«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

期数:

2012年08期

页码:

181-184

栏目:

应用开发研究

出版日期:

1900-01-01

文章信息/Info

Title:

Steel Bar Adaptive Counting Based on Gray-level Co-occurrence

文章编号:

1673-629X（2012）08-0181-04

作者:

刘娜娜

江苏科技大学计算机科学与工程学院

Author(s):

LIU Na-na

Institute of Computer Science and Engineering,Jiangsu University of Science and Technology

关键词:

棒材自适应计数; 局部二元模式; 二值化; 聚类方法

Keywords:

steel bar adaptive counting;  local binary pattern;  binarization;  clustering method

分类号:

TP391

文献标志码:

A

摘要:

针对棒材生产车间环境的复杂性和现有棒材计数方法的不精确性，文中提出了一种准确，高效的棒材自适应计数方法。该方法采用局部二元模式描述棒材的截面纹理，有效地将棒材截面图像与复杂的背景进行分离，使用阈值渐增的二值化方法获取图像中的局部灰度极大值，将此最大值点作为棒材的中心点。最后对棒材的半径使用聚类方法对误判点进行过滤，进而统计计数。实验结果表明，该方法极大地提高了棒材计数的效率，并将识别准确率提高到了98％

Abstract:

In order to resolve the problem of the complexity of the bar production plant environment and the inaccuracy of the existing counting methods, an accurate and efficient adaptive bars counting method is presented in this paper. This method uses the principle of local binary pattern to describe the texture of the bar cross-section, which effectively separates the background of the steel bars. And it uses binarization method with increasing threshold to get the local maximum value of the gray level image, as the center point of the bar. At last, on the radius uses clustering method to filter the misjudged points and then counts statistics. The experiments show that this method greatly improves the efficiency and the accurate to ninety eight percent

备注/Memo

备注/Memo:

国家自然科学基金（60573063,60573064）刘娜娜（1986-），女，硕士，主要研究领域为图像处理、模式识别

常用功能

更新日期/Last Update: 1900-01-01