自适应双线性广义模糊增强的图像轮廓提取

Contour Extraction of Image Using Adaptive Bilinear Generalized Fuzzy Enhancement

在模糊集合论(FS)和广义模糊集合论(GFS)的基础上,构造出用于模糊增强图像区域对比度的新型线性广义模糊算子(LGFO),理论分析并验证了其性质。提出了一种自适应双线性广义模糊增强算法用于提取图像边缘轮廓,该算法利用线性广义隶属度变换函数与LGFO,实现了灰度图像空间与广义模糊空间的相互转换,空间转换过程中对线性广义模糊隶属空间实施了线性广义模糊增强处理,最终在灰度图像空间中使用"MIN"算子提取图像轮廓。该算法还使用可检测边缘度与噪声标准差之商作为图像增强的评估标准,自动选择模糊参数D,实现了线性广义模糊增强图像的自适应优化。实验表明,该算法快速无失真,适用于彩色图像,提取的图像轮廓准确、层次分明。

A new linear generalized fuzzy operator （LGFO） used for fuzzy enhancement of regional contrast is constructed based on the theory of traditional fuzzy set （FS） and generalized fuzzy set （GFS）, and theoretical analyses and testing the characters LGFO. And then an algorithm for adaptive bilinear generalized fuzzy enhancement used for extracting edge contour of images is put forward. The algorithm is to achieve space transform between gray image space and generalized fuzzy space by using linear generalized membership transform function and LGFO. Using the algorithm, the linear generalized fuzzy membership space is enhanced fuzzily by LGFO in the process of space transform. Finally, the contour of image is extracted using ＂MIN＂ operator in gray image space. The algorithm is also to achieve adaptive optimization of the image after linear generalized fuzzy enhancement by selecting fuzzy parameter D automatically, using the quotient between detectable edge degree and noise standard deviation as the evaluation standard of image enhancement. The experiments show that the algorithm is efficient and has no distortion, and can be applied to color image. By using the algorithm, the contour of image is accurate and structured.