簡易檢索 / 詳目顯示

回結果列表
研究生： 黃子峻
HUANG, TZU-CHUN
論文名稱： 基於 TSDF 重建之相機路徑規劃應用於機器人非平面繪圖
TSDF-Based Reconstruction and Camera Path Planning for Robotic Drawing on Non-Planar Surface
指導教授： 連震杰
Lien, Jenn-Jier James
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 資訊工程學系
Department of Computer Science and Information Engineering
論文出版年： 2025
畢業學年度： 113
語文別： 英文
論文頁數： 91
中文關鍵詞： 相機路徑規劃 、三維重建 、截斷有號距離函數 、非平面繪圖 、機械手臂
外文關鍵詞： Camera path planning, 3D reconstruction, Truncated Signed Distance Function, Non-planar drawing, Robotic arm
相關次數： 點閱：33  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 隨著大尺寸顯示面板在汽車設計中的普及，此趨勢促進了運用機械手臂針對曲面玻璃與工件進行高精度自動化塗膠的技術需求。然而，實際生產中常因缺乏工件三維 CAD 模型，或受製造公差影響，導致在執行精確塗膠前，必須先獲得工件的三維幾何資訊。本研究基於搭載低成本 Intel RealSense D415 RGB-D 相機的眼在手機械手臂系統，提出一套整合三維掃描重建與繪圖功能的完整架構。系統會依據工件幾何規劃相機路徑，進行多視角深度圖擷取，再以截斷有號距離函數（TSDF）與行進立方體（Marching Cubes）演算法重建非平面工件表面的三角網格，並根據重建結果生成繪圖路徑，控制機械手臂執行繪製以模擬塗膠作業。
    本研究的主要貢獻有三。首先，提出兩階段式 TSDF 重建與相機路徑規劃方法，以克服人工規劃相機路徑耗時且缺乏系統性策略的問題。其次，將 RealSense D415 的可接受範圍（acceptable range）概念整合至相機路徑規劃與掃描重建流程中，以過濾較不可靠的深度數據，提升重建結果的品質。最後，提出一種基於數位指示量表的低成本誤差評估方法，可沿繪圖路徑量測重建表面於估計法向量方向的誤差，提供與實際任務高度相關的精度驗證依據。
    實驗結果顯示，本研究方法相較於簡單的橫向密集相機路徑，能有效降低曲面形狀變化對三維重建誤差的影響。整體而言，本系統可於缺乏 CAD 模型或存在不可忽略的製造公差的情況下，實現非平面工件掃描重建與繪圖，並為後續開發非平面工件的機械手臂塗膠應用提供可靠的技術基礎。

    With the increasing prevalence of large-size display panels in automotive design, there is a growing demand for robotic systems capable of high-precision gluing on curved glass and workpieces. In practice, however, the absence of 3D CAD models or the presence of manufacturing tolerances necessitates the acquisition of accurate 3D geometric information before precise gluing can be performed. This study presents a comprehensive framework that integrating 3D scanning, reconstruction, and drawing functionalities for an eye-in-hand robotic system equipped with a low-cost Intel RealSense D415 RGB-D camera. The system plans camera paths based on the workpiece geometry, captures multi-view depth images, reconstructs the non-planar workpiece surface as a triangular mesh using the Truncated Signed Distance Function (TSDF) and Marching Cubes algorithms, and subsequently generates drawing paths to control the robot arm to perform drawing that simulates the gluing task.
    The main contributions of this study are threefold. First, a two-stage TSDF reconstruction and camera path planning method is proposed to overcome the time-consuming nature and lack of an effective guiding strategy in manual camera path planning. Second, the concept of the RealSense D415's acceptable range is integrated into the camera path planning and scanning–reconstruction processes to filter unreliable depth data and improve reconstruction quality. Third, a low-cost error evaluation method based on a digital indicator is introduced, which measures the reconstruction surface error along the estimated normal direction of the drawing path, providing a task-relevant accuracy validation.
    Experimental results demonstrate that the proposed method effectively mitigates the influence of surface curvature variations on 3D reconstruction errors compared with a simple dense transverse camera path. Overall, the system enables scanning, reconstruction, and drawing of non-planar workpieces in the absence of CAD models or when non-negligible manufacturing tolerances are present, and provides a reliable technical foundation for the subsequent development of robotic arm applications for gluing non-planar workpieces.

