3D Point Cloud Real Texture Generation System Based On Binocular Line Laser Scannin

At present,3D reconstruction technology based on line laser has been widely used in many fields,such as reverse engineering design,cultural relics restoration and workpiece detection.With the in-depth research of this technology,in the face of the demand for 3D reconstruction.People are not only satisfied with obtaining the point cloud data of the measured target,but also hope to get the real texture and depth information of the 3D object to make the reconstruction effect more realistic.Therefore,the detailed research on how to improve the accuracy of line laser calibration and how to establish the mapping relationship between 3D information and texture information is conducted in this paper,and a 3D point cloud real texture generation system based on binocular line laser scanning is built in this paper.Firstly,in order to make the texture information obtained from the 3D reconstruction of the measured object the same as the real texture information on the surface,the system built in this paper obtains the real texture of the measured object surface based on a dual color industrial camera and a 780 nm near-infrared laser whose wavelength exceeds its spectral response range and does not affect the acquisition of texture information.Secondly,aiming at the problem that the color camera cannot obtain the near-infrared laser fringe,the IR-CUT optical element of the dual-color camera is improved in this paper so that the camera can obtain the high-definition line laser fringe images.Combined with the principle of binocular outer polar line constraint,an optical plane calibration method which can improve the calibration accuracy and stability of the system is proposed.Thirdly,the algorithm of Steger is used to extract the light stripe center of the laser stripe image collected by the color camera.In view of the low efficiency of extracting the stripe center and the large amount of image data processing,the ROI region of the image is segmented according to the shape characteristics of the line laser stripe before running Steger algorithm.The experimental results show that,the average speed of extracting the center of light bar by this method is approximately 3.3 times higher than that before improvement.Finally,aiming at the problem of matching 3D information and texture information,a method combining dual camera field center correction and binocular stereo calibration to establish the matching relationship between point cloud camera and texture camera is proposed in this paper.The accurate mapping of surface 3D shape information and surface texture information of the measured object is realized.The experimental results show that the binocular laser 3D point cloud real texture generation system built in this paper is in the measurement range of 400 mm ～ 600 mm,the deviation obtained by plane fitting using the optical plane calibration method proposed in this paper is 0.0452 mm.The accuracy of 3D reconstruction can reach0.0472 mm,and the matching error of pixel points is 0.3234 pixel,which meets the needs of the system.