Moving Target Detection In Complex Background

Intelligent visual monitoring is a new field of computer vision,it integrates computer intelligent application,pattern recognition,artificial intelligence,image processing applications,math and other key technologies.It wide range is applied to all kinds of robot and intelligent vehicle navigation,medical image analysis,intelligent transportation,various types of video monitoring and other fields.Because various background environment,different lighting conditions,complex shape,and motion pattern of the complexity of the barrier between,and the detecting and tracking multiple purpose keep is one of the tasks in the field of computer vision,challenging.Road traffic management in order to obtain the vehicle motion data,often used for early hardware measuring induction coil.If you use the camera,and then use software to deal with all kinds of resources,can improve the processing of convenience and flexibility.In this paper,the moving target detection and tracking research how to make use of computer data object motion video sequences is the key of the research.Moving object detection and tracking can be divided into background extraction,mobile group of extraction,extraction and tracking moving objects move the point group.In this article,comparison of various algorithms for each step analysis:(1)In the study of moving target detection algorithm,the introduction of several commonly used moving object detection method,and the validity of each algorithm are analyzed in the experiment.(2)In the research of moving object segmentation algorithm,author describes the basic knowledge of the target segmentation involved and key technologies,including image binarization,morphology,connectivity analysis.Target detection method using image binarization methods to improve the mathematical morphology and corrosion connected domain,and the realization of the scanning algorithm and extend,reduce computational complexity,and made the algorithm's efficiency.(3)In the complex urban traffic environment composed of trees,lamp poles and viaducts,a variety of multi target mobile detection,segmentation,extraction and trackingprocessing are studied under the background of the environment.By comparison and analysis,the target tracking algorithm based on KLT feature prediction is adopted to achieve automatic multi target tracking and detection.The results show that the algorithm has better results.