Study of 3D Target Replacement in AR Based On Target Tracking

Jiahui Bai, Guang-Yu Nie, Weitao Song, Yue Liu, Yongtian Wang

Abstract: Augmented reality application faces the problem of 3D target replacement for better mixing effect, however, the existing methods have such problems as large amount of calculation and high hardware requirements. Inspired by the development of deep learning in the target detection and target tracking, this paper introduces a neural network and trains a detector to identify the target from the binocular picture to generate the three-dimensional position of the target. By using the difference of the positions between the two images and the camera parameters, the depth calculation formula is used to generate the position of the target. Experimental result shows our method can realize the 3D position generation of the target, which provides a new idea for solving the replacement of objects in the augmented reality system.