一种基于计算机视觉的炼铁原料自动识别方法
吴清潇; 陈奕璇; 朱枫; 贝文诚; 欧锦军; 郑建忠; 郝颖明; 周静; 付双飞
2010-06-23
专利国别中国
专利号CN101751552B
专利类型发明授权
产权排序1
权利人中国科学院沈阳自动化研究所
其他题名Automatic identification method of ironmaking raw materials based on computer vision
中文摘要本发明涉及一种基于计算机视觉的炼铁原料自动识别方法,包括以下步骤:采集各炼铁原料及原料运送装置空载时的彩色图像,根据彩色图像的样本建立特征库;在各炼铁原料及原料运送装置的彩色图像中进行特征提取,得到各彩色图像的特征值;将上述提取的彩色图像特征值与特征库进行匹配,确定当前原料的种类。本发明方法综合利用颜色特征和频谱特征与预先离线建立的特征样本库进行在线匹配以实现物料图像的分类和识别,对物料的识别准确无误,提高了工作效率,对后续准确分别计量物料提供了先决条件,具有非接触、柔性强、自动化程度高等优点。
是否PCT专利
英文摘要The invention relates to an automatic identification method of ironmaking raw materials based on computer vision, which comprises the following steps: collecting the color images of the ironmaking raw materials and an idle raw material conveyer, and establishing a feature library according to the samples of the color images extracting features in the color images of the ironmaking raw materials and the raw material conveyer to obtain the feature value of each color image matching the extracted feature values of the color images with the feature library and determining the sorts of the current raw materials. The method of the invention comprehensively uses color features and frequency spectrum features to match with the feature sample library on line, wherein the feature sample library is previously established off line, so that the classification and identification of the material images are realized, the identification of the materials is correct and inerrant, the working efficiency is improved, the prerequisite conditions of subsequently counting the materials respectively and accurately are provided, and the method has the advantages of non contact, strong flexibility, high degree of automation, and the like.
公开日期2012-06-27
申请日期2008-12-05
语种中文
专利申请号CN200810229325.1
专利代理沈阳科苑专利商标代理有限公司 21002
内容类型专利
源URL[http://ir.sia.cn/handle/173321/14273]  
专题沈阳自动化研究所_光电信息技术研究室
推荐引用方式
GB/T 7714
吴清潇,陈奕璇,朱枫,等. 一种基于计算机视觉的炼铁原料自动识别方法. CN101751552B. 2010-06-23.
个性服务
查看访问统计
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。


©版权所有 ©2017 CSpace - Powered by CSpace