自动数据采集环境下的沥青路面评价指标研究

One of the key factors of the scientific pavement maintenance management is to build a reasonable pavement evaluation index system. The increasingly refined pavement decision-making process requires clear pavement evaluation indicators, which can also describe the pavement conditions comprehensively. At the same time, pavement distress composition tends to be single for high-level asphalt pavements, but the deterioration rates of different distress are significantly different. Besides, pavement distress data information is largely enhanced under currently popular automated inspection environment both in the content and the professional degree. That is, whether it is from the application, research or the technology point of view, the requirements for the asphalt pavement evaluation system have tended to be single, independent and complete. However, the existing asphalt pavement evaluation system in China comes from the traditional method of artificial visual pavement survey and is consequently mainly based on comprehensive index and also with an overlap in index connotation, which has been unable to match the current needs. Therefore, it is proposed in this study to establish a new asphalt evaluation index system to adapt to the automated inspection environment. The first step is to optimize the data collection platform to fulfill the automated pavement damage and three-dimensional deformation detection. Then to build the new pavement evaluation system, which is consisted by mutually independent evaluation indicators with single connotations. And in this new system, pavement damage and deformation description could be completely separated. At the same time, the distress shape and location information could be associated with the evaluation indicators to characterize the causes of distress to support the refined pavement maintenance decision-making. At last, a large amount of pavement survey will be carried out to verify the reasonableness and effectiveness of the new index system. The testing and verification process will be conducted repeatedly according to the feedback to finally establish the stable and reasonable index system.

合理的路面评价指标体系是路面科学养护管理的关键之一。日趋精细化的养护决策要求评价指标针对性明确、指标构成能全面描述路况；同时高等级沥青路面的损坏形式逐渐趋于单一、不同损坏的发展速率却差异显著；当前普及的路面数据自动采集也使损坏数据内容丰富、专业性增强。也即无论是从应用、研究还是技术实现的角度，对沥青路面评价指标的要求都趋于单一化、独立化和完整化。而我国现行的评价体系脱胎于传统的人工目测调查方法，以综合性指标为主、指标内涵重叠，已无法适应这一需求。因此本课题拟建立自动数据采集环境下的沥青路面评价指标体系。首先完善数据采集平台，实现路面破损和三维变形检测；然后基于此框架提出指标内涵单一且互相独立的评价指标体系，彻底分离破损与变形，并在指标中关联损坏形态和位置信息，以表征病害成因、支持精细化养护决策；最后，在路面实测的基础上进行指标体系的合理性和有效性验证，通过多次反馈形成稳定合理的最终方案。

借助综合检测设备对沥青路面损坏进行自动检测已成主流趋势，但传统的沥青路面评价指标与之不完全匹配，无法充分利用路面破损信息。本课题自主开发并升级了沥青路面自动化检测平台，实现了对路面破损病害的高清成像和基于双激光结构光的路面三维变形测量，并研发了相应的图像处理算法，实现了对路面开裂的精确识别和基于激光结构光图像的三维变形识别。随后在图像分析的基础上，提出了多个适用于精细化路面评价的指标，包括表征路面破损类病害的开裂指数、等效裂缝率、横缝贯穿率、横缝间距、等效修补率等，以及表征路面变形类的车辙包络线深度、车辙填充面积等，并对这些指标的适用性进行了研究。为了验证所建立指标体系的有效性，本项目利用从其它检测平台中获取的超过400公里的路面实测数据进行了验证，同时开展了较大范围的人工路况调查（10公里）进行数据对比，证明现有的自动化检测手段能够较为准确地获取路面中等程度以上的开裂和车辙变形信息，但对轻度损坏识别仍存在不可忽视的信息丢失，导致其难以直接指导对数据精细度有较高要求的项目级路面养护决策，而本课题建立的指标体系能够有效弥补这一空白。