Research on Image Fusion Technology Based on Representation Learning

Image fusion technology is a concept put forward in the late 70s of the 20 century, is a category of information fusion in image as the research object, the information fusion technology based on the theory of image processing technology. This paper is in the framework of image fusion based on, combined with the structure is widely used in the field of signal processing and its applications -sparse, puts forward a said learning for image fusion method based on, through learning, the image is divided into the low frequency part of the sparse and can be approximately regarded as the high frequency part of sparse, then frequency part of sparse representation, are fused by using different fusion rules, finally the fused low frequency part and high frequency part after inverse discrete wavelet transform for, the final fusion image can be obtained.