Identifying moving objects from a video sequence is a fundamental and critical task in many computer-vision applications. A common approach is to perform background subtraction, which identifies moving objects from the portion of a video frame that differs significantly from a background model. Most methods for foreground region detection in videos are challenged by the presence of quasi-stationary backgrounds flickering monitors, waving tree branches, moving water surfaces or rain. Additional difficulties are caused by camera shake or by the presence of moving objects in every image. In this paper, we proposed a background subtraction algorithm based on background reconstruction method for both stationary and dynamic background video sequences. Firstly, pre-processing is used to obtained the appropriate video frames from video sequences. Secondly, background is reconstructed by averaging and filtering method . Finally, the initial video object is derived in each frame by subtracting the background from this image, after that, mathematic morphology post-processing is used to get an accurate video object. Experiments on typical sequences have successfully demonstrated the validity of the proposed algorithm.
Index Terms— Video segmentation, Background reconstruction, Background subtraction, Moving object.
Video segmentation refers to the identification of region in a frame of video that are homogeneous tin some sense. Different features and homogeneity criteria generally lead to different segmentation of same data; for example, color segmentation, texture segmentation, and motion segmentation usually result in subdivision maps. Furthermore, there is no assurance that any of the resulting segmentation will semantically meaningful, since semantically meaningful region may have multiple colors, multiple textures, or multiple motions. Generally motion segmentation is closely related to two other problems, motion (change) detection and motion estimation. Change detection is a special case of motion segmentation with two regions, namely changed and unchanged region (in case of static cameras) or global and local motion region (in the case of moving cameras)
Tracking moving objects from a video sequence is a fundamental and critical task in video traffic surveillance, human detection and tracking, and gesture recognition in human-machine interface. A common approach to extracting the moving objects is background subtraction method. With a simple theory and an uncomplicated design, the background subtraction method can quickly detect the moving objects, getting a relatively accurate objects information, especially suitable for video surveillance with stationary cameras. However, when the background model is reconstruct, there will be problems like background disturbance, changing of external light condition and moving objects long stayed in the background.