A Novel Edge-Aware Image Fusion Approach for Thermal and Visible Images

Abstract

In this work, we present a novel image fusion approach aimed at enhancing the quality of fused images generated from thermal (infrared - IR) and visible image pairs. We compare our proposed method with the conventional Discrete Wavelet Transform (DWT) approach, using a comprehensive set of image quality metrics, including Peak Signal-to-Noise Ratio (PSNR), Spatial Frequency (SF), and Human Visual System (HVS)-based metrics. Our experimental results unequivocally demonstrate the superior performance of our proposed fusion method. Notably, our approach excels in preserving and enhancing edge information, a vital aspect in various image fusion applications such as surveillance and medical imaging. The consistently higher PSNR values across all sample images reflect our method's exceptional edge enhancement capabilities. Its remarkable edge enhancement capabilities, computational efficiency, and adaptability position it as an efficient and versatile alternative to conventional DWT-based approaches and deep learning-based methods.