1. Infrared and visible image fusion based on rolling guided filter and ResNet101
- Author
-
Kaizhi Shang, Guangqiu Chen, Shuai Wang, and Yucun Chen
- Subjects
Fusion ,Image fusion ,Computer science ,business.industry ,Feature extraction ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,Filter (signal processing) ,Residual ,Image (mathematics) ,Gaussian filter ,symbols.namesake ,Computer Science::Computer Vision and Pattern Recognition ,Fuse (electrical) ,symbols ,Computer vision ,Artificial intelligence ,business - Abstract
An infrared and visible image fusion algorithm based on rolling guided filter and residual network is proposed. The source image is divided into the base and the detail layer images by using the rolling guided filter. For the base image fusion, the source image is sent to the residual network, the multilayer depth features are extracted and integrated into a weight map, and the fused base layer image is obtained by weighted fusion algorithm. The absolute value max of the coefficient and Gaussian filter is used to fuse the detail layer image. Finally, the fusion image is obtained by multi-scale inverse transform. The experimental results show that, compared with the existing fusion methods, the proposed method has achieved good results in both subjective and objective evaluation.
- Published
- 2021
- Full Text
- View/download PDF