端銑刀刃口平整度之影像檢測系統開發

Development of Inspection System for Cutting Edge Flatness on End-mill by Image Processing

林琮祐(Cong-You Lin)

指導教授 ： 李炳寅

共同指導教授 ： 陳進益(Jenn-Yih Chen)

國立虎尾科技大學/工程學院/機械與電腦輔助工程系碩士班/碩士(2018年)

摘要

隨著科技水平的提高，工業生產過程中的自動化要求不斷增加，生產精度的要求也不斷提高，故需要利用自動化與測量精度高的非接觸測量技術來取代傳統人工且測量精度低的測量方法，因為使用自動化檢測，有著快速與準確的優點，傳統的人工檢測，常常會受到檢測人員視力、體力等先天體能條件的限制，而使得檢測的品質及效率降低，且可靠度也不及使用自動化檢測。在工業檢測上的應用，使用機器視覺的技術已經非常普遍，其中以尺寸量測、瑕疵檢測及動態攝影等為主，並且不會破壞待測物的特性。
本論文是以X、Y、Z、C四軸PC-based控制平台與使用C++程式語言，研發一套刃口平整度影像檢測系統，搭配控制器擷取機械座標系統，並且實現量測與對焦自動化等人性化功能，整合成為一個機器視覺檢測平台。光源部分使用同軸光源與自行設計蜂巢式分區環形光源，並搭配控制器PLC控制環形光源的光源照射區域，以解決採光不足的問題；利用CCD焦距特性與對焦原理並使用Sobel運算子來計算梯度值擷取機台機械座標；人機介面以教導式圖片視窗介面提示使用者操作流程步驟，量測功能可量測刀具外徑、螺旋角、刃口平整度等尺寸檢測。由實驗結果顯示與德國Zoller genius 3五軸自動刀具量測儀和日本Olympus高精度工具顯微鏡之量測數據相比之下，本文的量測數據誤差皆在2.2%之內，故本檢測系統的精度在合理範圍並可應用於檢測不同的刀具幾何。

關鍵字

機器視覺 ； 刃口平整度 ； 自動對焦 ； 索柏運算子

並列摘要

Due to the improvement of modern scientific and technological level, the automation and production accuracy requirements are increasing in the industrial production process. The modern industrial production needs to use the automation and high measurement accuracy of non-contact measurement technology to replace the traditional, artificial, and low measurement accuracy of the measurement methods. By using the automated inspection, it has advantages of fast and accuracy, the traditional artificial inspection is limited by the detectors vision, physical and other congenital physical conditions. Thus, the quality and efficiency of the inspection are reduced and the reliability in the area of automated inspection is not good enough to be trusted. The machine vision inspection technology is widely applied in dimension measurement, flaw inspection, dynamic photography, etc. The major advantage of the vision inspection is it won’t damage devices under test during the measuring process.
This thesis presented an inspection system for the flatness of cutting edges by image processing. The system is developed in C++ environment and built an X, Y, Z, and C four-axis PC-based control platform. We used the controller to obtain the machine coordinate and implemented the humanized functions of automatic and autofocus measurements, integrated as a machine vision inspection platform. The co-axis light source and an LED ring light with different angles were applied to solve the problem of insufficient lighting. The focusing characteristics and principles of CCD, and Sobel operator were utilized for calculating the gradient value to get the machine program coordinate. The human machine interface has a tutoring manner for users to operate the measuring procedures. The measurement functions can measure the diameter, helix angle, flatness of cutting edge, and other dimension measurements. Experimental results show that the measurement error is within 2.2% in comparison with the measurement instrument Zoller genius 3 and Olympus high-precision tool microscope. Therefore, the accuracy is reasonable for the proposed system which can be applied to measure the geometry of different cutting tools.

並列關鍵字

machine vision ； flatness of cutting edge ； autofocus ； Sobel operator

參考文獻

[14] 林祺勛，2011，“五軸工具磨床線上刀具幾何影像檢測系統開發”，國立虎尾科技大學，創意工程與精密科技研究所，碩士論文。

[15] 盧鈜俞，2013，“端銑刀之模組式檢測系統開發”，國立虎尾科技大學，機械與電腦輔助工程系研究所，碩士論文。

[17] 吳易濃，2013，“端銑刀之刃口研磨精度最佳化設計”，國立虎尾科技大學，機械與電腦輔助工程系研究所，碩士論文。

[19] 涂鈞凱，2010，“應用機器視覺於連接器光學檢測系統之開發”，國立台北科技大學，機電整合研究所，碩士論文。

[5] 蔡博智，2002，“影像追蹤方法應用在監控系統之研究”，中原大學，機械工程學系，碩士論文。