节点文献

组合砌体墙损伤红外热像检测方法研究

Research on the Infrared Thermal Image Detection Approach for Composite Masonry Wall Damage

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 李瑞陈飞

【Author】 LI Rui;CHEN Fei;Geological Survey of Jiangsu Province;

【机构】 江苏省地质调查研究院

【摘要】 轻钢芯柱-保温夹层组合砌体墙的裸露部分受自然因素影响易出现墙体损伤,影响墙体的使用寿命,为此提出组合砌体墙损伤红外热像检测方法。该方法通过超声红外热成像技术,获取组合砌体墙损伤的红外图像,基于双边滤波方法,提升了组合砌体墙损伤图像的低照度;基于图像处理技术和形态学检测技术,对组合砌体墙损伤边缘进行精准提取,确定损伤程度,实现精确检测。结果表明:当图像数量为20~100张时,该方法检测的mAP值与召回率均高于90.00%,墙体损伤检测较为完整且损伤支线明显,直观表现了检测效果;准确度均高于95.00%,提升了墙体损伤检测方法的检测完整度;平均IoU均高于85.00%,有效提高了损伤区域与标定区域之间的重叠率。该方法的检测效果好、性能高。

【Abstract】 The exposed part of the light steel core column insulation sandwich composite masonry wall is prone to wall damage due to natural factors, which affects the service life of the wall.Therefore, an infrared thermal imaging detection method for damage of composite masonry walls is proposed.This method uses ultrasonic infrared thermal imaging technology to obtain infrared images of damage to composite masonry walls, filtering out pixels below the optimal threshold.Based on the bilateral filtering method, Based on image processing technology and morphological detection technology, accurate extraction of damaged edges in composite masonry walls is carried out to determine the severity of the damage and achieve precise detection.The results showed that when the number of images was 20 to 100,the mAP value and recall rate detected by our method were both higher than 90.00%.The wall damage detection was relatively complete and the damage branches were obvious, which intuitively demonstrated the detection effect; The accuracy is higher than 95.00%,which improves the completeness of the wall damage detection method; The average IoU is higher than 85.00%,effectively improving the overlap rate between the damaged area and the calibration area.This method has good detection effect and high detection performance.

  • 【文献出处】 粉煤灰综合利用 ,Fly Ash Comprehensive Utilization , 编辑部邮箱 ,2024年05期
  • 【分类号】TU364;TU312.3
  • 【下载频次】5
节点文献中: 

本文链接的文献网络图示:

本文的引文网络