The combination of visible and infrared imaging is a significant and common issue. In order to merge the characteristics seen in visible and infrared pictures into a single image, many fusion techniques have recently been developed. These cutting-edge techniques are extensively employed in several applications, such as target detection, picture classification, and image pre-processing. Finding important elements in the original photos and combining them to create the fused image is the main challenge in image fusion. Discrete wavelet transform (DWT), contourlet transform, shift-invariant shearlet transform, quaternion wavelet transform, and other traditional signal processing techniques have been used for decades in the image fusion area to extract picture characteristics. for the purpose of fusing visible and infrared images. However, the fused picture could contain artefacts from various techniques. Optimization-based fusion techniques are suggested as a solution to these issues. To get the best answer (fused picture), these approaches require several iterations. Due to the numerous rounds, these optimisation techniques could oversmooth the fused image. Furthermore, edge-preserving picture fusion techniques are also gaining popularity. In order to identify the borders of visible and infrared sensor images, this work used anisotropic filtering. For the objective of fusion, this approach uses edge-preserving smoothing filtering/process. Additionally, machine learning has emerged as a very active research technique with applications in several image processing domains. In order to create a single image that incorporates all of the information from both visible and infrared images, we suggest an efficient image fusion technique based on an unsupervised machine learning framework based on principal component analysis (PCA). Anisotropic filtering is used to first break down the source pictures into their fundamental components and detail content. We employ PCA to extract multi-layer features for the detail content. We create the final fused detail content using the weighted-average technique and these attributes.

The combination of visible and infrared imaging is a significant and common issue.