大專生計畫（科技部）· 2022–2023NSTC Undergraduate Grant · 2022–2023
深度影像機械手臂人機互動介面Depth-Vision Robotic-Arm HMI
結合六軸機械手臂與深度影像辨識，使用者只要在螢幕點選物體，即可將其移至指定位置。A six-axis arm with depth-vision recognition: tap an object on screen and the arm moves it to a target.
概述Overview
開發結合六軸機械手臂與深度影像辨識（搭配深度學習）的人機互動介面。已完成手臂通訊、正逆向運動學、像素轉實際長度與物體辨識，並能以純深度影像辨識並抓取物體。設計理念是讓「取什麼、放什麼」由人決定——只要在螢幕點一下物體，手臂即可將其移至指定位置；後續朝物體堆疊、非平面與圓錐形等複雜情境擴展。 A human–machine interface combining a six-axis arm with depth-vision recognition (plus deep learning). It implements arm communication, forward/inverse kinematics, pixel-to-real-length mapping, and object recognition, and can grasp objects from depth images alone. The design lets a human decide what to pick and place — tap an object on screen and the arm moves it — with future work targeting stacked, non-planar, and conical objects.
重點Highlights
- 人機介面：螢幕點選物體 → 自動移至指定位置。HMI: tap an object on screen → it is moved to the target automatically.
- 純深度影像辨識並抓取物體。Grasping from depth images only.
- 手臂通訊與正／逆向運動學。Arm communication and forward/inverse kinematics.
- 像素 → 實際長度座標校正。Pixel-to-real-length coordinate calibration.
- 深度學習物體辨識；朝物體堆疊／非平面情境擴展。Deep-learning object recognition; extending toward stacked / non-planar scenes.