    摘要 I Abstract II 誌謝 III Table of Contents IV List of Tables VI List of Figures VII Chapter 1 Introduction 1 1.1 Motivation and Objective 1 1.2 Global Framework 2 1.3 Related Works 5 1.3.1 Robotic 3D Scanning and Reconstruction 5 1.3.2 3D Reconstruction Techniques and the TSDF Method 7 1.3.3 Robotic Drawing 8 1.4 Contributions 10 Chapter 2 System Setup and Function Specifications 11 2.1 System Setup 11 2.1.1 Coordinate Frame Definitions 11 2.1.2 Tool Coordinate Frame Calibration 12 2.1.3 Camera Calibration and Hand-Eye Calibration 14 2.1.4 Definition of World Coordinate Frame and Operational Bounding Box 15 2.2 Hardware Specifications 16 2.2.1 Intel RealSense D415 Depth Camera 16 2.2.2 Robotic Arm System 18 2.2.3 Computer Specifications 18 2.2.4 Drawing and Measurement Tools 19 2.2.5 Non-planar Workpieces 20 Chapter 3 Camera Path Planning and 3D Reconstruction 23 3.1 Camera Path for Coarse TSDF Reconstruction 23 3.2 TSDF Representation and its Open3D Implementation 28 3.2.1 From SDF to Voxel-Based TSDF Representation 28 3.2.2 Implementation of the Scalable TSDF Method in Open3D 29 3.3 Camera Path Planning and Depth Image Acquisition for Refined TSDF Reconstruction 36 3.3.1 Candidate Camera Pose Generation 38 3.3.2 Feasible Camera Pose Filtering 40 3.3.3 Camera Path Planning for Fine Scanning 41 3.4. Generating the 3D Surface from the Refined TSDF 48 Chapter 4 Drawing Path Planning 50 4.1 Drawing Path Creation 50 4.1.1 Creation and Projection of 2D Stroke Points 50 4.1.2 Tool Pose Calculation 51 4.1.3 Conversion from Stroke Coordinates to Robot Pose Parameters 52 4.2 G-code Generation 53 Chapter 5 Experimental Results 54 5.1 Data Collection 54 5.2 Matrices 56 5.2.1 Error of 3D Stroke Points Along the Predicted Normal Direction 56 5.2.2 Cloud-to-Mesh Signed Distance 58 5.3 Experimental Results and Analysis 61 5.3.1 Error of 3D Stroke Points Along the Predicted Normal Directions and Cloud-to-Mesh Signed Distance 61 5.3.2 Trend of the Digital Indicator's Depression on 3D Stroke Points 63 5.3.3 Comparison Between Proposed Method and Dense Transverse Camera Path 69 5.3.4 Analysis of the Fine Scanning Camera Path 72 5.3.5 Line Stroke Drawing Experiment on the Acrylic Curved Surface 76 Chapter 6 Conclusions and Future Works 77 6.1 Conclusions 77 6.2 Future Works 77 Chapter 7 References 79

