摘要
边缘检测是图像分割中的一个重要部分 .本文依据 B样条修匀公式的图像边缘检测的基本思想 ,但又因为用 B样条函数光滑后的图像对原始图像的逼近性并不十分精确 ,即它们的误差图像中残留着一些可能造成检测到的边缘线不连续或有用特征可能检测不到的边缘信息 ,为减少残量误差 ,本文对图像原型值采取盈亏修正 ,然后再对修正的图像使用 B样条修匀公式做全息光滑曲面 ,重新得到的曲面改进了逼近性 。
The basic concept of an edge detector is an important method in image areas. Application of the edge detector of a B spline smoothing surface in this paper,the smoothed image can not exactly approximate the original image in other words, it remains some edge information in their error image and the remained information maybe induces that the detected edge line can not join together and valuable fatures can not be detected. In order to solve this problem, this article proposes that the value of the original image is amended through adding and minusing. A smoothing surface is presented by using the uniform B spline function in the amended image. The new camber is improved approximation and still has smoothness and convexity by putting the improved scheme in order.
出处
《哈尔滨师范大学自然科学学报》
CAS
2003年第2期72-75,共4页
Natural Science Journal of Harbin Normal University
