A Multiple Objects Contour Extraction Using Initial Edge Critical Inflection Edge Points Detection and Segmental Bi-directional Edges Searching and Pruning

• Main Authors: Wei-Chieh Lai, 賴維杰
• Other Authors: Roy Chaoming Hsu
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/68250832870361644880

碩士 === 國立嘉義大學 === 資訊工程學系研究所 === 99 === Contour extraction is one of the image processing technologies, in obtaining get the contour of the objects from an unknown image. It is also one of the most important pre-processing in image processing application. The first obtain the morphological representation of the object using skeletonization, and then detects the initial inflection points iteratively for each object. The difficulty curvatures problem of contour extraction is that if the shape of the object contains complicated, the accuracy of contour extracted could be affected. Therefore, this paper proposes a fast and accurate contour extraction method obtain multiple-objects from an unknown image. The proposed contour extraction method is consisted of two steps, Initial Edge Critical Inflection Edge Points Detection (IECIEPD) and Segmental Bi-directional Edges Search、Linking and Pruning(SBESLP). After the initial contour points are found, the origin and target points pruning and linking in two directions on each contour segment with the edge searching, and finally, the segmented contours are connected to obtain closed contour for each object in image.