Research On Moving Object Detection And Tracking Algorithms Based On Video Sequences

Detection and tracking of moving object is a frontier topic to be subjected to closeattention in the domain of digital image processing and computer vision.It has an importantapplication value.However,due to the complexity and diversity of the scene,there are stillmany problems to be solved.In this paper,the elementary knowledge of digital image processing firstly are discussed,including histogram equalization,median filter,Gaussian filter,threshold segmentation,edgesegmentation and watershed segmentation algorithm.Through these treatment,the image qualitycan be effectively improved,and provide a guarantee for follow-up work.On the research of moving object detection three main methods are analyzed,includingFrame Subtraction,Background Subtraction and Optical Flow.According to the two framessubtraction appears“double”and“empty”, We used a three frames subtraction to improve.Background Subtraction discusses gaussian mixture model and non-parameter model.Combining Frame Subtraction and Background Subtraction was proposed based on thecharacteristics of the two methods at last.Experiment shows that the algorithm can obtainmore complete target area.On the research of moving object tracking we mainly analyze the Kalman filter theoryand the Mean Shift algorithm and the Camshift algorithm,the latter is improved Mean Shift.Camshift algorithm characterized by color histogram,can effectively solve the target distortionand occlusion problem and Kalman algorithm can predict the location of the target in the nextframe of video sequences. Using the comprehensive advantages of the two methods,wepropose the Camshift combined with Kalman filter algorithm for object tracking. This methodensures the right track even when the shelter and similar objects interfere．...