Research on image fusion algorithm based on Compressed Sensing theory

Abstract

Visual effect and fusion quality of the traditional image fusion method is relatively poor. Compressed Sensing (CS) as a new image fusion method is proposed, which is simple and easy to implement. At first, the method uses wavelet transform to the original images and gets sparse matrix by sparse processing of wavelet coefficients. And then coefficient matrix is fused by the method of coefficient absolute value maximum. After that we use the method of random observation to get compression sampling to the coefficient matrix after fusion. Image restoration is obtained from compression sampling by solving the optimization problem. The method can recover the image with a small number of sample points because we have handled wavelet coefficients by sparse processing. Experimental results show that the image quality obtained by this method is better than the method of traditional coefficient absolute value maximum fusion at the same sampling point and we can achieve better results by using this method at a small number of sampling points.