Fusion of infrared and visible images based on multi-features

In allusion to the lower overall contrast and smaller detail contrast of a fused image from the conventional fusion methods, an effective multi-feature weighted multi-resolution image fusion algorithm was proposed. Firstly, the edge features and average gradient features were extracted from a low frequency coefficient after multi scalar decompose, while the correlated signal intensity ratio feature was extracted from a high frequency coefficient. Then, the high frequency coefficient of the fused image was obtained from the pixel-level weighted average image fusion conducted by the edge feature fusion. Furthermore, a novel combination map was proposed to process the frequency coefficient from the same place with two patterns to solve the problem that the simple weighted method is not effective for retaining the edge and texture information. Finally, the low frequency coefficient of the fused image was obtained by adaptive weighted method based on regional average gradient and the target image was obtained by inversing multi-scale transformation for low frequency and high frequency coefficients. The experiments on fusing infrared and visible images show that the proposed algorithm is better than the classical methods. And the fusion quality indexes, such as standard deviation, spatial frequency, information entropy and average gradient have increased by 15.12%, 4.30%, 6.15% and 3.44%, respectively.