本文選題：汽車組合儀表 + 視覺(jué)檢測(cè)　； 參考：《湖北工業(yè)大學(xué)》2017年碩士論文

【摘要】：隨著汽車工業(yè)和電子信息技術(shù)的快速發(fā)展,汽車組合儀表的信息化程度越來(lái)越高,質(zhì)量要求日趨嚴(yán)格。因此,汽車組合儀表在出廠前要經(jīng)過(guò)嚴(yán)格的質(zhì)量檢測(cè),但人工檢測(cè)的效率和精度低,不適合自動(dòng)化生產(chǎn)。針對(duì)目前人工檢測(cè)中效率低和精度差的問(wèn)題,本文利用了機(jī)器視覺(jué)領(lǐng)域相關(guān)的研究成果,設(shè)計(jì)了一套汽車儀表視覺(jué)檢測(cè)系統(tǒng)。在深入研究信號(hào)驅(qū)動(dòng)單元硬件設(shè)計(jì)方案和汽車儀表液晶符號(hào)的定位與識(shí)別算法后,本文完成了以下方面的工作。(1)以STM32為主控制器,采用集成式的設(shè)計(jì)方式,設(shè)計(jì)了汽車儀表信號(hào)驅(qū)動(dòng)單元硬件電路和相應(yīng)的軟件。每種類型的信號(hào)驅(qū)動(dòng)電路互相獨(dú)立,并由主控制器控制其產(chǎn)生輸出信號(hào)。主控制器首先與上位機(jī)建立通信連接,當(dāng)接收到上位機(jī)發(fā)送的指令后,主控制器根據(jù)指令控制相應(yīng)的驅(qū)動(dòng)電路產(chǎn)生信號(hào),這種集成式的設(shè)計(jì)方式大大提高了系統(tǒng)的穩(wěn)定性和精度。(2)針對(duì)液晶顯示區(qū)域的定位問(wèn)題,根據(jù)液晶顯示區(qū)域的特征,利用一種改進(jìn)的ACSS角點(diǎn)檢測(cè)算法,對(duì)區(qū)域進(jìn)行粗略定位,再通過(guò)邊緣測(cè)量算子進(jìn)行精確定位。改進(jìn)的ACSS算法增加了對(duì)二值邊緣的篩選步驟,篩選后二值邊緣的數(shù)量減少,去掉了大量的偽角點(diǎn),計(jì)算量變小,與目前主流的角點(diǎn)檢測(cè)算法相比,實(shí)驗(yàn)結(jié)果表明改進(jìn)的ACSS算法檢測(cè)速度和精度都有一定提高。(3)針對(duì)液晶符號(hào)識(shí)別問(wèn)題,運(yùn)用BP神經(jīng)網(wǎng)絡(luò)算法對(duì)液晶字符進(jìn)行識(shí)別。首先對(duì)液晶字符進(jìn)行預(yù)處理,濾除噪聲干擾,提取液晶符號(hào)的二值圖像,接著在垂直方向上使用投影法進(jìn)行字符分割,最后使用訓(xùn)練好的BP網(wǎng)絡(luò)對(duì)單個(gè)符號(hào)進(jìn)行識(shí)別。實(shí)驗(yàn)結(jié)果表明BP網(wǎng)絡(luò)的字符識(shí)別率高,實(shí)時(shí)性強(qiáng)。另外,本文還對(duì)其他液晶符號(hào)、LED指示燈和小表指針角度的識(shí)別算法進(jìn)行了分析和實(shí)現(xiàn),最終的識(shí)別結(jié)果達(dá)到了設(shè)計(jì)要求。通過(guò)以上工作,本文實(shí)現(xiàn)了汽車儀表視覺(jué)檢測(cè)系統(tǒng),檢測(cè)系統(tǒng)能實(shí)時(shí)完成汽車儀表的全部功能檢測(cè),系統(tǒng)的拓展功能強(qiáng),兼容性好,并在汽車儀表生產(chǎn)流水線上進(jìn)行了應(yīng)用,達(dá)到了預(yù)期的效果
[Abstract]:With the rapid development of automobile industry and electronic information technology, the information level of automobile combination instrument is becoming more and more high, and the quality requirement is becoming more and more strict. Therefore, the automobile combination instrument must pass through the strict quality inspection before leaving the factory, but the manual inspection efficiency and the precision are low, is not suitable for the automation production. Aiming at the problems of low efficiency and poor precision in manual inspection, a visual inspection system of automobile instrument is designed by using the related research results in the field of machine vision. After deeply studying the hardware design of the signal drive unit and the localization and recognition algorithm of the liquid crystal symbol of the automobile instrument, the following work has been accomplished in this paper. (1) the STM32 is used as the main controller and the integrated design method is adopted. The hardware circuit and corresponding software of vehicle instrument signal drive unit are designed. Each type of signal drive circuit is independent of each other and is controlled by the master controller to generate the output signal. The main controller first establishes a communication connection with the host computer. After receiving the instruction sent by the host computer, the master controller controls the corresponding driving circuit to generate the signal according to the instruction. This integrated design method greatly improves the stability and accuracy of the system. (2) aiming at the location of liquid crystal display region, according to the characteristics of liquid crystal display region, an improved ACSS corner detection algorithm is used to roughly locate the region. Then accurate location is carried out by edge measurement operator. The improved ACSS algorithm increases the screening steps of binary edges, reduces the number of binary edges, removes a large number of pseudo-corner points, and reduces the amount of calculation, compared with the current mainstream corner detection algorithm. The experimental results show that the speed and accuracy of the improved ACSS algorithm are improved. (3) to solve the problem of liquid crystal symbol recognition, BP neural network algorithm is used to recognize liquid crystal characters. First, the liquid crystal character is preprocessed, noise interference is filtered, the binary image of liquid crystal symbol is extracted, then the character is segmented by projection method in the vertical direction, and the single symbol is recognized by the trained BP neural network. The experimental results show that BP neural network has high character recognition rate and high real-time performance. In addition, this paper also analyzes and implements the recognition algorithm of LED indicator lamp and small meter pointer angle of other liquid crystal symbols, and the final recognition results meet the design requirements. Through the above work, this paper has realized the automobile instrument vision inspection system, the detection system can complete the automobile instrument all function inspection in real time, the system expansion function is strong, the compatibility is good, and has carried on the application in the automobile instrument production line. Achieved the desired effect
【學(xué)位授予單位】：湖北工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 倫向敏;侯一民;;運(yùn)用迭代最大熵算法選取最佳圖像分割閾值[J];計(jì)算機(jī)工程與設(shè)計(jì);2015年05期

