Image processing algorithm for extracting the phase map from structured lights

Abstract

In structured light three-dimensional measurement systems, the position of stripe in the structure light image reflects the three-dimensional shape information of the object being scanned. In this paper, we propose a semi-automatic 3D image processing algorithm that extracts the phase map from the distorted line pattern. The proposed method uses the stripe extraction approach to segment and cluster the stripe projected over the inspected object after Gaussian convolution denoising. Then a noise-aware smoothing spline fitting process is applied to the result to obtain single continuous stripe and the coordinates of the clustered stripes are calculated to form a phase matrix through a cubic polynomial interpolation procedure. After the camera and projector calibration, the 3D coordinates of the object are computed accurately and robustly. The experimental results show that the work is particularly effective.