一种改进多尺度形态边缘检测算法

Improved multi-scale morphological algorithm for edge detection

提出一种改进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图像的不同尺寸的边缘信息,运用本文提出的基于直方图配准的自适应方法对所获不同尺寸的图像边缘进行融合,获得理想的图像边缘,有效地消除了噪声。实验结果表明当图像中加入了强度超过20%的椒盐噪声后,检测到的边缘仍然非常清晰。

An improved multi-scale morphological edge detection algorithm is presented. Edges of different size were detected by using different scale operator, and then they were combined to produce all the edges of interest in an image with self-adaptive fusion method based on histogram matching. Meanwhile, fuzziness of image edge detection was improved and noise was removed efficiently. Experiment results show that distinct edges can still be obtained when intensity of salt ＆ pepper noise is over 20%.