平稳小波的侧抑制网络边缘提取方法

侧抑制网络具有“突出边框,增强反差”的功能,但其增强反差的同时也增强了噪声,其抗噪性较差。正交小波阈值去噪法对图像噪声有一定抑制作用,然而对于图像中噪声幅度较大的情况下,正交小波变换会使图像边缘失真,甚至模糊。基于平稳小波变换的图像去噪法,可以有效降低噪声的同时较好地保持图像的质量,与基于正交小波变换的阈值降噪方法相比,有明显的优越性。因此利用新Bubb le函数将平稳小波和侧抑制网络结合起来,用平稳小波变换去除噪声、侧抑制网络突出边框,从而准确、有效地提取出图像的边缘。

Edge Detection based on Stationary Wavelet and Lateral Inhibition Network

Lateral inhibition has the function that can enhance the contrast and give prominence to changes,but it is sensitive to noise.Although orthogonal wavelet threshold denoising can restrain the noise,yet it sometimes causes image edge distortion seriously and blurs details when there is strong noise in image. Compared with orthogonal wavelet transform,stationary wavelet transform has some advantages,such as protecting the image from distortion while denoising.So we use bubble function to combine lateral inhibition with stationary wavelet to eliminate noise and enhance the contract.By doing that,we can pick up the edge exactly,effectively and quickly.