A 3D Scene Information Enhancement Method Applied in Augmented Reality

• DOI: 10.3390/electronics11244123
• Publication Date: 2022-12-12T09:05:22Z
• Authors: Bo Li; Xiangfeng Wang; Qiang Gao; Zhimei Song; Cunyu Zou; Siyuan Liu
• Keywords: RANSAC; Scale-invariant feature transform; Feature (linguistics)

Aiming at the problem that detection of small planes with unobvious texture is easy to be missed in augmented reality scene, a 3D scene information enhancement method grab for proposed based on series images real taken by monocular camera. Firstly, we extract feature points from images. Secondly, match different images, and build three-dimensional sparse point cloud data camera internal parameters. Thirdly, estimate position size cloud. The can used provide extra structural reality. In this paper, an optimized extraction matching algorithm Scale Invariant Feature Transform (SIFT) proposed, fast spatial recognition RANdom SAmple Consensus (RANSAC) established. Experiments show achieve higher accuracy compared Oriented Fast Rotated Brief (ORB), Binary Robust Scalable Keypoints (BRISK) Super Point. effectively solve missing faces ARCore, improve integration effect between virtual objects scenes.