DOE-based Structured Light For Robust 3D Reconstruction

Abstract

Spatial structured light (SL) enables three-dimensional measurements in a single shot by projecting patterns with coded information. However, conventional structured light systems typically employ a digital light processing projector (DLP) to project specific light patterns or coded patterns onto the surface of the object under examination. This reliance on projectors limits the miniaturization potential of such structured light systems, thereby hindering their suitability for industrial applications. The main contributions of the work in this paper include: 1) We introduce a novel and practical depth-aware miniature solution for smart terminals that employs a diffractive optical element (DOE) instead of a digital projector to project coded patterns. 2) We propose a one-dimensional SL pattern generation method that is successfully etched onto a DOE. 3) In contrast to traditional image processing techniques, all crucial steps in the pattern decoding process, such as corner detection and block recognition, are executed using deep learning methods, ensuring heightened robustness in real-world applications. The experimental results verify the effectiveness of the proposed method.