Research Of Movingtarget Videotracking System Based On S3C2440

With the continuous improvement of productivity, the increasing ofmanufacturing industry competition and cost pressures, the video trackingapparatus for personnel monitoring in industrial production and management,will be able to improve production efficiency, reduce labor costs. So designed asimple and reliable, and low cost tracking system, to promotion of video trackingapplication in industrial production, especially the application of production linesand product quality testing field is very meaningful.So this paper puts forward a video moving target tracking system that basedon LVQ neural network algorithm, and it adopt the embedded chip ARM9as thehardware platform and the Windows CE6.0operating system as the softwareplatform.Through pretreated the detected image to obtain the input eigenvalue ofLVQ neural network, and then makes the LVQ neural network obtain the abilityto judge the target after the network learning the study sample. Thus the systemcan be used for product quality testing and other links in industrial production torealize the automation of production management.Firstly, the research status and development trend of video tracking system isintroduced in this paper. And through analysis the functions that the videotracking system can realize to ultimately determine the design of based on theLVQ neural network algorithm.In order to develop and upgrade system moreconvenient in the future, Windows CE6.0operating system is chose as thesoftware developing platform and the LVQ neural network algorithm is used torecognize and track the target.The system hardware is consisted by main controlmodule, video signal acquisition module, data transmission module, externalmemory module and power supply module.The system porting and driver loadingis explained in detail in this paper, and the design of system software is designedin the embedded software platform. Through the simulation results show that, the video tracking system basedon the LVQ neural network algorithm can be realized to recognize and track theobject at a certain speed, and it is suitable as the video surveillance in theindustrial production line and production manipulator.