Paper Delivery System for Calligraphy-Writing Robot

奕澂賴
2021年1月12日
讀畢需時 2 分鐘

Extend Taiwan Ministry of Science and Technology (MOST) Project

Advisor: Prof. Pei-Chun Lin

Besides the Ink-Grinding Robot, my team continued to set up a paper feeding system. After finishing one part of the writing, the system helps to deliver the rice paper (a kind of paper used for calligraphy writing) on the table to the next position for writing. The reason is that the workspace of the arm couldn't cover the entire paper. The system can solve that problem and make the calligraphy-writing process fully automatic.

We planned to apply the way of computer vision to fulfill the task. Firstly, we carried out camera calibration for the webcam equipped on the arm and constructed a LabVIEW system for arm motion control. The arm would automatically move the paper to the next position with its end-effector every time it finished a part of writing.

However, it won't always move the paper to the exact spot accurately. A few angle errors appeared after delivery. Thus, I implemented a SURF algorithm into the system with OpenCV to detect the error angle. The pictures of the paper at the ideal spot were input into the system in advance. After delivery, the webcam on the arm would take a picture of the paper position right away and compare it to the one at the ideal position. The function could match the features on the calligraphy characters and capture the homography matrix. The error angle could then be known from some calculation of the entries in the homography matrix. The research result was then submitted as an abstract and accepted by the 2020 International Conference on Advanced Robotics and Intelligent Systems.