Research On Algorithm Of Infrared Target Tracking Based On Compressed Sensing

With progress of social modernization,increase of urban population and motor vehicle ownership rate,the traffic is getting worse.The highly saturated car ownership rate makes environmental pollution worse,and the exceeded carbon emission causes more and more fog and haze days.In the face of these urgent problems,the idea of intelligent traffic emerged.Traffic intelligence can use the existing transport facilities effectively,to better guide traffic stream,and reduce traffic load through intelligent tools.It can also reduce traffic accident rate,track illegal vehicles,handle traffic accidents quickly,mitigate environmental pollution,and improve transport efficiency,to ensure traffic safety.This thesis studied the compressed-sensing-based video tracking and detecting method,and proposed a more effective idea.The detailed work is as follows:1)It described the advantages and disadvantages of infrared imaging and general imaging,and described the reasons for selecting the infrared video tracking.2)Due to the importance of the algorithm in today’s intelligent traffic,artificial intelligence,as well as the massive traffic data to be processed,the study focused on the compression sensing algorithm.By combining the random Haar-Like feature extraction with compressed sensing theory and Naive Bayesian classifier,it carried out target tracking by infrared video.In this method,it improved the feature extraction part.3)It conducted simulation of the algorithm on the four groups of infrared videos under different experimental backgrounds;with oscillogram and table,it showed the tracking effect intuitively.The experimental results showed that the compression sensing algorithm can be used to track and detect the infrared target,and the improvement measures are effective.4)There are many algorithms for video tracking,and we compared the KCF algorithm with our method.The simulation results showed that our method has better target tracking accuracy.This paper conducted experiments on the proposed algorithm in the Matlab and VC ++development environment.This compressed-sensing-algorithm-based video tracking and detecting method designed by us has advantages of simple operation,low cost,and good costperformance.In addition,as infrared video provides wider monitoring scope,plus good robustness in experimental results and the effective improvement measures,the tracking accuracy is higher than that of KCF algorithm.