基于横断面特征的沥青路面车辙损伤主层位判别

Identification of Main Layers of Rutting Damage on Asphalt Pavement Based on Cross-Sectional Characteristics

为探究沥青路面车辙表面形态与面层损伤主层位间的相关关系,采用13点激光车辙仪采集了3条车辙共1.1km数据,计算了7类表面横断面指标,并以10m评定单元结合钻芯数据标定了车辙损伤主层位。然后,构建了集成随机森林、逻辑回归、朴素贝叶斯的多分类器识别分类模型进行损伤层位自识别并通过交叉验证测试结果,最终建立路面横断面指标与面层破损主层位的关系。试验结果表明:损伤主层位识别模型整体平均预测准确率达到97%,整体预测准确性良好;筛选车辙最大深度、最大水平宽度、负面积、正负面积比、凸凹比为指标时,平均预测准确率最高;损伤主层位的误判主要发生在车辙损伤主层位过渡段或存在其他病害的区域。

To study the correlation between rutting surface morphology and the main surface damage of asphalt pavement, profile data for 3 rutting sections in a total of 1.1km were collected by 13-point laser rutting instrument, 7 kinds of transverse profile indexes were calculated, and the main level of rutting damage was demarcated with the 10m evaluation unit combined with the drill core data.Then, a multi-classifier identification and classification model integrating random forest, logistic regression, and naive Bayes was constructed for self recognition of damaged layers. Through cross validation of test results, the relationship between the pavement crosssection indexes and the surface damage main horizon was finally established.The experimental results show that the overall average prediction accuracy of the damage main horizon identification model reaches 97%, and the overall prediction accuracy is good;When the maximum rutting depth, the maximum horizontal width, the negative area, the ration of positive and negative area, and the convex concave ratio are taken as index subset, the average prediction accuracy is the highest. The misjudgment of the main layer of rutting damage mainly occurs in the transition section of the main level of rutting damage or in areas with other diseases.