基于机器视觉的工件检测系统若干问题研究.pdf

合肥工业大学
硕士学位论文
基于机器视觉的工件检测系统若干问题的研究
姓名:曹勇
申请学位级别:硕士
专业:信号与信息处理
指导教师:鲁昌华
2011-04
基于机器视觉的工件检测系统若干问题的研究

摘要

国内多数厂家对O型密封橡胶圈质量的检测多使用卡尺测内、外径及截面
的方法,这种测量方法误差大、检测效率低,同时检测的精度和速度也达不到
生产的要求。机器视觉技术因其非接触、速度快、检测精度高等优点越来越受
到厂商的重视。本文设计了一套基于机器视觉技术的O型密封橡胶圈在线检测
系统,其主要的研究内容如下:
,设计了O型密封橡胶圈的检测装置,并搭建了
系统的硬件实验平台。
、图像分割、边缘检测等经典算法,比较分
析各种算法的优劣性,选择了适合本课题的Curvelet混合滤波,有效地去除了图
像中的混合噪声,大大提高了信噪比;采用大津法进行图像分割、形态学进行
边缘检测,能很好地提取出边缘信息。
3. 根据橡胶圈的检测要求,采用边缘像素扫描法进行厚度检测、Hough变
换法对橡胶圈图像进行圆心定位和内外径测量,并利用特征参数快速形成模板
的方法进行缺陷检测。
4. 采用了DSP技术提高检测系统效率,完成了图像处理算法由PC机到DSP
的移植,并对检测系统进行了测试和误差分析。



关键词:机器视觉;橡胶圈;Curvelet; DSP
Research on Several Problems of Workpiece
Inspection System Based on Machine Vision

ABSTRACT

Calipers have monly used by the majority of domestic manufacturers as the
method to measure and detect thickness, section, inner and outer radius of O-shape sealing
rubber ring. However, this measurement has a disadvantage of large error, low detection
efficiency, meanwhile detection precision and speed also could not reach the requirements
of manufacturing. The machine vision technology, owing to the advantages of non-contact,
fast speed, and high detection precision, has been valued by more and more manufacturers.
In this paper, we designed a set of on-line detection system for O-shape sealing rubber ring
based on machine vision technology, its main contents were as follows:
1. Formulated systematic design scheme, designed O-shape seal rubber ring detection
device, and constructed experimental hardware platform.
2. Through the experiments, we verified the classical algorithms of image
preprocessing, image segmentation, edge detection, etc. After parative analysis of
merits and inferiorities between different algorithms, we present the method of Curvelet
hybrid filter adapt to the project. It’s proved that this method could effectively remove the
mixed noise in image and increase the SNR. T