本文選題：機(jī)器視覺(jué)　切入點(diǎn)：數(shù)學(xué)圖像處理　出處：《機(jī)械科學(xué)研究總院》2016年碩士論文

【摘要】：在工業(yè)智能化和高科技化的當(dāng)代社會(huì),汽車(chē)作為代步工具的重要性越來(lái)越明顯。目前在我國(guó),汽車(chē)廠商越來(lái)越重視生產(chǎn)質(zhì)量、生產(chǎn)成本、生產(chǎn)效率,相關(guān)部門(mén)也出臺(tái)了各種有利措施,來(lái)推進(jìn)汽車(chē)產(chǎn)業(yè)轉(zhuǎn)型升級(jí),向高效、高質(zhì)的目標(biāo)發(fā)展。在這些政策、技術(shù)發(fā)展革新的大背景下,汽車(chē)廠商對(duì)發(fā)動(dòng)機(jī)裝配質(zhì)量的要求也越來(lái)越高,本論文研究開(kāi)發(fā)的發(fā)動(dòng)機(jī)滾子搖臂機(jī)器視覺(jué)檢測(cè)系統(tǒng)就是為了提高生產(chǎn)質(zhì)量和效率,應(yīng)運(yùn)而生的。發(fā)動(dòng)機(jī)滾子搖臂是汽車(chē)發(fā)動(dòng)機(jī)的重要零件,它隸屬于發(fā)動(dòng)機(jī)中的氣門(mén)機(jī)構(gòu),作用是把發(fā)動(dòng)機(jī)中凸輪軸傳來(lái)的力和位移傳遞到氣門(mén)上,完成推開(kāi)或關(guān)閉氣門(mén)的任務(wù)。由于非常重要,其裝配質(zhì)量必須經(jīng)過(guò)檢測(cè)。傳統(tǒng)的檢測(cè)發(fā)動(dòng)機(jī)滾子搖臂裝配質(zhì)量的方法是,人眼識(shí)別是否合格,容易產(chǎn)生誤判。因此,需要研發(fā)一套機(jī)器視覺(jué)系統(tǒng)來(lái)自動(dòng)識(shí)別檢測(cè)滾子搖臂位置是否放偏、漏裝及未完全到位。本論文主要的研究?jī)?nèi)容有以下幾點(diǎn):1、按照設(shè)計(jì)目標(biāo)和任務(wù)進(jìn)行硬件選型,包含光源、工業(yè)相機(jī)、與工業(yè)相機(jī)配合使用的鏡頭、圖像采集卡、工業(yè)PC等;2、選擇合適的機(jī)器視覺(jué)軟件,建立恰當(dāng)?shù)乃惴?用來(lái)處理采集到的圖片及數(shù)據(jù),確保設(shè)備軟件與設(shè)備硬件配套兼容,從而很好的檢測(cè)出發(fā)動(dòng)機(jī)滾子搖臂在裝配過(guò)程中是否放偏、是否缺失以及是否未完全到位,進(jìn)而判斷出是否符合裝配質(zhì)量要求;3、進(jìn)行在線測(cè)試,對(duì)機(jī)器視覺(jué)檢測(cè)系統(tǒng)中的算法和各個(gè)參數(shù)進(jìn)行優(yōu)化;4、實(shí)現(xiàn)整體視覺(jué)檢測(cè)系統(tǒng)中PLC以及PC通訊。確保系統(tǒng)相機(jī)能得到PLC給的觸發(fā)信號(hào)并且檢測(cè)完成的OK/NG信號(hào)能夠傳輸?shù)絇LC,確保發(fā)動(dòng)機(jī)滾子搖臂的檢測(cè)數(shù)據(jù)能無(wú)誤差的傳送到PC進(jìn)行保存。本論文的新穎之處是利用機(jī)器視覺(jué)相關(guān)技術(shù),在發(fā)動(dòng)機(jī)滾子搖臂圖像的采集及分析、滾子搖臂放偏、缺失或未完全到位的識(shí)別、檢測(cè)數(shù)據(jù)和檢測(cè)圖片的處理等方面進(jìn)行創(chuàng)新研究。目前該系統(tǒng)已經(jīng)成功應(yīng)用在一汽大眾EA211發(fā)動(dòng)機(jī)裝配生產(chǎn)線中,系統(tǒng)檢測(cè)準(zhǔn)確率可達(dá)到98%以上,處理時(shí)間為2秒/臺(tái),提高了發(fā)動(dòng)機(jī)滾子搖臂裝配檢測(cè)的效率,保證了發(fā)動(dòng)機(jī)產(chǎn)品裝配質(zhì)量。
[Abstract]:In the modern society of industrial intelligence and high technology, the importance of automobile as a walking tool is becoming more and more obvious. At present, automobile manufacturers pay more and more attention to production quality, production cost and production efficiency. The relevant departments have also issued various favorable measures to promote the transformation and upgrading of the automobile industry to the goal of high efficiency and high quality. In the context of these policies, technological development and innovation, Automobile manufacturers have higher and higher requirements for engine assembly quality. The purpose of this paper is to improve the production quality and efficiency of the engine roller rocker arm machine vision inspection system. The engine roller rocker arm is an important part of the automobile engine. It belongs to the valve mechanism in the engine, and its function is to transfer the force and displacement from the camshaft in the engine to the valve. Complete the task of pushing or closing the valve. Because of its very important, its assembly quality must be tested. The traditional method of testing the assembly quality of the engine roller rocker arm is that the human eye is able to recognize whether or not it is up to standard, which is prone to misjudgment. It is necessary to develop a machine vision system to automatically identify and detect whether the position of the roller rocker arm is biased, missing or not in place. The main research contents of this paper are as follows: 1, hardware selection according to the design objectives and tasks, including light source. Industrial camera, lens used in conjunction with industrial camera, image acquisition card, industrial PC, etc., select the appropriate machine vision software, set up the appropriate algorithm to process the collected pictures and data. Make sure the equipment software is compatible with the hardware of the equipment, so that the engine roller rocker arm in the assembly process is deflected, missing and not fully in place, And then to determine whether or not to meet the assembly quality requirements, to conduct on-line testing, The algorithms and parameters in the machine vision detection system are optimized to realize the communication between PLC and PC in the whole vision detection system. The system camera can get the trigger signal from PLC and detect the completed OK/NG signal. Input to PLC to ensure that the detection data of the engine roller rocker arm can be transmitted to PC without error. The novelty of this paper is the use of machine vision related technology. In the engine roller rocker arm image collection and analysis, roller rocker arm deviation, missing or not fully in place recognition, At present, the system has been successfully applied in FAW Volkswagen EA211 engine assembly line. The detection accuracy of the system can reach more than 98%, and the processing time is 2 seconds per unit. The efficiency of assembly detection of engine roller rocker arm is improved, and the assembly quality of engine product is ensured.
【學(xué)位授予單位】：機(jī)械科學(xué)研究總院
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：U464.13;TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前2條

1 王艷華;管一弘;;基于模糊集理論的醫(yī)學(xué)圖像分割的應(yīng)用[J];計(jì)算機(jī)技術(shù)與發(fā)展;2008年11期

2 韓峰;李仁發(fā);曾慶光;劉彥;;圖像與視頻提取技術(shù)綜述[J];中國(guó)科技論文在線;2007年10期

，

本文編號(hào)：1682229