模糊推理在边缘检测中的应用

针对传统边缘检测算法边缘轮廓模糊、易产生伪边缘等缺点,提出一种基于模糊逻辑推理策略的边缘检测新方法,在无需确定阈值的情况下对数字图像进行边缘提取。给出一种以变化的3×3模板为基准的模糊推理规则,通过对周围像素点的垂直和水平方向以及连续四点像素的黑白分布的判断,提取出图像中的黑点、白点和边缘部分。实验结果表明:与Robert算法相比,该方法对图像中噪声和双重边缘的抑制以及对曲线的圆度轮廓的提取都有很好的效果。

Application of fuzzy inference in image edge extraction

According to the problems that the traditional algorithm of edge detection exists, a new edge detection method based on the fuzzy logic is presented in this paper, and the edge of image could be extracted without the threshold by using this method. A mutative 3＊3 template is adopted for the fuzzy inference rules, in order to extract the black point, white point and edge of the image, it is necessary to determine the distribution of black and white pixels by detecting the vertical direction, horizontal direction and the four continuous points of the surrounding pixels. Experiments show that this method has a great performance in inhibiting the noise and double edge, forever when extracting the contour of curve, it also works.