| With the development of computer vision technology,3D vision measurement technology has made great progress,and it has been widely used in industrial manufacturing,medical diagnosis,cultural relics protection and so on.Among them,3D visual measurement technology based on linear structured light has always been the focus of research because of its advantages of high precision and high measurement efficiency.On the basis of combining the advantages of existing visual measurement technology and system characteristics,in order to further improve the integrity and accuracy of object measurement,this paper studies the object surface reconstruction technology based on linear structured light visual measurement,and proposes a certain innovative measurement scheme.The built measurement system can realize the full picture scanning of objects without relative pose.The main research contents are as follows:1.First of all,the underlying principle of linear structured light vision measurement system is studied and the direct laser triangulation method is selected as the principle of measurement in this paper.Then the image processing algorithm involved in this system is introduced in detail.Aiming at the research on the algorithm of extracting the center line of linear structured light fringe,this paper firstly introduces the traditional algorithm of extracting the center line of light fringe which is widely used.On the premise of summarizing the advantages and disadvantages of traditional algorithm,an algorithm of extracting the center line of light fringe which can adapt to various width changes is proposed.This algorithm obtains the initial center of optical fringe based on skeleton thinning,and then uses Bazen method to calculate the normal direction to obtain the optical fringe of nearly equal width.Finally,Steger algorithm is used to extract the optical fringe accurately.Under the premise of guaranteeing the extraction accuracy of the algorithm,the possible problems of line breaking and noise in the optical fringe extraction algorithm are solved.Improved computing efficiency by about 62%.2.The coordinate conversion model of linear structured light vision measurement system is established,the camera imaging principle involved in the system and the parameter calibration method required for coordinate conversion are described in detail,and the appropriate calibration method is selected according to the actual application scene of this paper,including camera calibration,optical plane calibration,direction of motion calibration and binocular vision calibration.And based on this design of the first set of visual measurement system:monocular visual displacement measurement system,the specific experimental steps and measurement results of the system are introduced and analyzed in detail,after many tests proved that the accuracy of the measurement system is stable within 0.3mm.3.Point cloud processing technology and related algorithms are studied,with emphasis on point cloud noise reduction,point cloud density reduction sampling and point cloud splicing algorithms.In order to solve the blind area problem of monocular vision measurement system,binocular vision measurement system is built,and binocular vision calibration and ICP point cloud precise splicing algorithm are introduced to realize three-dimensional object measurement.Finally,combined with stereo matching technology,the binocular vision measurement system was built,which realized the secondary measurement and point cloud stitching of visual blind areas in multiple perspectives on the premise of no position and pose conversion relationship.After many tests,it has been proved that the measuring accuracy of the binocular vision measuring system designed in this paper is within 0.4mm.Based on this,the improved binocular vision full picture measuring system realizes the full picture scanning of objects in the picture.Through multiple scanning,the surface integrity of the reconstructed objects can be greatly improved.Finally,the three-dimensional surface reconstruction of the object is realized,which meets the design requirements of the system and has certain practical value.In terms of the measurement scheme,the characteristics of the measurement system are not the same in different scenarios. |
