关键词: 小波变换, 各向异性扩散, 状态权, 边缘检测

Abstract: Edge detection based on wavelet and anisotropic diffusion is proposed. It should be carried out the way the denoising image is decomposed by wavelet so as to gain high frequency and low frequency wavelet coefficients. Anisotropic diffusion method is used to achieve state weights after the high frequency wavelet coefficients are normalized. The new higher frequency wavelet coefficients both to remove noise and keep the structure unchanged are achieved when state weights act on original wavelet coefficients. The low frequency wavelet coefficients are denoised by means of wavelet threshold. The high frequency and low frequency wavelet coefficients are detected through detecting the local maximum of the wavelet transformation coefficient modulus. The final edge images are obtained. Owing to introducing the mixed method of combining wavelet with anisotropic diffusion, it is indicated from the experimental result that the noises are restrained efficiently and the edges are consecutive and clear.

Key words: wavelet transform, anisotropic diffusion, state weights, edge detection