基于机器视觉的汽车组合仪表检测系统的设计与实现

发布时间：2018-05-12 13:05

本文选题：机器视觉 + 仪表　；参考：《湖北工业大学》2017年硕士论文

【摘要】：本文基于机器视觉技术的研究,针对汽车组合仪表中需要被检测的繁杂信息,包括指针、灯和字符等,设计并实现了一整套检测系统来对其进行智能检测。相比于传统的人工目检,不仅降低了检测误差,而且还提高了生产效率。主要工作如下:1.通过研究汽车仪表自身特点,结合实际的检测过程,提出了整个仪表检测系统的设计方案。硬件方面,根据实际的检测需求,采用集成式的方式进行仪表驱动设计,并选择合适的图像采集装置;软件方面,由检测的内容,设计了汽车组合仪表整个检测的流程。2.针对用传统的图像减影法提取指针存在的不足,提出了一种基于高精度模板匹配的减影法,为有效的获取指针图像提供了保障。然后分析各种直线提取的方法,根据仪表自身特点,提出了结合用最小二乘法来拟合椭圆的方法,提取椭圆长轴作为直线,测量该直线与水平线之间的夹角,再由指针中心与旋转中心的相对位置,确定指针的实际偏转角度。3.研究了仪表LED灯的检测方法,通过灰度值确定其亮度,转换到HSV色域空间内测量色调值确定其颜色,结合集合的思想确定其是否存在连锡的情况。研究了段式灯的检测方法,通过拟合出段式灯的最小外接矩形进行来段式灯的检测。研究了字符的检测方法,通过字符预处理、字符分割和支持向量机的方法识别液晶屏上的数字。4.由提出的硬件、软件的设计方案和图像处理算法完成对整个检测系统的实现。操作检测系统,对东风汽车康明斯组合仪表进行检测实验。检测实验结果表明,本文所设计实现的基于机器视觉的汽车组合仪表检测系统能够完成对仪表较为全面的检测,在时间消耗较少的情况下能够保证各个检测项目的精准检测,满足实际的工业检测要求,具有一定的发展前景和实用价值。
[Abstract]:Based on the research of machine vision technology, this paper designs and implements a set of detection system to detect the complex information, including pointer, lamp and character, which need to be detected in the automobile combination instrument. Compared with the traditional manual inspection, it not only reduces the detection error, but also improves the production efficiency. The main work is as follows: 1. The design scheme of the whole instrument detection system is put forward by studying the characteristics of the automobile instrument and combining the actual testing process. In the hardware aspect, according to the actual testing demand, the instrument driving design is carried out in the integrated way, and the appropriate image acquisition device is selected. In the software aspect, the whole detection process of the automobile combination instrument is designed by the content of the detection. 2. Aiming at the shortcomings of traditional image subtraction method, a subtraction method based on high precision template matching is proposed, which provides a guarantee for obtaining pointer images effectively. According to the characteristics of the instrument, a method of fitting the ellipse with the least square method is put forward. The long axis of the ellipse is extracted as the straight line, and the angle between the line and the horizontal line is measured. Then the relative position of the center of the pointer and the center of rotation is used to determine the actual deflection angle of the pointer. 3. The detection method of instrument LED lamp is studied. The brightness of instrument LED lamp is determined by gray value, and the color of instrument LED lamp is determined by converting it to the measurement hue value in HSV gamut space, and the existence of tin is determined by combining with the idea of set. The detection method of segment lamp is studied, and the detection of segment lamp is carried out by fitting the minimum external rectangle of segment lamp. The method of character detection is studied. The method of character preprocessing, character segmentation and support vector machine (SVM) is used to identify the number. 4 on the LCD screen. By the proposed hardware, software design and image processing algorithm to complete the implementation of the whole detection system. Operate the detection system and test the Cummins instrument of Dongfeng Automobile. The test results show that the vehicle combination instrument detection system based on machine vision designed in this paper can complete the more comprehensive detection of the instrument and ensure the accurate detection of each test item in the case of less time consumption. To meet the actual industrial testing requirements, has a certain development prospects and practical value.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;U463.7

【参考文献】