Moving Object Detection Based On Structured Low-Rank And Sparse Decomposition Model

Moving object detection,as a branch in the field of image processing,is the first step of a series of middle-level or high-level visual processing such as target tracking,target classification,behavior understanding,semantic description and so on.Therefore,moving object detection has become a basic and indispensable subject in computer vision research with great theoretical and practical significance.Until now,extensive effective methods have been put forward for detecting moving objects.However,in practical application,due to various challenges in the scene,it still faces lots of challenges such as sensitivity to noise and occlusion,false detection and missed detection in complex scene,collapsing with large or tiny objects and so on.Aiming of relieving above issues,the main work and contributions of this thesis include the following two points:(1)Aiming at the problems such as low detection efficiency and low robustness of the current moving object detection methods with relatively large object sizes,we propose a novel approach,called Collaborative Low-Rank And Sparse Separation(CLASS),for moving object detection.Given the data matrix that accumulates sequential frames from the input video,CLASS detects the moving objects as sparse outliers against the low-rank structure background while pursuing global appearance consistency for both foreground and background.The sparse and the global appearance consistent constraints are complementary but simultaneously competing,and thus CLASS can detect the moving objects with different sizes effectively.The smoothness constraints of object motion are also introduced in CLASS for further improving the robustness to noises.Moreover,we utilize the edge-preserving filtering method to substantially speed up CLASS without much losing its accuracy.The extensive experiments on both public and newly created video sequences suggest that CLASS achieves superior performance and comparable efficiency against other state-of-the-art approaches.(2)In the fact that the existing low-rank and sparse separation based moving object detection methods tend to produce false positive detection and incomplete boundary of the objects in complex scenes,we propose to enforce the spatial compactness and appearance consistency in the low-rank and sparse separation framework in this thesis.Given the data matrix that accumulates sequential frames from the input video,our model first detects the moving objects as sparse outliers against the low-rank structure background.Furthermore,we explore the spatial compactness by enforcing the consistency among the pixels within the same superpixel.This strategy can simultaneously promote the appearance consistency since the superpixel is defined as the pixels with homogenous appearance nearby the neighborhood.The extensive experiments on public GTFD dataset suggest that,our model can better preserve the boundary information of the objects and achieves superior performance against other state-of-the-arts.