processing architecture in the pipeline.The main research work of this paper is as follows:(1)This paper analyzes the design method of video image processing system based on FPGA,and adopts the top-down design method and processing structure in the pipeline.Then the paper realizes the acquisition,preprocessing and display of high-definition video image through Verilog hardware description language programming to build a real-time processing of high-definition video images hardware platform based FPGA for the research of moving targets detection and tracking algorithm in FPGA.(2)This paper analyzes the common algorithms in moving objects detection,regional mark and target tracking,and compares them in the time-consuming,resources consumption and algorithm effect according to their basic characteristic.And the paper realizes the detection and tracking of moving objects in real-time high-definition video streams by an FPGA implementation scheme based on background difference method,connected domain tagging algorithm and histogram matching algorithm.(3)This paper analyzes the inherent timing relation of real-time video stream and designs the parallel architecture of the moving objects detection and tracking algorithm module based on the temporal parallelism of real-time video data.Thus,the processing speed of the algorithm module can be improved by using the large-scale parallel processing mechanism of FPGA.The hardware implementation circuit of the moving targets detection and tracking algorithm module is designed in detail on FPGA by using Verilog hardware description language to realize the real-time detection and tracking of the moving targets.(4)This paper verifies the design of real-time moving targets detection and tracking system based FPGA.Through the test results,the system can complete the detection and tracking of multiple moving targets in the 1080 p video in 30 fps rate.