首页> 外国专利> METHOD OF INVESTIGATING STATIONARY GAS DETECTION DEVICE INSTALLATION SITE

METHOD OF INVESTIGATING STATIONARY GAS DETECTION DEVICE INSTALLATION SITE

机译:固定式气体检测装置安装地点的调查方法

摘要

To prevent, when a non-image type gas detection device is used as a stationary gas detection device, the stationary gas detection device from allowing that there exists any place where gas might leak.SOLUTION: A method of investigating a stationary gas detection device installation site comprises: a first step of capturing, for a predetermined first period, infrared images of a predetermined region 31 where gas leakage is predicted in a monitor area 30, and executing image-based gas detection using the infrared images captured for the period; a second step of executing a learning phase of machine learning with, as teacher data, abnormal images (gas images and images of abnormal temperature positions) extracted from the infrared images by executing the first step, such that a learning model 16 for recognition of abnormal images is bui a third step of capturing infrared images in the monitor area for a predetermined second period; and a fourth step of performing, using the learning model 16, a recognition phase of the machine learning on the infrared images captured in the third step.SELECTED DRAWING: Figure 2
机译:为防止将非图像类型的气体检测设备用作固定式气体检测设备时,该固定式气体检测设备不允许存在任何可能泄漏气体的地方。解决方案:一种调查固定式气体检测设备安装的方法该站点包括:在预定的第一时间段内捕获在监视区域30中预测有气体泄漏的预定区域31的红外图像,并使用在该时间段内捕获的红外图像执行基于图像的气体检测的第一步;第二步骤,通过执行第一步骤,以从红外图像中提取的异常图像(气体图像和温度位置异常的图像)作为教师数据,执行机器学习的学习阶段,使得用于识别异常的学习模型16图像被建立;第三步,在监视区域中捕获红外图像达预定的第二时间段;第四步是使用学习模型16对在第三步中捕获的红外图像执行机器学习的识别阶段。选定的图:图2

著录项

  • 公开/公告号JP2020016527A

    专利类型

  • 公开/公告日2020-01-30

    原文格式PDF

  • 申请/专利权人 KONICA MINOLTA INC;

    申请/专利号JP20180139211

  • 发明设计人 KUISEKO MANAMI;

    申请日2018-07-25

  • 分类号G01M3/02;

  • 国家 JP

  • 入库时间 2022-08-21 11:35:44

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