基于沉降点占比的非错台型桥头沉降检测方法

A Method for Detecting Non-staggered-step Bridgehead Settlement Based on Settlement Points’ Ratio

现行国家标准未对非错台型桥头沉降的检测方法作出规定,而现有方法主观低效、普适性低且易受路面病害干扰。针对上述问题,利用车载式检测系统快速采集道路纵坡数据与路面图像,对纵坡数据和行驶距离进行数据预处理后形成道路纵断面线,并将其与图像位置进行了对应。基于此建立了一种非错台型桥头沉降计算模型与算法,其中提出沉降点占比SPR这一计算指标。在3种车速下,分别对各15组距离梯度进行了试验。结果表明:车载式检测系统符合国家规程有关纵坡自动化检测的误差要求。同时对试验对象选取和计算模型取值作出了规定,在宁波境内对100座桥开展了试验。结合试验数据,利用所建模型与算法输出了200组(上、下桥)SPR,并将其与人工测量的桥接缝前后200组纵坡差绝对值进行了相关性和回归分析,两者的相关系数为0.957 4,回归模型决定系数为0.916 6,表明所提方法的输出结果与人工试验结果间有很强的相关性,且回归模型精度较高。因此,检测方法对于检测和定量描述非错台型桥头沉降状况是科学合理的,有效填补了桥头沉降检测的研究空白。

The current national standards have not stipulated the detection method of non-staggered-step bridgehead settlement, and the existing method is subjectively inefficient, low in universality and easily disturbed by road surface diseases. In view of the above problems, the road longitudinal slope data and road images are quickly collected by using vehicular detection system. The data of longitudinal slope and driving distance are preprocessed to form the road profile line, which corresponded to the pavement image position. Based on this, a calculation model and algorithm of non-staggered-step bridgehead settlement are established, in which a calculation indicator named Settlement Points’ Ratio(SPR) is proposed. The 15 groups of distance gradients at 3 speeds are tested respectively. The result shows that the vehicular detection system meets the error requirements in national regulations on automatic detection of longitudinal slope. Meanwhile, the selection of test objects and the values of calculation model are specified, and 100 bridges in Ningbo are tested. According to the test data, 200 groups of SPR(both sides of the bridge) are output by using the established model and algorithm. The correlation and regression analysis between the 200 groups of manual measured absolute values of longitudinal slope difference at front and rear of the bridge joint and the output data are conducted. The correlation coefficient is 0.957 4 and the determination coefficient of the regression model is 0.916 6,indicating that there is a strong correlation between the output result obtained by the proposed method and the manual test result, and the regression model has high accuracy. Therefore, the proposed method is scientific and reasonable for the detection and quantitative description of non-staggered-step bridgehead settlement, it effectively filled the research gap of bridgehead settlement detection.