- 學位論文
機械手臂整合機械視覺於智能製鞋之研究
The Study on an Intelligent Robot System with Mechanical Vision for Shoe Making
摘要
鞋類製造是勞力密集型之傳統產業，一般普遍上被認為較難以實現自動化，然而在勞力成本持續上漲同時，製鞋廠商須追求高品質與嚴格控管並降低製造成本為目標，促使製鞋廠商必須投資並導入生產自動化及設備控制智能化之製程，減少產線上人工需求，提升整體生產效益，因此本研究基於製鞋製能自動化研究項目包含:(1)成型前段之前幫智能系統開發、(2)智能製鞋打粗塗膠機械手臂系統開發。 成型前段之前幫智能系統開發，本研究主要解決傳統人工前幫幫鞋入料之技術，此傳統技術需要訓練有術的幫鞋師傅才能夠準確地將入楦的鞋楦正確入料完成幫鞋動作，以此人力技術是難以取代的，本研究開發一專家視覺系統，以標準入料拍攝之視覺影像為基礎，利用影像特徵比對機械手臂入料的影像，再透過前幫智能系統辨識GO、NG或可調。 打粗塗膠程序是鞋面與鞋底貼合的重要程序，然而由於鞋樣設計特別是運動鞋十分複雜，打粗塗膠製程往往只能依靠手工進行。鞋業近年來已開始走向3D的數位開發，不像以往皆只在2D平面作業，因此很多的鞋型在進入量產前便已有鞋面鞋底的3D模型 。本研究利用鞋型的3D模型，在Rhino 3D之CAD平台上開發插件程式，使得複雜鞋型之打粗塗膠路徑可在插件程式中進行規畫，並在製程中結合實際鞋面的三維視覺掃描結果，修正CAD規畫之路徑後轉換成機械手臂的路徑格式，驅動機械手臂進行複雜鞋型打粗塗膠製程，完成一以3D CAD為基礎並整合三維視覺之智慧型打粗及塗膠機械手臂系統。
並列摘要
Shoe manufacturing is a labor-intensive traditional industry and generally considered more difficult to be automated. But labor cost continues rising in years and footwear manufacturers have still to keep products high-quality and reduce cost continuously. Therefore, the footwear manufacturers must invest and introduce automation equipments and intelligent processes to reduce artificial demand in production line and increase the overall efficiency of production. This research is based on the automation projects of shoe-making such as (1) the development of expert system for shoe lasting process, and (2) the development of upper rounghing & gluing system with intelligent robotic arm respectively. In the expert system for shoe lasting process, this paper aims to solve the traditional artificial loading into a lasting machine. The traditional loading needs a well-trained master to adjust a shoe properly and put it accurately into machine. It is difficult to be replaced by automation. In the study, a visual expert system is developed. Based on visual images of a standard & correct sample of loading, features from images of shoe with a robotic arm loading are compared. Then the judgement such as GO, NOGO and ADJUSTABLE is done by the expert system according to comparison. Roughing & gluing are important procedures in the shoe assembly process. These procedures often rely on manual because design of shoes is very complex especially for sneakers. Unlike there were only 2D patterns in the past, shoe industry has begun to use 3D system to create shoe modeles in recent years. In the study, these 3D models of uppers and soles are appled to roughing & gluing automation. Plug-in programs in Rhino 3D are developed and roughing & gluing paths can be pre-planned in the programs. Then real geometry of a shoe upper is captured by a 3D scanner, and the pre-planned paths are modified according to the captured data. And paths are converted and input into to a robotic arm. Finally, the robotic arm with 3D CAD based pre-plannig system and 3D scanner can run complex paths automatically for shoe roughing & gluing process.