Based On EmguCV Of Moving Object Dete Ction Methed In Video Sequence Images Research

With the development of computer vision and digital image processing technology at home and abroad,has a large number of scholars of target detection algorithm research and demonstration analysis,however,in the field of computer vision research also has many questions waiting for solving.Video sequence images of moving objects detection method research target tracking and recognition is one of the basis of many key research and application.This paper first the basic theory and method of moving target detection were described briefly.Then using actual photography collection of video sequence and video sequence of motion target detection public library,has carried on the experiment;Compares the test results analysis,the advantages and disadvantages of several common algorithms of moving targets,the improved algorithm is put forward finally.Based on improved hybrid gaussian model and frame difference of moving target detection algorithm,is research in recent years is rich in target detection algorithm.This article is based on Emgu CV open source image processing library,the design of an improved video sequence image motion target detection algorithm,the purpose is to improve the frame difference algorithm for complex gaussian mixture model and noise background environment change,target detection is incomplete and serious problems such as to make improvement;Finally three frame difference algorithm,this paper proposes a into improved gaussian mixture modeling,and combined with K-means clustering algorithm to get the testing results of the calculation,remove the shadow parts such as noise in the test results;Among them,according to the improved gaussian mixture algorithm update strategy,the inhibition of the noise effectively.Using VS2013 and Emgu CV2.9 build test environment and platform,design the experimental process,and then analyzes the improved algorithm research,the experimental results show that the improved algorithm,effectively improve the three frame difference moving targets detection results of incomplete fault,mixed gaussian algorithm to detect relative noise larger faults,the improved algorithm is relatively good adapt to illumination changes,environmental changes,such as background mutation problem,solve the less sensitive to changes in the background as well as the characteristics of good real-time and anti-jamming capability,test results are also in a variety of data sets higher robustness and accuracy,and to some extent,can obtain better results,but can be by simple closure operation,with a rectangular box,displaying video image sequences detected moving target in the area.At the same time,the use of c # in the experiment of unsafe code technology,can effectively improve the efficiency of the moving target detection algorithm.