在現今變化快速的3C產品市場中,整個供應鏈被要求快速地設計產品與生產製造商品,但是在螺絲、螺帽等傳統產業中,還是跳脫不了使用人力來從事品質檢測工作,使得產品的良率與人力成本的支出不一定能成正比。為了有效解決異型螺帽之內紋檢測問題,本研究提出一個架構,運用無背隙 (Zero Backlash) 減速機的高精度六軸機械手臂,結合二值化影像處理與Otsu(大津展之)統計式門檻值(Threshold)決定法,進行螺帽內紋檢測的系統。並且透過實驗,驗證此架構的可行性。本研究的目的,在於運用高精度六軸機械手臂搭配機器視覺,提升產業界中異形螺帽、盲孔螺帽的檢測技術,若能藉由此架構來替代現有的人工視覺瑕疵檢測方式,即可改善人工檢測的種種缺失,大幅減少過去抽樣檢測或人工目視的誤判,進而提升檢測良率與確保產品品質的一致性。
In today's fast-changing market, 3C products, the entire supply chain is required to quickly design products and manufacturing goods, but in the screws, nuts and other traditional industries, can not escape or to engage in the use of human quality inspection work, making the product yield and labor costs expenses necessarily proportional. In order to effectively solve shaped pattern detection problem within the nut, this study proposes a framework for the use of backlash (Zero Backlash) reducer precision six-axis robotic arm, combined with the binary image processing and Otsu (Otsu exhibition of) Statistics thresholding (Threshold) decided to France for the nut pattern detection system. And through the experiment to test the feasibility of this architecture. The purpose of this study is to use high-precision six-axis robotic arm with machine vision, enhance industrial circles shaped nut, nut blind hole detection technology, if by this framework to replace the existing artificial vision defect detection methods, namely, can improve the detection of various artificial missing, substantially reduced over the last sampling or artificial visual misjudgment, and thus enhance the detection yield and ensure product quality and consistency.