Neural network-based visual system calibration method

Abstract

The invention discloses a neural network-based visual system calibration method. The method comprises the steps of deriving a Faugeras calibration algorithm suitable for a Delta robot vision system according to a position relationship between a robot and a video camera; solving a linear intrinsic parameter and a linear extrinsic parameter for vision system calibration by utilizing the Faugeras calibration algorithm, and taking the linear intrinsic parameter and the linear extrinsic parameter as an initial weight value and an offset of a neural network; and deriving a 2-2-3-1 four-layer neural network structure suitable for Delta robot vision system calibration by utilizing Faugeras, wherein a transitive relation between an input layer and a hidden layer of the network represents an extrinsic parameter for video camera calibration, a transitive relation between the hidden layer and an output layer of the network represents the extrinsic parameter for the vision system calibration, and an activation function of the neural network is nonlinear due to existence of nonlinear factors such as distortion and the like. In an output of the neural network, an X axis and a Y axis of a coordinate system of the robot have different network characteristics, so that the 2-2-3-1 four-layer distributed neural network structure is adopted; and in addition, the experimental calibration precision of the method is 0.109mm, and the precision of a conventional Faugeras calibration algorithm is 0.535mm.

Description

technical field [0001] The invention relates to a neural network calibration method, in particular to a calibration method for a vision system used to control material grabbing of an industrial Delta robot. Background technique [0002] There are many influencing factors in the imaging process of the Delta robot vision system, such as radial distortion, tangential distortion, and measurement errors. The final camera imaging model has become a complex nonlinear model, and the mapping relationship between object points and image points has also become a non-linear mapping relationship. Many researchers have conducted in-depth research on the imaging relationship of cameras and proposed many new calibration methods. Traditional calibration methods cannot include all nonlinear factors in the imaging process, and can only select the main factors, while ignoring other uncertain factors. However, an accurate mathematical calibration model will lead to cumbersome calculations and ...

AI Technical Summary

Problems solved by technology

[0003] In order to overcome the problems of low accuracy, slow iteration speed and cumbersome calibration process of the traditional calibration...

Images