Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2321
標題: 機器視覺於加工邊緣毛邊量測之應用
Application of machine vision in edge burr measurement
作者: 蔡曜仲
Tsai, Yao-Chung
關鍵字: machine vision;機器視覺;burr;measurement;毛邊;量測
出版社: 機械工程學系
摘要: 
大部分的零件在加工製造過程中,皆或多或少於加工面附近殘留一些未移除或多出的毛邊。它們的存在將損害產品的性能及外觀,需要在流入後方生產線前予以檢知,並加以處理,否則將對產品的品質造成不可忽視的影響,為此需要對毛邊建立一套檢測的方式。
過去利用肉眼進行毛邊量測,雖然簡單,但卻有人為誤差的問題存在。針對於此,本研究以銑削加工為對象,建立一套機器視覺量測方法。以比對加工前後二維的影像,根據邊緣輪廓變化,並搭配本文之毛邊偵測方式,找尋影像中毛邊最高點的位置至次像素的精度。再利用校正所得的函數關係式,將由影像中的高度值換算為實際的高度。
經實驗測試,機器視覺量測系統的誤差在正負0.01mm以下,唯須注意毛邊翻覆後對誤差的影響。

During the manufacturing process , the cutting surface of most components would remain burr . This would affect the external surface and functionality of the product . Producers should identify this defect and strive to maintain quality , so we should establish a method to inspect burr .
Traditionally , the measurement tool of burr is our naked eyes . The disadvantage of this method is unnoticeable error caused by human . In order to avoid this drawback , we can take advantage of the machine vision system . First , compare the contours of the two 2D images , which are taken before and after the workpiece has been machined. Second , find the highest location of burr in precision of sub-pixel with our approach presented in this paper . This could be done according to the difference between contours . Finally, transfer the value of sub-pixel to the real height of burr by making use of the mapping function established in the calibration process .
When the burr had not rolled seriously , the error of machine vision was between -0.01mm and 0.01mm in our experiment .
URI: http://hdl.handle.net/11455/2321
Appears in Collections:機械工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.