Study On Position Measurement Of H-Shaped Steel Using Line Structured-Light

It is a key step for the steel cutting system to measure the position of the H-shaped steel accurately. Traditional methods of manual and contact measurement seem to be lack of efficiency, and the measurement accuracy is influenced by human factors. What's worse, the manual methods can't be integrated with automatic equipments. In this dissertation, the method of line structured-light measurement, which has been widely used in the industrial field, is used to obtain the accurate positions of four points in the cross section of the H-shaped steel. After a thorough study about the basic principle of the line structured-light measurement, some improvements on system calibration and image processing are proposed in order to make the measurement system to meet the need of fast speed and high accuracy. The following aspects are the main contents:Firstly, the dissertation studies the principles of several popular methods in 3-D vision measurement. Then the main attention is paid to analyze the mathematical model of line structured-light measurement. According to the need of the engineering application, a system that can automatically measure the position of H-shaped steel is set up based on the principle of line structured-light measurement. The system consists of a software platform and various hardware modules. The calibration of line structured-light system can be mainly divided into two aspects: the calibration of camera parameters and the calibration of light plane parameters. In the camera calibration stage, a fully automatic calibration method using an improved checkerboard target is proposed. The entire calibration process does not need any human intervention, which greatly reduces the workload in the calibration process. As the experiment results show, the calibration accuracy of the proposed method is almost the same with the calibration method that needs considerable human intervention. A template matching algorithm with different weights is proposed. First, the template matching is used to extract the initial position of the point to be measured,and then continuous iterations are implemented to obtain the sub-pixel coordinates. In order to take full advantages of the parallel computing resources in modern computers,the original image processing program is optimized using the multi-thread technology and the Open MP standards respectively.