二进小波与改进的形态学融合的边缘检测算法

Image edge detection algorithm based on dyadic wavelet transform and improved morphology

针对小波变换边缘检测算法抗噪能力差、图像边缘不连续等缺点,提出一种将二进小波变换与形态学算子融合的边缘检测算法。利用新构造的二进小波滤波器边缘检测算法对含噪图像进行边缘检测,可以保留较多的边缘细节;利用新设计的多结构抗噪形态学算子对含噪图像进行边缘检测,抑制噪声良好;将两种算法得到的边缘结果按一定规则进行融合,利用Laplace算子锐化融合后的图像,得到最终的边缘检测结果。实验结果表明,该融合算法在抑制噪声的同时显示较多的图像细节,检测的图像边缘连续且准确。

Aiming at the shortcomings of wavelet transform edge detection algorithm,such as poor anti-noise ability and discontinuous image edges,an edge detection algorithm combining dyadic wavelet transform and morphological operator was proposed.Edge detection was performed on the noisy image using the newly constructed dyadic wavelet filter edge detection algorithm to retain more edge details.The newly designed multi-structure anti-noise morphological operator was used to perform edge detection on the noisy image to suppress noise.The edge results obtained using the two algorithms were fused according to certain rules,and the merged image was sharpened using the Laplace operator to obtain the final edge detection result.Experimental results show that the fusion algorithm displays more image details while suppressing noise,and the detected image edges are continuous and accurate.