连续钢桁梁桥涂装试板损伤红外检测  

Study on Infrared Detection Method of Coating Test Plate Damage of Continuous Steel Truss Bridge

作　　者：王猛 谭剑锋 WANG Meng;TAN Jianfeng(Guangdong Jiaoke Testing Co.,Ltd.,Guangzhou Guangdong China 510550)

机构地区：[1]广东交科检测有限公司,广州510550

出　　处：《广东公路交通》2022年第6期82-86,92,共6页Guangdong Highway Communications

摘　　要：连续钢桁梁桥涂装试板损伤检测对判断桥梁腐蚀程度具有重要意义,为此提出连续钢桁梁桥涂装试板损伤红外检测方法。分析了红外检测原理以及涂装试板损伤检测的温度特征,采用红外探照设备获取涂装试板红外图像,并得到去噪后的图像采集结果。根据微分计算理念,建立微分形态学边缘检测算子,增强原始图像损伤边缘。通过多尺度小波变换得到增强图像包含的细节特征信息,将其输入增量式极限学习机(Incremental Extreme Learning Machine, I-ELM)损伤检测网络结构,生成涂装试板损伤检测结果。工程实例分析结果表明:红外损伤检测方法的标准差为0.02,具有较好的检测性能。The damage detection of coating test plate of continuous steel truss bridge has been of great significance to judge the corrosion degree of continuous steel truss bridge, therefore, an infrared detection method of coating test plate damage of continuous steel truss bridge has beenproposed. After clarifying the infrared detection principle and the temperature characteristics of the coating test plate damage detection, the infrared image of the coating test plate has been obtained relying on the infrared detection equipment, and the image acquisition results after denoising have been obtained. According to the concept of differential computing, a differential morphological edge detection operator has been established to enhance the damaged edge of the original image. The detailed feature information contained in the enhanced image has been obtained by multi-scale wavelet transform, which has been input into the damage detection network structure of Incremental Extreme Learning Machine(I-ELM), and the damage detection results of coating test panels have been generated. The example analysis results have shown that the standard deviation of the proposed infrared damage detection method is 0.02, which has good detection performance.

关 键 词：连续钢桁梁桥 涂装试板 损伤检测 红外图像 边缘增强 特征提取 

分 类 号：U448.36[建筑科学—桥梁与隧道工程]