【作者】 王毅； 郝重阳； 韩培友； 顾德明；

【Author】 Wang Yi Hao Chong-yang Han Pei-you Gu De-ming (Institute of Electronic & Information Engineering,Northwestern Polytechnical University,Xi’an 710072,China)

【机构】 西北工业大学电子与信息工程研究所；

【摘要】 采用多边形拟合物体图象边缘轮廓线,即实现从图象到图形的转换的方法,与直接记录原图象数据相比可大大减少轮廓线上所需记录的边缘点的个数,即减少数据量。但是这种采用图形形式的图象数据记录方式存在对“突刺”敏感和对所围面积无法控制的问题。本文提出的基于聚合的多边形拟合物体图象轮廓线的方法,与传统的拟合方法相比,既保留了原来的减少数据量的优点,又能解决传统方法对检测出来的边界“突刺”较为敏感的问题,还能有效地控制多边形各边与轮廓线围成面积的大小。更多还原

【Abstract】 Using polygon to fit the contour of the object’s image can implement the transform fromimage to graphics.Compared with recording the source image data directly,this method can largelydecrease the number of points at the contour,which must be recorded.It is,however,very sensitive tothe “burr” of the image.And the area encircled by the contour and the fitting polygon is hard to becontrolled.This paper presents an polygonal fitting method to image contour based on mergence.Compared with the traditional methods,it not only has the advantage of decreasing the record data,butalso can solve the problem of being too sensitive to the “bur” at the boundary,and control the areaencircled by the contour and the fitting polygon efficiently.更多还原