| Moving object detection is one hotspot of computer vision research field. As the saying goes "where plum blossom falls, the wind blow mountain full one night". Since the ancient times, people tend to focus on moving object. Moving object detection is broadly applied in military command, intelligent monitoring, pattern recognition, etc. Moving object detection is closely related to computer vision field. How to design a precise, real-time and portable moving object detection system has clearly become the important research topic.This paper has a deep research on the algorithms of moving object detection at home and abroad, finding the existing problems. We propose one algorithm about moving object detection, based on CSLBP and frames difference. This algorithm has good robustness to external environment and can accurately detect moving object.On this basis, this paper designs and realizes a moving object detection system with module design thoughts, facilitating system development and later maintenance. This whole system is divided into video capture module, preprocessing module, moving object detection module, post-processing module and video output module.Now, most of moving object detection algorithms are realized on PC. So it is slower and has small suitable scope. We realize the moving object detection algorithm on FPGA platform. FPGA has natural parallelism and can satisfies the requirement of high-speed image processing. We chose xilinx'Virtex-â…¡Pro as development platform, integrated hardcore PowerPC405 processor. We design the moving object detection system,using hardware and software collaborative design idea and SOPC technique. So this system is flexible and convenience for later hardware and software upgrade. |
