改良式立體視覺演算法應用於機械手臂之三維影像識別研究

隨著科技日新月異，電腦立體視覺的應用已廣泛使用於日常生活中，如防盜監視系統、工廠自動化組裝生產線及自動化檢測設備等，皆屬於無人自動化應用中不可或缺的一部分。電腦立體視覺系統可代替人類從事一些繁複且需非常專注而消耗大量體力和眼力的工作，並提升其正確性、精確度及效率的功能。本研究使用改良式立體視覺演算法，搭配具有雙CCD立體視覺系統及末端夾爪功能之5軸機器手臂，利用影像處裡、立體視覺系統對環境進行偵測及影像識別，藉此引導機械手臂的動作，在真實環境可識別特定顏色物體或偵測移動目標物，並由末端夾爪精準的抓取空間中的目標物。研究包含三個重要部分，第一部份為具備類似人類的視覺系統，經由雙CCD及一顆步進馬達所組成，來模擬人類的頭部及雙眼運動，對空間中的三維影像進行識別及偵測，並搭配改良式立體視覺演算法來並計算出其目標物與機器手臂的三維相對位置。第二部份為PC-based控制系統，包含了多軸運動控制卡及影像擷取卡。第三部份為機械手臂運動控制系統，機械手的第一軸採中空旋轉平台設計，仿人類的腰部功能，負責調整機器手臂的工作面向。而手臂部份則使用四顆步進馬達設計成右手型式，藉由機械人運動學及機器手軌跡路徑規劃模組，轉換而得知機械手臂各軸馬達所需轉動的角度，使之抓取工作範圍內目標物並移動至指定目標位置。

關鍵字

影像識別；機械手臂；電腦立體視覺；改良式立體視覺演算法

並列摘要

As the rapid change of science and technology, computer stereo vision can be applied extensively in daily life. For example like security system, industrial automatic assembly line and automatic visual inspection. Computer stereo vision is indispensable for unmanned automatic application. The impact of using computer stereo vision system improved the accuracy and efficiency of production. It can competence what complex and eyesight consuming jobs than human. In this study is using modified stereo vision algorithm on a robot system which contain dual CCD camera stereo vision system、5 degrees of freedom and an end-effector attached to the end of the robot arm. The system is using image processing and stereo visual system to detect and recognize the environment. It also could recognition an object’s colors or detection moving object, and then to guide the end-effector of the robot arm to gripping the object accurately under the real environment. This study, it consists of three major parts. The first is the vision system, which composed of two cameras and a stepper motor to simulate human’s vision system. The system computing coordinates between object and robot arm by the modified stereo vision algorithm. The second is PC-based control system, including the image capture card and multi-axis motion control card. The third is the motion control system, there is a hollow rotary stage that control the direction. And the other is right-hand function composed of four stepper motors. This system adopts robotic kinematics and path planning module to compute how many angles of each motor need to move. Finally, the end-effector gripped the object and move to the appointed destination.

並列關鍵字

Computer Stereo Vision ； Modified Stereo Vision Algorithm ； image recognition ； robot arm

參考文獻

[11] 謝奇均，具遠端控制服務型移動機器人之研究，中原大學機械工程學系碩士學位論文，2011

[8] 林駿青，立體視覺擲球機械手臂之研究，碩士論文，中原大學機械學系，2012。

[10] K. Y. Chen, C. C. Chien, W. L. Chang, and C. C. Hsieh, “Improving the accuracy of depth estimation using a modified stereo vision model in binocular vision,” The 10th International Symposium on Measurement Technology and Intelligent Instruments, Daejeon, Korea, Jun 29 - Jul. 2, 2011.

[2] K. Y. Chen, Y. P. Chang, and B. Fu, ”A stereo vision-guided mobile robot for real-time image tracking system,”(in Chinese) Chung Yuan Journal, Vol. 31, No. 1, pp. 21-29, 2003.

[3] K. Y. Chen, C. C. Chen, and C.T. Tseng, ”Improving the accuracy of depth estimation in binocular vision for robotic application,” Applied Mechanics and Materials, Vols. 284-287, pp 1862-1866, 2013.

國際替代計量

改良式立體視覺演算法應用於機械手臂之三維影像識別研究

全文下載

主題瀏覽