Research And Implementation Of Motion Target Detection And Recognition Technology Under Dynamic Background

Mobile target detection and recognition is an important topic in the field of Computer Vision.It has wide application prospects in medical diagnosis,transportation,military safety and aerospace.But in actual production and life,most of the moving targets are in the dynamic background.Therefore,it is of great theoretical and practical significance to study the detection and recognition technology of moving target in the dynamic background.At present,there are some difficulties in the research of moving target detection and recognition technology under dynamic background,such as irregular background change and smaller size of target to be detected.These results will result in long time consumption and low accuracy of target detection.This paper analyzes the reasons for the long time of mobile target detection and recognition technology.It proposes a target set algorithm and an improved three frame difference algorithm,and designs and implements a mobile target detection and recognition system in dynamic background.The main work of this article is:(1)Under the dynamic background,the detection speed of the existing target detection and recognition algorithm is slow,and the algorithm of screening possible target sets is proposed.By filtering the target set algorithm,the background contour can be filtered out from the picture to reduce the area to be processed in the picture so as to speed up the system’s running speed.(2)Analyze the original three frame difference method,combine the basic knowledge of logic operation,and make full use of the information between adjacent frames,and propose an improved frame difference method of three frames.The improved algorithm can improve the accuracy of target detection.(3)Taking the flying unmanned aerial vehicle(UAV)in the dynamic background as the research background,the mobile target detection and recognition system is designed and realized.And by optimizing the system code,the running time is reduced.According to the experimental results of this paper,the system detection accuracy is increased from 78%to 81%,and the speed of the system is raised about 20 times after adding the improved algorithm.