A novel edge detection approach using multi-resolution image

Focusing on the problem of multi-focus image fusion, a wavelet image fusion method based on Multi-scale Edge Detection is proposed. With the characteristics of detecting multi-directional edge and exiguous edge by Frei-chen template, this paper put forward fusion fuel based on the edge activity measurement of wavelet coefficients. The specific integration process is to make wavelet decomposition for two multi-focus images to obtain respective high-frequency and low-frequency coefficients and calculate corresponding edge active wavelet measurement of each wavelet coefficient, which is quadratic sum of the projections respectively from edge subspace and line subspace of Frei-chen template. The article adopts root mean square error and standard deviation as objective evaluation standard and combines subjective evaluation to compare the fusion methods proposed respectively by this article and other two methods. Results demonstrate the fused image obtained by the method proposed by the paper has clean image and better quality.