Study On The Target Auto-detection System Based On ARM-Linux Machine Vision

Firing practice is the basical unit of the military training. It is important to improve the training condition and the training quality in the process of the ARM construction. First, the paper discusses the development process of the Automatic Target-scoring System, and based on the present resaerch situation of the domestic and foreigen of the Automatic Target-scoring System, this paper designs an Intelligent Target-Scoring System with the embeded machine vision.This paper discusses the embeded target-scoring system based on machine vision, and correlative picture format. It analyses the embeded Linux operating system, Qt/E and how to transport the ARM9 processor using the Open Computer Vision Base(OpenCV). It respectively researches the picture correction,the picture grayscale and the image binaryzation,how to trim and divide up the picture,and the correlative algorithm recognizing the target. Based on these knowledges, the paper puts forward the judgement target theory by color mould. In the research process, the paper utilizes the implementation scheme based on the function module design notion, and uses the ARM embeded image recognization module as the core; meanwhile it also achieves the procedure from the basement operation system and transportation of the correlative softwares to the transmission of the picture collection,picture processing and recognizing the target. The paper also resolves the problem how to utilize the ARM processor with the USB camera to automatically recognize the picture and record the target's sorces. The paper instructs the Target-Socring System's hardwares and its designing process, and introduces the used softwares including the QT based on the Linux operating system and the OpenCV and how to use these softwares to achieve the projects, then enlarges the experiment device utilized in this paper and the detailed experimental results.