Camera calibration for linear structured light system based on BP neural network

Abstract

In this paper, BP neural network algorithm is applied to camera calibration for the linear structured light 3D digital measurement system and BP network model of camera calibration with double-input double-output is established based on the Levenberg-Marquardt algorithm (LM algorithm). The simulation tests which verify the data of characteristic points obtaining from 3D motion platform show that BP network with double-input and double output has advantages in terms of generalization and approximation ability of nonlinear function, and high accuracy in calibrating linear structured light 3D measurement system.