摘要
针对远距离数字照相所采集的隧道衬砌裂缝图像自身特点,结合实验数据拟合标定曲线的相机标定方法,采用图像增强预处理、循环迭代法裂缝区域提取、数学形态学修整、亚像素边缘检测等图像处理手段,提出隧道衬砌裂缝识别及宽度量测新算法.经室内试验验证,当拍摄距离小于8m时,算法精度较高,裂缝宽度量测误差一般小于0.40mm;根据现场实际情况可知,该算法误差最大为0.37mm,最小为0.08mm,平均误差为0.20mm,满足工程测量的要求.
According to its own characteristics of longdistance-taken image with cracks on tunnel lining, a new identification and width-measuring algorithm for tunnel lining crack was developed by combining the calibration curve of camera fitted with experimental data, on the basis of some image processing means, such as image enhancement preprocessing, region extraction of crack with loop iteration method, mathematical morphological analysis, edge detection of crack with sub-pixel method. Verified by laboratory tests, when the shooting distance is less than 8 m, the algorithm is of high precision and the measurement error of crack width is generally less than 0.40 mm. According to its application, the maximum error of this algorithm is 0.37mm, the minimum error is 0. 08mm, the average error is 0. 20mm which meets the requirements of engineering measurement.
出处
《同济大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2012年第6期829-836,共8页
Journal of Tongji University:Natural Science
关键词
数字图像处理
隧道衬砌裂缝
区域提取
亚像素
边缘检测
标定方法
digital image processing
crack on tunnel lining region extraction
sub-pixel
edge detection
calibration method
