Overview of 3D Reconstruction Methods Based on Multi-view

摘要

3D reconstruction based on multi-view is a hot research topic in computer vision due to more information and wide scope of the view of multiple views. Feature detection and matching, fundamental matrix estimation, camera self-calibration, 3D reconstruction and dense surface reconstruction are key proportions of 3D reconstruction. This paper summarizes the main algorithms in these parts, analyses and compares merits and drawbacks of the methods. The aim is to provide a concise, complete understanding of the subject so that the researchers in this field can have a holistic view of the problem and provide more efficient and unique solutions in the future.