Design And Implementation Of Structured Light Scanning Device

The 3-D line-structure light scanning has a wide application for its advantage such as low cost, small volume, light weight, ease and flexibility in the field of 3-D scanning of object. The study was developed based on line-structure light scanning. To realize automatic acquisition of 3D information of objects surface and the data registration of multidimensional scanning results, a calibration method was proposed based on simple plane to calibrate the system posture and a new line-structure light scanning system was constructed based on triangulation in this study. The main study content and conclusions are as follows:(1) In order to determine the posture of the scanning system, this paper has designed a self-calibration board to obtain 3D information and calibrate the system. This board contains 10 black circles, on the four corners placed with four big circles, in each big circle surrounding with 0, 1, 2, 3, small circles. These small circles are used to calibrate the system with four kinds of posture, and the big black circles are extracted as feature points, and these feature points are used to transform the two-dimensional image information to 3D information. Experimental results show that, the result of the system calibration posture obtained by the method is very accurate and the error of the 3D coordinate data calculated by that is less than 1.15 mm. It is satisfied the requirements of the calibration system attitude and the data solved.(2) For extracting the laser stripe center. A common extraction algorithm that is the gray barycenter method was used in this study to solve the problem of how to extract the laser stripe center which is a difficult problem for all active 3D vision measurement system. The calculation time was reduced greatly by setting a threshold to extract the next center point on the premise of extracting a prior center point. The multigroup experiments indicate that the average time is 16 ms for light extraction in 640 ? 480 image, which is satisfied the requirement of extracting the laser stripe center for the system.(3) A self-made calibration board was designed based on calculating the transformation parameters to solve the problem of how to make the data scanned with four positions into the same coordinate system. It sets four ranges to scan the object because of the need for scanning around the object to be measured and the data acquired by each range scanning should be in an independent coordinate system. The conversion matrix was computed using the planar calibration target attitude, and the scanning results of the four ranges was made to a unified coordinate system, which is the point cloud registration described in this paper. Through results of four groups of experiments by using this method, it shows that the relative error is less than 10 pixels.