Data Acquisition and Representation of Mechanical Parts and Interfaces to Manufacturing Devices
NRC '97 Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling
3-D modeling from range imagery: an incremental method with a planning component
NRC '97 Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling
Automatic 3-D digitization using a laser rangefinder with a small field of view
NRC '97 Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling
Active modeling of 3-D objects: planning on the next best pose (NBP) for acquiring range images
NRC '97 Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling
A portable three-dimensional digitizer
NRC '97 Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling
Shape reconstruction and texture sampling by active rectification and virtual view synthesis
Computer Vision and Image Understanding
| Hi-index | 0.00 |
Nowadays, 3D data are popularly performed in computer, and 3D browsers manipulate 3 D model in the virtual world. Till now, 3D scanner has been adopted to capture 3D information of an object in real world. Laser 3D scanner is the most popular one of 3D input devices. Though there had a large amount of H/W and S/W systems applied in many fields, the analysis of the best view is still an open problem that developed for acquiring complete data in a system. In this paper, in order to enhance the performance of a 3D laser scanning system, the low occlusion approach is analyzed to find the best view for scanning via considering the object position instead of the sensor position. Based on a 2-axis 3D color scanner developed by OES/ITRI, the best position (BP) of an object on the rotary table will be focused to meet the best view requirements for the low-cost scanner. Since the 2-axis 3D scanner only consists of linear and rotary moving apparatuses, in general, it needs 2 or more scanning processes to capture multiple layers to register and merge to a complete 3D file. In order to optimize the scanned layers and minimize the data amount, we propose an algorithm to effectively decrease the amount of scanned layers and to acquire the best-view data in the 2-axis 3D scanner based on the proposed best-position criteria. In the experimental results, the emphasized problems of the worst case really occur when the object is set at an infeasible position. And the best-view scanning is illustrated after the object is reallocated suitably with the best position by the proposed algorithm. Hence, beside the NBV problem, the low-occlusion criteria proposed in this paper also play an important role for acquiring complete 3D data. In future, we will focus on the development of the automatic procedure to combine the NBV solution to find out a more efficient scanning for any complicated 3D model.