The Research On Object Segmentation Algorithm Based On Edge Detection And Watershed

Along with the development of information technology, information technology application in medical care more and more widely. As the most direct visualization video information in medical care plays an important role, it can reflect the behavior and mental state of patients. But the large amount of video data presents a large difficult problem to video data transmission and storage. So it is significant to study video object processing.Video object segmentation which base on object is a part of video processing, it is a hotspot and difficulty in video study at present. The object of video object segmentation is to extract the motion objects from video image sequences .It plays a key role in following image analysis, storage and retrieval. Due to the image complexity and diversity of environmental factors, but now it still lacks a universal image segmentation algorithm for all video images. It creates a huge obstacle to the promotion of application of video image processing. So this article focuses on the research for a universal image segmentation algorithm and study on the edge information and region information. An improved video motion segmentation algorithm based on edge detection region merging based on video object segmentation algorithm is proposed. The main contents include the following aspects:Firstly, in view of the difficulty of the large calculation of the current image segmentation algorithm, an improved video motion segmentation algorithm based on varying edge detection is improved. The algorithm selects the appropriate image difference algorithm to get difference image based on the image change rate and keep more complete edge information of video object, then extracts the initial edge of video object, and finally extracts video object by the edge connected operator and post-processing techniques.Secondly, in view of the problem of shadow and incoherent motion imaging, an improved video object segmentation algorithm based on the edge detection and watershed is proposed. This algorithm extracts motion objects based on fusion edge detection operator of difference image and gets the contour of video object using the watershed region merging algorithm. More over, the algorithm extracts video object through combining the spatial and temporal image sequence information segmentation and effectively improves the problems of shadow and incomplete edge.In conclusion, this article brings the varying edge detection operator based on the difference image to the watershed transform and presents a video object segmentation algorithm based on the edge detection and watershed. It takes standard sequence image for research object, many experiment results validate that this method could effectively reduce the noise, improve the motion inconsistent, video object's shadow problem and the accuracy of video object segmentation.