使用模板追蹤技術之即時模型式三維物體姿態估測方法與實現

Implementation of a Real-Time Model-Based 3D Object Pose Estimation Method Based on Template Tracking

施博仁(Bo-Ren Shih)

指導教授 ： 蔡奇謚

淡江大學/工學院/電機工程學系碩士班/碩士(2016年)

https://doi.org/10.6846/TKU.2016.00473

摘要

三維物體的姿態估測在電腦視覺於機器人的應用上，擔任著重要的位置。本文結合模板追蹤方法與PnP演算法完成三維物體的姿態估測方法。所使用的方法是先透過擷取並匹配物體與參考影像的特徵點，並將特徵點經過單應性矩陣轉換出物體與影像的相對關係。再利用模板追蹤方法進行追蹤。最後再將追蹤的結果利用事先建立好的模型進行點匹配，並利用PnP方法將物體的三維姿態估測出來。在實驗中進行旋轉與平移的實驗。在平移實驗上有良好的結果。在旋轉實驗上，在正負20度旋轉都有不錯的結果，此外估測出來的角度相對不穩定。

關鍵字

PnP演算法 ； 模板追蹤 ； 三維姿態估測

並列摘要

3D object pose estimation plays a crucial role in computer and robotic vision. In this thesis, a novel model-based object pose estimation algorithm is proposed by integrating template matching and PnP pose estimation methods to deal with this issue efficiently. The proposed method firstly extracts and matches keypoints of the scene image and the object reference image. Using the matched keypoints, a 2D mapping between the reference image and the observed object can be formulated by a homography matrix, which is used as an initial setting for a template-based tracking algorithm. Based on the visual tracking result, the corresponding points between image keypoints and control points of the CAD model of the object can be determined efficiently. Finally, the 3D pose of the object with respect to the camera is estimated by adopting the PnP algorithm based on the corresponding points between 2D image keypoints and 3D model control points. The experimental results validate the estimation accuracy and real-time performance of the proposed model-based object pose estimation algorithm.

並列關鍵字

PnP algorithm ； template tracking ； 3D pose estimation

參考文獻

[2] Changhyun Choi and Henrik I. Christensen, “Real-time 3D Model-based Tracking Using Edge and Keypoint Features for Robotic Manipulation,” International Conference on Robotics and Automation Anchorage Convention District IEEE, pp. 4048-4055, 2010.

[4] Samuel Dambreville, Romeil Sandhu, Anthony Yezzi, and Allen Tannenbaum, "Robust 3D Pose Estimation and Efficient 2D Region-Based Segmentation from a 3D Shape Prior," In Proc. 10th European Conference on Computer Vision, 2008.

[6] D. G. Lowe, "Distinctive image features from scale-invariant," International Journal of Computer Vision, Vols. 60,No. 2, pp. 91-110, 2004.

[7] F.Jurie, M.Dhome, Hyperplane Approximation for Template Matching, vol. 24, 2002., pp. 996-1000.

[8] M. A. F. &. R. C.Bolles, "Random Sampling Consensus: a Paradigm for Model Ftting with Application to Image Analysis and Automated Cartography," Communications of the ACM, vol. 24, pp. 381-395, June 1981.