手繪寫生視圖影像處理系統

Abstract

機械設計師通常是以紙和筆的交叉運用，以寫生視圖的表達方式繪出外形 ，但是目前電腦輔助設計系統是以交談式輸入方式，使設計活動過冗長繁 雜，所以本研究發展一機械零件手繪草圖處理系統，結合CAD及影像處理 技術，將零件之手繪寫生視圖，以掃描機直接輸入電腦，即能在電腦中重 新構建零件之精確外觀設計圖形。本研究所發展之系統共有五項模組:(1) 輸入模組，(2)細線化模組， (3)分群模組，(4)求出各群幾何基元表達法 模組，(5)圖形修飾模組。系統之進行是先將圖形影像做細線化處理，然 後找出零件圖形之端點及連接點，以端點和連接點為分割點，將圖形分割 成直線、橢圓弧及橢圓等幾何基元，並求出各幾何基元之表達方程式，當 所有的幾何基元均以方程式表達後，即計算兩幾何基元方程式之新交點。 最後，即可重繪原設計圖形。 Freehand drawings of pictorial views are generally used by designers in describing their design intentions. The freehand drwing input scheme is user friendly; however it has to be transformed into a precise sketch for further storage and retrival. This research develops a system for automatically transforming a freehand drawing of pictorial views into a precise sketch. The system involves five procedures: (1) inputting the freehand drawing by a scanner, (2) applying a thinning algorithm to characterize the skeleton of the drawing, (3) grouping the pixels, (4) determing geometric primitives for each group of pixels, (5) embellishing the geometric primitives. The system has been implemented in C language on a PC486 machine with a scanner as one of its input device.