2 王競(jìng)雪;宋偉東;趙麗科;王偉璽;;改進(jìn)的Freeman鏈碼在邊緣跟蹤及直線提取中的應(yīng)用研究[J];信號(hào)處理;2014年04期

3 許松南;;基于STM32和CAN總線的雙通道正弦信號(hào)發(fā)生器設(shè)計(jì)[J];計(jì)算機(jī)測(cè)量與控制;2014年02期

4 吳曉軍;鄒廣華;;基于邊緣幾何特征的高性能模板匹配算法[J];儀器儀表學(xué)報(bào);2013年07期

5 劉雄偉;王曉;;與平面曲線方向相關(guān)的幾個(gè)注記[J];高等數(shù)學(xué)研究;2013年02期

6 張冀;王俊宏;尉遲明;丁明躍;侯文廣;;基于計(jì)算機(jī)視覺(jué)的汽車儀表指針檢測(cè)方法[J];計(jì)算機(jī)工程與科學(xué);2013年03期

7 申森;李艾華;姚良;馬治明;;基于小波包和Niblack法的槍號(hào)圖像二值化算法[J];光子學(xué)報(bào);2013年03期

8 劉松濤;殷福亮;;基于圖割的圖像分割方法及其新進(jìn)展[J];自動(dòng)化學(xué)報(bào);2012年06期

9 冉令峰;;基于垂直投影的車牌字符分割方法[J];通信技術(shù);2012年04期

10 鄧婕;;CAN總線通信原理分析[J];電子設(shè)計(jì)工程;2012年07期

相關(guān)碩士學(xué)位論文 前2條

1 孟祥雪;基于機(jī)器視覺(jué)的汽車儀表讀數(shù)檢測(cè)技術(shù)研究[D];哈爾濱工業(yè)大學(xué);2012年

2 張燕;基于機(jī)器視覺(jué)的藥用玻璃瓶缺陷檢測(cè)的研究[D];合肥工業(yè)大學(xué);2008年

，

本文編號(hào)：2061395