Segmentation Based Fine Registration of Very High Resolution Multi-Temporal Images

This procedure seeks to provide an overview of both traditional and contemporary picture registration techniques. The practise of overlaying photographs from the same scene that were acquired at various times, from various angles, and/or by various sensors is known as image registration. Two photos are geometrically aligned via the registration (the reference and sensed images). The examined techniques are divided into two categories-area-based and feature-based-and four fundamental Image registration phases. After colour correction, colour blending is performed to further smooth the colour transition between neighbouring photos. The mean values of all the overlapping areas between each pair of adjacent images must be extracted in order to create the global error function. To put it another way, this results in unneeded delays because the colour correction and panorama stitching processes cannot begin until the entire image sequence has been captured. The color/luminance inconsistency issue that existed in panoramic imaging is addressed in this process with an effective image stitching approach. The suggested method is intended for panorama stitching on mobile devices, and it enhances the colour blending and correction processes that were previously proposed in order to speed up processing.