基于三维自动化检测数据的沥青路面旧路评价方法研究

Reasonable evaluation method for existing road is basis of large and medium scale repair of the road, thus higher request to the fine evaluation of the existing road is needed. Existing road condition evaluation indexes based on automatic detection data are often aimed at the macro indicators for network level management. However, the existing evaluation index system which is mainly composed by comprehensive indexes can not meet the needs of the existing road evaluation and the decision making of large and medium scale repair. Project grade evaluation method of road structure based on excavation, core drilling and indoor test has destroyed the original structure of the road, and the representative of is not so good because of its locality, which can not meet the needs of large-scale detection. With the development and application of three-dimensional automatic detection technology, research on the model for evaluation of existing road got a practical basis. Therefore, this paper intends to establish the evaluation method of asphalt pavement based on 3-D detection technology. Firstly, collect pavement distress and deformation data by 3-D detection method with typical road sections; secondly, make a correlation analysis among deflection testing, deformation, cracks and other distress with the typical pavement structure damage; finally, based on evaluation on road facilities in different stages of its life cycle, repeated verification the evaluation method to make it reasonable and effective and form a complete evaluation scheme.

合理的旧路评价方法是路面大中修及专项养护维修决策的基础，专项养护维修方案的制定对在役道路的精细化评价有着更高的要求。现有的基于自动化检测的各项路况评价指标往往是针对网级管理的宏观指标。然而，以综合性指标为主的现有评价指标体系无法满足项目级旧路评价及大中修决策的需求。以开挖、钻芯及室内试验为主的项目级道路结构评价方法则破坏了道路原结构，且局部检测代表性差、无法满足大规模检测的需求。随着路面三维检测技术自动化的发展和应用，适合于旧路评价的检测指标的模型研究具备了实践基础。因此，本课题拟建立基于三维检测技术的沥青路面旧路评价方法。首先，选取典型路段，获取路面破损及变形三维检测数据；其次，在结构弯沉检测及应力应变分析的基础上，针对路面变形、裂缝2类典型损坏同道路结构层损坏进行关联分析；最后，通过对处于生命周期内不同阶段的道路设施进行检测评价，完成评价方法合理性和有效性的反复验证，形成完整的评价方案

合理的旧路评价方法是路面大中修及专项养护维修决策的基础，专项养护维修方案的制定对在役道路的精细化评价有着更高的要求。本项目引入路面三维检测自动化技术，改进了课题组前期研发的三维检测装置，拓展了其检测尺度范围，增加了修正检测的循迹系统。关联分析了不同路面结构类型、不同路龄的病害特征，引入病害分类特征作为监督学习损坏特征降维的算法，并证明了其有效性。将道路损坏程度根据维修需求初步分为表面性能退化、面层结构破坏、基层结构破坏、路基结构不稳定 4 个层面，综合利用道路结构信息、路龄、病害特征以及平整度、磨耗等其它指标共同用于路面损坏程度判别，建立了损坏程度判别的双层决策树算法，采用现场调研及局部弯沉检测的方法验证了损坏程度判别结果的准确性。此外，拓展分析不同病害的成因分析，开发了基于修正随机森林算法的成因分析算法。