Image Fusion Technology Based on MSVD and Realization of Matlab

Image Fusion Technology Based on MSVD and Realization of Matlab

Image fusion refers to combining two or more image information from different sources, different spatial resolutions, different sensors or captured at different times to form a new image to improve image information, image quality, data visualization and Ease of analysis and other aspects of performance. In practical applications, image fusion can be used in medical image analysis, remote sensing image processing, security monitoring, virtual reality and other fields. This paper mainly introduces an image fusion technology based on multi-resolution singular value decomposition (MSVD), and gives the corresponding matlab code implementation.

1. Algorithm principle

Multi-resolution singular value decomposition (MSVD) is a classic matrix decomposition method, which can decompose a large matrix into multiple small matrices, so as to realize the analysis and processing of the matrix. In image fusion, we can decompose the two images to be fused by MSVD respectively, and then process their eigenvalues ​​and eigenvectors to finally obtain the fused image information.

Specific steps are as follows:

1. The two images to be fused are decomposed by MSVD to obtain their eigenvalues ​​and eigenvectors.

2. Sort the eigenvalues ​​of the two images from large to small, and select the top k (k is a custom parameter) as the fused eigenvalues.

3. Based on the selected eigenvalues, construct a new matrix of eigenvectors.

4. The fused image information can be obtained by multiplying the new eigenvector matrix with the original image data.

Two, matlab code implementation

The following is the matlab implementation code of the MSVD-based image fusion algorithm:

% Image read
I1 = imread('image1.jpg');
I2 = imread('image2.jpg');

% Grayscale
I1_gray = rgb2gray(I1);
I2_gray = rgb2gray(I2);

% Convert image to double