摘要
依据数字图像技术,构建图像灰度值表征参数与真实粗糙度的定量关系,研究图像前处理、相机感光度、拍摄距离、环境照度、光照角度、采样间距及插值方法等因素对混凝土表面粗糙度图像特征参数的影响.结果表明:图像前处理可有效降低输入误差;相机感光度和拍摄距离对图像灰度平均值、灰度值起伏均方根影响较小;环境照度和光照角度对图像特征及表征参数影响显著;采样间距与灰度差均值负相关;插值方法对图像特征参数影响较小,但合理选取采样间距和插值方法可显著提高运算效率.推荐控制条件下隧洞管片粗糙度计算值与测量值平均相对误差、均方根误差和平方根分别为0.79%、0.3110 mm和0.9804.研究成果可有效促进数字图像技术在混凝土粗糙度测量领域的应用.
The quantitative relationships between the image characterization parameters of gray value and the real roughness were constructed by digital image technology.The influences of image pre-processing,light sensitivity of camera,shooting distance,ambient illumination,illumination angle,sampling interval,and interpolation method on the image characteristic parameters of the surface roughness of concrete were investigated.Results indicate that the image input error can be limited by image pre-processing.The value of international standard organization(ISO)and the shooting distance have less effect on the gray mean value and root mean square(RMS)of fluctuation of gray value.The ambient illumination and the illumination angle have a significant impact on the image features and characterization parameters.The sampling interval is negatively correlated with the mean gray difference.The interpolation methods have less effect on image feature parameters,but a reasonable sampling spacing and interpolation method can significantly improve the computational efficiency.Under the recommended control conditions,the absolute error,root mean square error and square root between the calculated and measured values of the roughness of water conveyance tunnel segments are 0.79%,0.3110 mm and 0.9804,respectively.The research results can effectively promote the application of digital image technology in the field of concrete roughness measurement.
作者
甘磊
金洪杰
沈振中
Gan Lei;Jin Hongjie;Shen Zhenzhong(College of Water Conservancy and Hydropower Engineering,Hohai University,Nanjing 210098,China;State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering,Hohai University,Nanjing 210098,China)
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2022年第3期497-505,共9页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(52179130,51609073)
江苏省自然科学基金资助项目(BK20201312).
关键词
混凝土
数字图像
粗糙度
灰度值
影响因素
concrete
digital image
roughness
grey value
influencing factors
