【摘要】 提出了一种基于灰色系统理论的边缘检测算法。该算法根据边缘点在像素邻域数列上表现出的与非边缘点的不同特征,分别选择典型非边缘像素的邻域数列和待检测图像所有像素的邻域数列作为参考数列和比较数列,利用斜率关联度计算两者之间的灰关联度,据此区分是否为边缘像素。实验表明,该算法对于椒盐噪声和随机噪声有较好的抗噪能力,提取边缘效果较好。更多还原

【Abstract】 This paper puts forward a novel algorithm for image edge detection based on gray correlation degree.First,this algorithm decides employing typical non-edge pixel neighboring sequence of gray level as referential sequence,simultaneously regards all pixel neighboring gray level sequence of disposed image as compared sequence,then computes gray correlation of referential sequence and compared sequence using slope correlation degree,finally judges whether a pixel is edge point based on the comparision of the computed gray correlation and the threshold value.To test this new algorithm the authors make experiments and the result shows this algorithm can gain more precise edge pixel than some traditional methods,futermore it can tolerate some noise as salt and stochastic.更多还原