3-D surface registration based on line light scanning

In the near-infrared tomography of medical imaging technology, the accuracy of geometric parameters about organism contour determines directly the detection reliability of heterogeneity in organism ^{[1, 2]}.Therefore, this paper takes the irregular convex body as research object to simulate the organism structure and uses the technology of computer vision to acquire the 3D point cloud in different viewing direction. Registrating all the views, the complete 3D model can be produced in the end, which laid the foundation for further study of the near-infrared diffuse optical tomography technology. In this paper, we discuss the method of 3D registration based on point laser markers, which has been verified in experiment. Experiment results showed that this method can reproduce precisely target object and improve the accuracy of the 3D scanning system. Besides, compared with the traditional marker method, this method can recognize the marker easily and avoid the error caused by pasting the markers on the surface of target object.