AFM显微图像多尺度形态边缘检测算法

A MULTI-SCALE MORPHOLOGICAL ALGORITHM FOR AFM MICROGRAPH EDGE DETECTION

提出一种适应于显微图像的多尺度形态边缘检测算法。用不同尺度大小的结构元素分别检测出显微图像不同尺寸边缘信息,进行图像规格化处理以消除不同尺度边缘图像的灰度值分布差异,运用取极小值方法对所获处理后图像边缘进行自适应融合。实验结果表明运用该算法获得理想图像边缘的同时,增强了模糊边缘,有效地消除了噪声,对显微图像处理十分有效,检测到的边缘非常清晰。

In this paper we present a new morphological edge detection method using a multi-scale approach. The method fits micrograph well. Information of micrograph edges in different fineness are detected using structure element operator at different scale, and then they are regulated in order to eliminate the difference of gray levels distribution among multi-scale edges, and then combined to produce all the regulated edges of interest in an image the adaptive fusion method of minimums is used. The experiment results show that ideal image edges are gained with the use of the method while the fuzziness of image edge detection is enhanced, noise is removed efficiently, and distinct edges can be detected efficiently for micrograph, all of these demonstrate the high validity of the algorithm.