基于非下采样Shearlet和WNMF的红外热波图像融合

Abstract

A fusion method of infrared thermal wave images based on nonsubsampled shearlet transform and weighted non-negative matrix factorization was proposed.After infrared thermal wave image sequences are decomposed by nonsubsampled shearlet transform,dynamic weighted non-negative matrix factorization algorithm is adopted for fusion processing of low-frequency coefficients.The weighted coefficients of algorithm are adjusted dynamically according to the mutation degree of image pixel.As a result,the defect areas of infrared thermal image are highlighted.The fusion rule based on area summodified-laplacian is used for high-frequency coefficients to preserve the defect edges and details.Experimental results show that,the method proposed in this paper has superior performance in subjective visual effect and objective quantitative evaluation indices such as edge preserving degree,correlation and running time.The proposed method is fast and effective.It can keep the edges of infrared thermal images more complete and clear.In addition,the proposed method can be adopted for image fusion of multi infrared thermal image sequences effectively.Therefore,the proposed method has high practical value in the thermal wave nondestructive testing.