- 學位論文
基於條紋結構光技術之自動化三維掃描系統
Automated 3D Scanning System Based on Structured Light Stripe Technique
摘要
本論文的目的是開發一套自動化的結構光掃描系統，應用於全口牙模掃描時，能在一分鐘內快速自動化完成。解決傳統數位全口牙模不易取得的問題，並提供更高解析度的模型，且縮短取得數位模型的時間。此系統包含自動化攝影機與投影機校正、影像編碼與解碼、三角量測法，以及多筆數位模型定位、整合與重建。 本論文開發直條紋演算法取得三維點資料，使用對極幾何(Epipolar geometry)原理簡化原始直、橫條紋演算法，也針對最近鄰點迭代(Iterative Closest Point ,ICP)演算法，進行參數設定與結果比較。 藉由量測標準球，分析直條紋演算法與直、橫條紋演算法的精度，透過誤差分析可得知兩種演算法結果相近，且精度都在±0.03mm內，符合數位牙模相關應用的需求。
並列摘要
The purpose of this thesis is to develop an automatic structured light scanning system which can complete automatic scanning in one minute when it is applied to full mouth dental cast scanning. To improve the difficulty of obtaining traditional digital full mouth dental cast, we provided higher resolution digital model and developed a system consuming less time. This system includes automatic projector-camera calibration, encoding and decoding images, reconstruction by triangulation, multi-digital models registration and integration, as well as the surface reconstruction from unorganized point clouds. In this study, we proposed Vertical Stripe pattern algorithm which adopts Epipolar geometry to obtain three-dimensional point data for simplifying Vertical and Horizontal Stripes pattern algorithm. We also set parameters to Iterative Closest Point (ICP) algorithm and compared to the results. Both Vertical Stripes pattern algorithm and Vertical and Horizontal Stripes pattern algorithm are used to analyze the error by measuring the standard ball. The result shows that two algorithms are similar, and the error is within ±0.03mm. As the result, the result satisfies the digital dental model requirements.