摘要
墙体结构的表面裂缝对建筑结构的健康和美观都有一定程度的影响,因此墙面裂缝的监测和检测工作较为重要。基于数字图像的检测方法人工参与度低,自动化程度高,能快速识别裂缝。本文对比了传统微分算子识别图像裂缝的效果,选取Canny算法为基本算法,在算法中引入导向滤波和同态滤波进行改进,达到消除噪声、保留和增强边缘细节信息的目的。在识别裂缝的基础上,通过边界像素的计算测量裂缝的像素长度和宽度,通过像素解析度求得实际长度。通过实验对比手工测量值和图像测量值,在普通相机标定情况下,图像测量的裂缝几何特征值精度较好,与手工测量值的误差在3%左右。因此该方法能够为后续裂缝的常态监测和检测的智能化提供一种思路。
The surface cracks of the wall have certain influence on the health and beauty of the building structure,so the monitoring and detection of the wall cracks is more important.The detection method based on digital image has the advantages of low manual participation and high automation,and can quickly identify cracks.In this paper,Canny algorithm is selected as the basic algorithm based on the comparison of the effect of traditional differential operator in identifying image cracks.In Canny algorithm,guiding filter and homomorphic filter are introduced to improve the performance of noise reduction,edge detail preservation and enhancement.Based on the recognition of crack,the pixel length and width of crack are measured by the calculation of boundary pixel,and the actual length is obtained by the resolution of Pixel.Comparing the measured values of cracks by site surveying with the that of our method,the precision of the geometric eigenvalue of the crack measured by image is better,and the error is about 3%with the measured value by site surveying.Therefore,this method can provide a new idea for the normal monitoring and intelligent detection of the follow-up fractures.
作者
缪盾
Miao Dun(Civil Engineering Department,Tongji Zhejiang College,Jiaxing 314000,China)
出处
《工程勘察》
2021年第10期49-53,共5页
Geotechnical Investigation & Surveying
基金
浙江省教育厅科研项目“基于数字图像的混凝土表面病害信息提取和测量”(Y202045082).
