Hash Indexing-Based Image Matching for 3D Reconstruction

Image matching is a basic task in three-dimensional reconstruction, which, recent years, has attracted extensive attention academic and industrial circles. However, when dealing with large-scale image datasets, these methods have low accuracy slow speeds. To improve the effectiveness of modern methods, this paper proposes an method for 3D reconstruction. The proposed can obtain high through hash index very short amount time. core includes two parts: creating table index. former used to encode local feature descriptors into codes, latter search candidates query points. In addition, extremely robust scaling transformation by using various verifications. A comprehensive experiment was carried out several challenging datasets evaluate performance matching. Experimental results show that HashMatch presents excellent compared state-of-the-art both computational efficiency accuracy.