基于GPR的道路结构层病害双视图联合标注数据集

本数据集隶属于交通运输、仓储和邮政业下道路运输业的公路管理与养护领域，依托三维探地雷达（GPR）技术构建。采集时依据《城市地下病害探测技术规程》（CJJ/T291-2018），设500MHz-2GHz工作频率与0.5-2m/s扫描速度，探测城市地下0-5m结构，获取电磁特性数据并生成雷达波形图，呈现横纵向裂缝、脱空、富水等五类病害信号特征；经异常信号筛选、数据空缺填补、格式转换及坐标脱敏预处理，结合时间戳与空间坐标等实现同路段数据无缝拼接，再由专业人员标注病害位置、范围及严重程度，形成可训练数据，格式以图片为主。其可助力市政部门制定养护计划、交通建设单位把控施工质量、检测机构提升报告科学性、科研院校开展技术研究，服务于城市道路等地下工程检测、评估、维护及科研场景。

This dataset belongs to the road management and maintenance subfield of road transportation under the transportation, storage and postal industry, and is constructed based on three-dimensional ground-penetrating radar (GPR) technology. During data collection, it follows the Technical Specification for Urban Underground Disease Detection (CJJ/T291-2018), with operating frequency set at 500MHz–2GHz and scanning speed set at 0.5–2m/s. It detects urban underground structures at a depth of 0–5m, acquires electromagnetic characteristic data, generates radar waveform maps, and presents the signal features of five types of diseases including transverse and longitudinal cracks, voids, and water-rich areas. After preprocessing procedures such as abnormal signal screening, data gap filling, format conversion and coordinate desensitization, seamless stitching of data from the same road section is realized by combining timestamps and spatial coordinates. Subsequently, professional personnel label the position, scope and severity of the diseases, forming trainable datasets that are primarily in image format. This dataset can assist municipal departments in formulating maintenance plans, transportation construction units in controlling construction quality, testing institutions in improving the scientific rigor of their reports, and research institutions and universities in conducting technical research, serving scenarios including urban road underground engineering detection, evaluation, maintenance and scientific research.

搜集汇总

背景与挑战

背景概述

该数据集是一个基于GPR的道路结构层病害双视图联合标注数据集，由多位个人共同登记，数据来源于企业，目前处于公示中状态。它专注于道路结构层病害的检测与标注，采用双视图联合方法，旨在提升病害识别的准确性。