Fusion of visible and thermal Solar Panels images

Abstract

A study has been conducted on image fusion methods in this field. After that, it was decided to adopt one type of method and follow it to develop an effective image fusion method. Image merging, which involves combining two or more images from different or similar sources to create a new image with more comprehensive information. The research specifically focuses on comparing the effectiveness of the convolutional neural networks (CNN) method with other techniques such as LATLRR and NSST when applied to thermal and color image fusion. By evaluating six factors, it is clear that the CNN method achieves the best results among the three methods. The study also includes a comprehensive discussion of the types of image fusion, levels of processing, and the importance of image fusion in multiple domains, in addition to visual and quantitative evaluations of the performance of different fusion methods