VLSI-Band Fusion Based VI-IR Image Fusion using Stationary Wavelet Transform

In recent times, the fusion of infrared (IR) and visible imaging (VI) has emerged as a crucial and frequently encountered challenge in various applications. Numerous fusion techniques have been introduced to amalgamate the distinctive characteristics present in IR and VI, ultimately producing a unified image. These cutting-edge methodologies find extensive utility in diverse domains such as image preprocessing, target recognition, and image classification. The central challenge in image fusion revolves around effectively extracting salient features from the source images and subsequently amalgamating them to create the fused image. This study primarily concentrates on the hardware-oriented implementation of VLSI-based VI-IR image fusion mechanisms. The workflow begins by reading the images within the MATLAB environment. Subsequently, the stationary wavelet transform (SWT) is employed to decompose both the VI and IR images into multiple bands. Specifically, the low-low frequency bands are converted into text files. The critical band fusion process is executed on these low-low frequency bands, employing a decoder-based adder to merge the information from both images. Finally, the fused image is reconstructed using MATLAB from the obtained text files. The results of our simulations reveal that the proposed approach not only significantly reduces hardware requirements but also enhances fusion performance compared to conventional methods.