一种基于空间编码结构光的稠密三维重建算法

In this paper, we propose a novel, simple and fast dense three-dimensional reconstruction method based on spatially encoded structured light. Firstly, for the problem that color stripe-based mothed are not suitable for colored objects, a structured light coding pattern composed of sinusoidal stripes and pseudo-random dots is proposed. All the stripes are coded with the same color, and pseudo-random dots are introduced to identify the stripes. Secondly, aiming at the low spatial resolution of the point cloud obtained by the traditional stripe-based mothed, a dense three-dimensional reconstruction method based on local phase matching is proposed. The center points extraction algorithm is used to accurately extract the stripes. Obtain the phase of pixels between adjacent stripes by Fourier transform. Finally, all the pixels between adjacent stripes are matched to achieve dense three-dimensional reconstruction. The experimental results show that the proposed algorithm in this paper can still effecA Dense Three-dimensional Reconstruction Algorithm Based on Spatially Encoded Structured Lighttively acquire the three-dimensional information of the object surface under the condition that the surface structure of the measured object is complex, and has better accuracy and higher spatial resolution.