The Design And Research Of The Video Moving Object Detection System Based On The FPGA

At present, the intelligent monitoring technique is applied to all walks of life, ismainly to monitor regional specific goals for the real-time detection and management.The video moving object detection algorithm is intelligent monitoring technology ofthe most important step in the system.In this paper, the edge detection algorithm withinterframe difference method as the core of video moving object detection algorithm,at the same time to improve the traditional classical Sobel operator, make detection toachieve better effect.In order to meet the requirement of high real-time imageprocessing, this article selects the has parallel processing ability and the processingspeed of the FPGA as the main control chip, to come in video image processing andextract moving targets, finally complete display on VGA display.First of all,This paper introduces the three mainstream video moving objectdetection algorithm, and expounds and analyzes their working principles andadvantages and disadvantages in detail,by comparing and analysising the results ofdifferent algorithm simulation for the same picture, select the interframe differencemethod with the more simple algorithm and better detection effect as the corealgorithm.At the same time in order to better target the movement the outline clearlydisplayed, this paper introduces the edge detection algorithm commonly used inimage processing area of extraction and segmentation. The thesis describes andanalyzes the principles and characteristics of four kinds of edge detection operatorand makes a software simulation for each operator. Finally select the sobel edgedetection with simple hardware implementation and the edge of the target locationprecise to operator to image edge.The second,this paper introduces the hardware design platform of thesystem,mainly divided into video image acquisition module, image cache module,algorithm processing module and video display module.At the same time the paperintroduces the hardware chip principle and function of each module,including DE2development board of Altera corporation,the MT9M011chip,the SDRAM chip andVGAvideo conversion chip.Then according to the overall design, the system is divided into modules, andthis paper describes the hardware realization of each module in detail.At the same time operation the waveform simulation on Modelsim for the video acquisitionmodule, format conversion module,the simulation results show the correctness of datatransmission in the process of Images from the collection to display.Then focus onanalysis the deficiency of traditional median filtering algorithm and classical Sobeledge detection algorithm and put forward the improvement,At the same time,simulatethe improved algorithm,do a comparative analysis with the traditionalmethod,concluded that the improved algorithm is better,and consume less hardwareresources.On this basis do frame difference to the processed image and then getmoving target,and display the processing results on a VGAmonitor.The experimental results show that the proposed in this paper based on theimproved Sobel edge detection of the interframe difference method can accuratelydetect moving objects in video, and the edge of the target location precise and clear,at the same time the algorithm compared with the traditional algorithm consumes lesshardware resources, fully meet the requirements of high real time capability.