首页> 外国专利> DEFECT DETECTION METHOD, DEFECT DETECTION APPARATUS, LEARNING METHOD, PROGRAM, AND RECORDING MEDIUM

DEFECT DETECTION METHOD, DEFECT DETECTION APPARATUS, LEARNING METHOD, PROGRAM, AND RECORDING MEDIUM

机译:缺陷检测方法,缺陷检测设备,学习方法,程序和记录介质

摘要

PROBLEM TO BE SOLVED: To execute defect detection of different images with time and accuracy of a substantially same level.SOLUTION: A defect detection method includes: a step for receiving an input image of an industrial component; a step for receiving a model type for the input image; a step for acquiring a model image corresponding to the model type; a step for estimating inter-image transformation between the model image and the input image; a step for acquiring a positioned input image and a positioned model image by transforming the model image and the input image to a common coordinate system based on the estimated inter-image transformation; a step for forming a plurality of image difference vectors for the input image and the model image; and a step for applying a statistical classification model for generating a labeled classification map for the input image to the plurality of image difference vectors.
机译:解决的问题:以基本相同的时间和精度执行不同图像的缺陷检测。解决方案:缺陷检测方法包括:接收工业部件的输入图像的步骤;接收输入图像的模型类型的步骤;获取与模型类型相对应的模型图像的步骤;估计模型图像和输入图像之间的图像间变换的步骤;根据估计的图像间变换,通过将模型图像和输入图像变换为共同坐标系,来获取位置输入图像和位置模型图像的步骤。为输入图像和模型图像形成多个图像差矢量的步骤;该方法还包括以下步骤:将统计分类模型应用到所述多个图像差矢量上,所述统计分类模型用于为所述输入图像生成标记的分类图。

著录项

  • 公开/公告号JP2012032370A

    专利类型

  • 公开/公告日2012-02-16

    原文格式PDF

  • 申请/专利权人 SHARP CORP;

    申请/专利号JP20110052176

  • 发明设计人 YANASE MASAKAZU;SIN YU SHU;CHAN YUEN;

    申请日2011-03-09

  • 分类号G01N21/88;G01B11/30;G01N21/956;G06T1/00;

  • 国家 JP

  • 入库时间 2022-08-21 17:42:37

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号