PROJECT TITLE :

Local-Adaptive Image Alignment Based on Triangular Facet Approximation

ABSTRACT:

Today, the primary challenge in panorama stitching research is achieving precise and efficient image alignment. Using triangular facet approximation, this work provides a local-adaptive picture alignment approach based on camera coordinates that is superior to the imaging model of cameras. When two weighting algorithms are combined to build a more robust planar transformation model, it becomes local-adaptive. The planar and spherical triangulation algorithms are introduced to align normal and fisheye photos more effectively by approximating the scene as a set of adjacent triangle faces. Both qualitative and quantitative tests on a variety of hard instances demonstrate the effectiveness of the proposed strategy.