Graph-Partitioning-Based Matching Algorithm for 3D Reconstruction in IoT Environments

Abstract

This paper deals with the 3D image reconstruction within an Internet of Things (IoT) environment. To increase feature matching accuracy, an optimization problem based on graph partitioning is applied on extracted keypoints. By assigning efficient resource, the algorithm becomes robustness and reliability in varying data density and environmental dynamics. By analyzing relationships based on graph partitioning, the algorithm reaches optimal partitions for reference and query keypoints, This algorithms improves the matching process and enhance the accuracy of  3D reconstruction.