基于AI分类和目标检测的三维探地雷达数据自动解译方法  

3D GPR Data Interpretation System Based on AI Classification and Target Detection

作　　者：黄钰琳 董珏 库捷峰 唐剑 Huang Yu Lin;Dong Jue;Ku Jie Feng;Tang Jian(Shenzhen Ande Space Technology Co.Ltd,Guangdong Province,Shenzhen 518000)

机构地区：[1]深圳安德空间技术有限公司,深圳518000

出　　处：《建设科技》2023年第3期75-78,共4页Construction Science and Technology

摘　　要：目前,三维探地雷达以其高精度、高效率、无损的工作方式,成为对道路隐患探测最为有效的手段之一。然而,数据解译工作严重依赖于专业解译人员,存在着标准不统一、效率低下、解译结果不稳定等问题,严重影响了道路坍塌隐患检测的效果。针对以上问题,本文提出了“基于AI分类和目标检测的三维探地雷达数据自动解译方法”。通过提高解译工作的智能化水平,有效提高了解译工作效率,保证了解译结果的质量,降低了人力成本,使大规模、周期性的道路塌陷隐患普查成为可能。Nowadays, 3D GPR becomes one of the most effective means to detect subsurface defects due to its high precision, efficiency and non-destructive working mode. However, data interpretation relies heavily on skilled interpreters, causing problems like inconsistent standard, low efficiency and unstable interpretation results. To solve the above issues, this paper proposes "3D GPR data interpretation system based on AI classification and target detection". By improving the intelligence level of data interpretation, it greatly improves work efficiency, ensures the stability of interpretation results, reduces a large number of labor costs, and makes it possible to conduct large-scale and periodic road survey.

关 键 词：城市道路塌陷隐患探测 三维探地雷达 人工智能 数据解译方法 

分 类 号：U418[交通运输工程—道路与铁道工程]