基于多尺度分析和数学形态学融合的边缘检测

EDGE DETECTION BASED ON FUSION OF MULTI-SCALE ANALYSIS AND MATHEMATICAL MORPHOLOGY

图像边缘检测的关键是在尽量多地检测到图像边缘的同时更有效地抑制噪声,为此提出了一种基于多尺度分析和数学形态学融合的边缘检测方法。首先利用小波变换将图像分为高频和低频部分;然后再分别进行处理:高频部分利用小波系数的层内和层间相关性对系数调整后实现边缘检测,低频部分利用数学形态学进行边缘检测;最后将各自得到的边缘图像通过多尺度边缘融合,从而实现边缘提取。实验结果表明,同单纯基于多尺度分析方法或基于数学形态学的边缘检测方法相比,提出的方法具有更好的噪声抑制和边缘细节保护功能。

The key to image edge detection is to suppress noises more effectively while detecting as many as possible the edges of image.For this purpose,this paper proposed an edge detection method based on the fusion of multi-scale analysis and mathematical morphology.Firstly,the image is decomposed into high and low frequency parts using wavelet transform.Then,they are processed respectively,for the high frequency part,the edge detection is realised after adjusting the wavelet coefficients using correlation between inter and intra wavelet coefficients,and for low frequency part,the edge detection is conducted using mathematical morphology;finally,the edge images derived respectively are merged by multi-scale edge fusion so that to accomplish edge extraction.Experimental results show that the proposed method has better effects in noise suppression and edge detail protection compared with single multi-scale analysis or mathematical morphology-based edge detection method.