The hybrid calibration of linear structured light system

Abstract

The calibration of structured light system is to identify the relationship between the image coordinates and the corresponding three dimensional (3D) world points, and is the key part of the structured light 3D vision inspection system. A hybrid calibration method of linear structured light system, based on linear calibration and back propagation (BP) neural network, is proposed in this paper. This hybrid method benefits from the neural network's ability that multilayer feed-forward networks are capable of approximating an arbitrary continuous nonlinear function and of solving least squares problem by compensating nonlinear errors of structured light system brought by lens distortion and other nonlinear factors. The results of experiments have shown that the proposed method is robust and effective.