钢结构表观裂纹监测技术对比与展望

Comparison and Prospect of Apparent Crack Monitoring Technology for Steel Structures

随着钢结构的广泛应用,其疲劳开裂问题也日益得到重视,尤其是对钢结构开裂情况把握的实时性需求日益迫切。首先总结和对比分析了常用钢结构疲劳开裂检测手段的技术特点和局限;而后结合钢结构开裂实时信息获取的现实需求,引入了钢结构表观裂纹监测的概念,提出了在统一时间尺度下疲劳裂纹发展过程追踪的自动化监测与识别方法。通过微视距成像设备裂纹信息采集、图像拼接、裂纹分割及基于深度学习的自动化识别,实现裂纹扩展形态信息的自动获取。分析与研究表明,计算机视觉技术在钢结构裂纹监测领域的应用与发展可以突破目前该领域的技术局限。

With the wide application of steel structures, the problem of fatigue cracking has become increasingly prominent, especially the de-mand for the real-time cracking of steel structures is becoming more and more urgent. Firstly, the technical characteristics and limitations of the commonly used cracking detection methods of steel structure fatigue are summarized and compared. Then in combination with the actual demand of real-time information acquisition of steel structure cracking, the concept of apparent crack monitoring of steel structure is intro-duced. The method of automatic monitoring and identification of fatigue crack development process tracking on a unified time scale is proposed. In this paper, the crack information acquisition,image splicing, crack segmentation and automatic recognition based on deep learning are real-ized by micro-vision imaging equipment and this method has realized the automatic acquisition of crack propagation morphology information. The analysis and research show that the application of computer vision technology in the field of steel structure crack monitoring can break through the technical limitations of the current field.