计算机与现代化

• 数据库与数据挖掘 • 上一篇    下一篇

一种基于Hough变换和图像融合的快速破损检测

  

  1. 针对工业破损检测的现状进行研究，提出运用数字图像处理系统代替传统的硬件检测。首先对薄膜进行图像预处理，运用Hough（霍夫）变换分割出连续2帧的感兴趣区域，运用小波变换增强图像
    对比度；其次计算各个像素点的EOI(拉普拉斯能量)和VI(对比度)，将归一化处理后的值作为图像融合的系数；最后把融合后图像转化到HSI（色度、饱和度、亮度）空间，进行阈值分割，进行破损判别
    。实验表明，这种检测方法具有抗干扰性强、成本低、准确性高、计算量小、识别速度快的优点，基本能实现包装的实时检测。
    关键词： 数字图像处理； Hough变换； 破损检测； 图像融合； 透明薄膜包装
  • 收稿日期:2014-06-05 出版日期:2014-10-10 发布日期:2014-11-04
  • 作者简介:刘娣(1989)，女，江苏盐城人，东华大学信息科学与技术学院硕士研究生，研究方向：图像处理与模式识别； 周武能(1959)，男，湖北荆州人，教授，博士，研究方向：神经网络，图像处理，工业 过程控制； 张详位(1989)，男，江西上饶人，硕士研究生，研究方向：神经网络与图像处理。

A Fast Breakage Detection Based on Hough Transform and Image Fusion

  1. The status of industrial breakage detection has been studied, we propose making the digital image processing system replace the traditional hardware detection, then
    the first is image preprocessing to membrane, using Hough transform to segment region of interest for two consecutive frames, image contrast enhancement using wavelet transform,
    the followed is to calculate the energy of Laplace(EOI) for each pixel and Visibility(VI), the normalized value of the processed image fusion as a factor, and finally fused
    image was transformed into HSI space, then is the threshold segmentation, to discriminate the damaged. Experiments show that this detection method has strong antiinterference,
    low cost, high accuracy, small amount of calculation to identify the advantages of speed, meets the requirements of the realtime detection in the industrial field.
  • Received:2014-06-05 Online:2014-10-10 Published:2014-11-04

被引次数

1

摘要：

针对工业破损检测的现状进行研究，提出运用数字图像处理系统代替传统的硬件检测。首先对薄膜进行图像预处理，运用Hough（霍夫）变换分割出连续2帧的感兴趣区域，运用小波变换增强图像
对比度；其次计算各个像素点的EOI(拉普拉斯能量)和VI(对比度)，将归一化处理后的值作为图像融合的系数；最后把融合后图像转化到HSI（色度、饱和度、亮度）空间，进行阈值分割，进行破损判别
。实验表明，这种检测方法具有抗干扰性强、成本低、准确性高、计算量小、识别速度快的优点，基本能实现包装的实时检测。

关键词:  , 数字图像处理； Hough变换； 破损检测； 图像融合； 透明薄膜包装

Abstract:

The status of industrial breakage detection has been studied, we propose making the digital image processing system replace the traditional hardware detection, then
the first is image preprocessing to membrane, using Hough transform to segment region of interest for two consecutive frames, image contrast enhancement using wavelet transform,
the followed is to calculate the energy of Laplace(EOI) for each pixel and Visibility(VI), the normalized value of the processed image fusion as a factor, and finally fused
image was transformed into HSI space, then is the threshold segmentation, to discriminate the damaged. Experiments show that this detection method has strong antiinterference,
low cost, high accuracy, small amount of calculation to identify the advantages of speed, meets the requirements of the realtime detection in the industrial field.

Key words: digital image processing, Hough transformation, breakage detection, image fusion, packaging of transparent film

中图分类号: 

版权所有 © 《计算机与现代化》编辑部
地址：江西省南昌市西湖区井冈山大道1416号　邮编：330003
电话：0791－86490996　 传真：0791-86492535　 Email：jgsdd@163.com
本系统由北京玛格泰克科技发展有限公司设计开发