Infrared and visible image fusion method based on anisotropic diffusion and content aware fusion strategy

Infrared and visible image fusion is a process whereby infrared and visible spectral images are combined to enhance the visual quality of the resulting image and to retain more detailed information. Traditional infrared fusion techniques tend to suffer from information loss and lack of clarity of texture details. To retain more detailed information and improve image quality in infrared and visible image fusion, a method is proposed which involves extracting the base and detail layers using Gaussian filtering and anisotropic diffusion, respectively. The base layer is then fused using an average fusion strategy, the detail layer is fused based on a content-aware fusion strategy, and the processed base and detail layer images are fused into a final image. Experiments were conducted on several scenes from publicly available datasets. The experimental results demonstrate that the method proposed in this paper can generate fused images with clear targets and rich details, while simultaneously exhibiting a notable enhancement in indicators such as average gradient and spatial frequency.