    [1] J.-F. Chauvette, I. L. Hia, R. D. Farahani, R. Plante, N. Piccirelli, and D. Therriault. “Non-Planar Multinozzle Additive Manufacturing of Thermoset Composite Microscaffold Networks, ”Composites Part B: Engineering, 256, pp. 110627, 2023.
    [2] B. Curless, and M. Levoy. “A Volumetric Method for Building Complex Models from Range Images,” in Proceedings of the Proceedings of the 23rd annual conference on Computer graphics and interactive techniques, pp. 303-312, 1996.
    [3] R. Liu, W. Wan, K. Koyama, and K. Harada. “Robust Robotic 3-D Drawing Using Closed-Loop Planning and Online Picked Pens, ”IEEE Transactions on Robotics, 38(3), pp. 1773-1792, 2021.
    [4] W. E. Lorensen, and H. E. Cline. (1998). Marching Cubes: A High Resolution 3d Surface Construction Algorithm. In Seminal Graphics: Pioneering Efforts That Shaped the Field (pp. 347-353).
    [5] R. Malhan, R. Jomy Joseph, P. M. Bhatt, B. Shah, and S. K. Gupta. “Algorithms for Improving Speed and Accuracy of Automated Three-Dimensional Reconstruction with a Depth Camera Mounted on an Industrial Robot, ”Journal of Computing and Information Science in Engineering, 22(3), pp. 031012, 2022.
    [6] R. K. Malhan, and S. K. Gupta. “Planning Algorithms for Acquiring High Fidelity Pointclouds Using a Robot for Accurate and Fast 3d Reconstruction, ”Robotics and Computer-Integrated Manufacturing, 78, pp. 102372, 2022.
    [7] C. Mineo, D. Cerniglia, V. Ricotta, and B. Reitinger. “Autonomous 3d Geometry Reconstruction through Robot-Manipulated Optical Sensors, ”The International Journal of Advanced Manufacturing Technology, 116(5), pp. 1895-1911, 2021.
    [8] R. A. Newcombe, S. Izadi, O. Hilliges, D. Molyneaux, D. Kim, A. J. Davison, P. Kohi, J. Shotton, S. Hodges, and A. Fitzgibbon. “Kinectfusion: Real-Time Dense Surface Mapping and Tracking,” in Proceedings of the 2011 10th IEEE international symposium on mixed and augmented reality, pp. 127-136, 2011.
    [9] S. Pagano, R. Russo, and S. Savino. “A Vision Guided Robotic System for Flexible Gluing Process in the Footwear Industry, ”Robotics and Computer-Integrated Manufacturing, 65, pp. 101965, 2020.
    [10] R. Škoviera, J. K. Behrens, and K. Štěpánová. “Surfman: Generating Smooth End-Effector Trajectories on 3d Object Surfaces for Human-Demonstrated Pattern Sequence, ”IEEE Robotics and Automation Letters, 7(4), pp. 9183-9190, 2022.
    [11] D. Song, and Y. J. Kim. “Distortion-Free Robotic Surface-Drawing Using Conformal Mapping,” in Proceedings of the 2019 International Conference on Robotics and Automation (ICRA), pp. 627-633, 2019.
    [12] R. Y. Tsai, and R. K. Lenz. “A New Technique for Fully Autonomous and Efficient 3 D Robotics Hand/Eye Calibration, ”IEEE Transactions on robotics and automation, 5(3), pp. 345-358, 1989.
    [13] D. Werner, A. Al-Hamadi, and P. Werner. “Truncated Signed Distance Function: Experiments on Voxel Size,” in Proceedings of the Image Analysis and Recognition: 11th International Conference, ICIAR 2014, Vilamoura, Portugal, October 22-24, 2014, Proceedings, Part II 11, pp. 357-364, 2014.
    [14] A. M. A. Zaki, M. Carnevale, C. Schlette, and H. Giberti. “On the Use of Low-Cost 3d Stereo Depth Camera to Drive Robot Trajectories in Contact-Based Applications, ”The International Journal of Advanced Manufacturing Technology, 128(9), pp. 3745-3759, 2023.
    [15] Q.-Y. Zhou, J. Park, and V. Koltun. “Open3d: A Modern Library for 3d Data Processing, ”arXiv preprint arXiv:1801.09847, 2018.

    QR CODE