您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(工学版)》

山东大学学报(工学版) ›› 2009, Vol. 39 ›› Issue (5): 68-72.

• 机械工程 • 上一篇    下一篇

基于RGB颜色空间的异性纤维识别检测算法

冯显英 张成梁 杨丙生 李蕾   

  1. 冯显英 张成梁 李蕾:山东大学机械工程学院, 山东 济南 250061;
    冯显英:高效洁净机械制造教育部重点实验室(山东大学), 山东 济南  250061;
    杨丙生: 山东天鹅棉业机械股份有限公司, 山东 济南 250032
  • 收稿日期:2009-05-19 出版日期:2009-10-16 发布日期:2009-10-16
  • 作者简介:冯显英(1965-),男,山东金乡人,教授,博士生导师,研究方向为智能检测与数控技术.E-mail:FXYing@sdu.edu.cn
  • 基金资助:

    引进国际先进农业科学技术计划(948计划)资助项目(2006-Z17)

Foreign fiber recognition and detection algorithm based on RGB color space

  1.  FENG Xian-Ying, ZHANG Cheng-Liang, LI Lei: School of Mechanical Engineering, Shandong University, Jinan 250061, China;                                                FENG Xian-Ying:Key Laboratory of High Efficiency and Clean Mechanical Manufacture(Shandong University), Ministry of Eduaction, Jinan 250061, China;
    YANG Bing-Sheng: Shangdong Swan Cotton Industrial Machinery Stock Co., Ltd, Jinan 250032, China
  • Received:2009-05-19 Online:2009-10-16 Published:2009-10-16

摘要:

研究了棉纤维的RGB色彩空间分布模型,提出了精确物理模型算法——彩色聚色算法.在此基础上,提出了实时检测算法.为了进一步提高算法效率,研究了改进后的实时检测算法,用于实现异性纤维检测系统中目标与背景的识别.由于识别目标的多样性以及复杂性,仅靠一两个特征无法得到满意的结果,通过提取颜色、面积、形状等特征,提出了基于组合特征的彩色图像多目标识别方法,通过多个目标特征的组合可以提高目标识别的质量.

关键词: 异性纤维;彩色图像分割;组合特征

Abstract:

The cotton fiber distributing model in RGB color space was studied in the article and an accurate physical model arithmetica color image clustering   algorithm was proposed. A real- time detecting algorithm was developed on this basis. In order to enhance the algorithm efficiency, the improved real-time detecting algorithm was presented to extract the target from backgrounds in the foreign fibers inspecting system. Due to the complexity and diversity of recognizing a foreign fiber, it is hard to meet the requirements using one or two characters. A multiple targets recognition method based on combined characters such as color, area and shape was presented, which can highly increase the quality of targets recognition.

Key words: foreign fiber; color image segment; combined characters

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!