Research On Target Detection Algorithm And Tracking Algorithm In Machine Vision

As artificial intelligence gradually enters People's Daily life,it promotes the rapid development of machine vision technology.Among them,moving target detection and tracking has always been a research hotspot in the field of computer vision.In practical engineering,image data transmission loss,noise interference,moving target deformation,occlusion,scale change and background mutation,light intensity change and other problems often exist.How to detect and track moving objects quickly and accurately and have stronger robustness under complex scenes has been a research hotspot in the field of moving target detection and tracking.In view of the above problems,this paper studied the research status of target detection and tracking,and studied the algorithms related to target detection and tracking from three aspects.First of all,in the aspect of image denoising,several commonly used image denoising methods are analyzed and studied in this paper.In order to improve the denoising ability of median filtering and the problem of fixed filtering window,an adaptive fuzzy median filtering algorithm is proposed.The simulation results show that the proposed algorithm has better denoising ability for different noise densities,and can better preserve the image edge details.Secondly,in terms of target detection,this paper in the process of moving objects were detected ViBe prone to "ghost" phenomenon is seen with prospects mistakenly identified problems,put forward the ViBe of radius of an adaptive threshold algorithm,abandoned the single frame background sampling modeling process modeling method,using multiple frame averaging method for background modeling,morphology processing methods of target detection results optimization,through simulation experiment in this paper,the algorithm has better robustness under complex background,the "ghost" phenomenon in eliminating target and raised the prospect of detection accuracy.At the end of the paper,according to nuclear related filtering algorithm(KCF)target dimension and tracking precision,position prediction combinedABSTRACTwith Kalman filtering features,on the basis of HOG features combined with color features,at the same time to join the target sample size set,and through the simulation experiment in this paper,the algorithm not only can solve the problem of target dimension,at the same time improve the tracking accuracy and robustness of the algorithm.