影像基礎書畫機器手臂系統之研製

本論文選用Raspberry Pi3單板電腦以及Arduino Mega2560開發板設計一套可以透過影像處理即可直接藉由機械手臂快速書寫文字與繪畫圖案，其中透過Raspberry Pi3單板電腦負責OpenCV影像處理的工作， Mega2560開發板上負責逆向運動學推導。為了驗證此系統的可行性，我們設計了圖形繪畫與文字書寫兩個任務，首先我們先將此系統應用在繪圖功能上，透過Canny邊緣檢測算法，偵測輸入圖案的邊緣特徵點並將之繪出，藉此驗證整套自動化系統之溝通的實現以及機械手臂定位的精確性。再來，本論文為了透過此套系統來實現自動書寫功能，在輸入文字影像後使用OpenCV影像處理來提取骨架及輪廓，用做判斷毛筆的抬筆高度，接著找出起始點及鄰近點的斜率來決定筆劃順序，最後再將資訊傳給Mega2560開發板來做馬達控制。由實際實驗結果可得知本論文所提出之影像基礎書畫機器手臂有一定程度之書寫與繪畫功能。

關鍵字

繪圖系統；書寫系統； Delta機械手臂；逆運動學；邊緣測算法；大津演算法；樹莓派

並列摘要

This thesis selects Raspberry Pi3 and Arduino Mega 2560 to design a painting and calligraphy robot arm system by using the image processing through the delta robot arm. The Raspberry Pi3 is used to do the image processing and Arduino Mega 2560 is used to do the inverse kinematics and control motors. To verify the feasibility design of the system, two missions are applied. One is Painting mission and the other one is calligraphy mission. First, we apply the system on the painting mission. By using the Canny edge detection, detecting the edge of the image that we input then drawing. Through the mission we can verify the communication of whole system and accuracy of the robot arm. After that in order to achieve the calligraphy function by using the system. We input the letter image and using OpenCV to get the skeletons and contours as height of our brush then do the path-planning. Finally transmit the information to Mega 2560 to control the motor. From the real time experimental result, it shows that the proposed robot arm system can achieve favorable performance.