隧道裂缝自动识别性能影响因素的研究

Study of Factors Affecting Automatic Detection Performance of Cracks in Tunnel

在室内环境下,利用混凝土试验模块,设计一套基于CMOS传感器件的装置来进行隧道裂缝自动识别试验。将裂缝图像灰度值分布和特征值相对误差作为衡量自动识别的重要指标,分析有效像素、检测距离、检测速度、光照强度等因素对自动检测性能的影响。模拟试验结果表明,有效像素和检测距离对图像分布的特征影响不大,但影响自动检测精度;有效像素增大,检测距离减小,相应的检测精度增大;检测速度和光照强度不仅对裂缝图像灰度分布特征影响大,对自动检测精度也有显著地影响,检测速度的增大会导致检测精度降低,光照强度过高或过低都会影响检测精度。

In the indoor environment,concrete test module is used to design a device to automatically identify cracks in the tunnel test based on CMOS sensors.The crack image gray value distribution and characteristic values are regarded as a measure of an important indicator of the relative error of automatic recognition,the impact of effective pixels,testing influence distance,test speed,light intensity and other factors on automatic detection performance are analyzed.The simulation results show that the effective pixel and detection distance of the features of the image distribution has little effect,but affect the automatic detection accuracy;with the increase of effective pixels,the detection distance is reduced,and the corresponding detection accuracy increases;the detection speed and light intensity distribution not only significantly influence the fracture image gray,but also have a significant impact on the automatic detection accuracy,detection rate increase will lead to lower detection accuracy,the light intensity is too high or too low will affect the detection accuracy.