Study On The Key Techniques For 3D Laser Color Scanner

Along with the development of computer technology, there is an urgent request inmovie, advertising, industry and national defence etc. That is how to obtain 3D colormodel of objects. 3D Laser Color Scanning System (3DLCS), as a new kind of scanner toacquire the full-color three dimension data of scanned object through laser, is a bestsolution for the requairment. It realtes to such subjects as optics, mechanics, andelectronics etc. 3DLCS has much significance on digital image processing and computergraphics. 3D reconstruction and Simplification are widely used in Virtual reality (VR),Graphic rendering, reverse enginnering to build CAD models, such as to build solidmodels based on industry computered tomography (ICT) images. Study on 3Dreconstruction from medial images is important to science and worthiness in practicalapplication. Our 3DLCS system has patented in China at 1996 (ZL 95 2 27897.9), and beensupported by National Natural Science Fundation (69775022) and National 863 Project(863-306-ZT04-06-3). It is no exaggerate to say that our 3DLCS is in the leading positionin our country. Unfortnately however, it is not enough for industrial application at present. In this dissertation, with the support of National Natural Science Fundation(69775022) and National 863 Project (863-306-ZT04-06-3), I have developed sometechniques to improve our system so that it can be used practically as early as possible. The main content I research includes 3D acquiring, 3D reconstruction, 3D meshing,3D data merging and 3D mesh simplification. Firstly, to the fields of 3D acquiring, it systematically concludes the popular 3Dacquiring methods. Moreover, in order to overcome the default of traditional constructedlighting system, we present the interactive constructed lighting system. Secondly,the dissertation pays deeply attention to the accuracy of 3DLCS, and IIIimproves the laser-extracting algorithm by "multi-frame-average " method. The resultsshow that it can improve the accuracy dramatically, and surmount many difficulties oforiginal one. Thirdly, to 3D data merging problem, the dissertation uses the color information tofind matching feature points, and give the solution by neural network. However, thismethod is not satisfactory. So I give another ICP (iterative closest point) method and showthe experimental results. Fourthly, to the 3D scattered data, the dissertation presents a new algorithm. Thescattered data is expanded and project into a plane; then the trianglization is executed. Theresults seem wonderful. Fifthly, in order to simplify the mesh while keeping the color attributes, thedissertation gives a novel way by processing the image data with graphics. All the imagedata are converted into graphics space and then simplification is performed. The practicalresults appear perfect, and some shortcomings are removed as well. Finally, I give the framework of our 3DLCS-200, and the prototype shows themethod is available.