URDD: An open dataset for urban roadway disease detection and classification

We present two urban road disease datasets: DURDD for road disease detection and CURDD for road disease classification. DURDD includes four main types of underground road diseases: cavity, detachment, water-rich, and looseness. It also contains disease detection datasets in three base formats: COCO, Pascal VOC, and YOLO. In CURDD, the dataset is divided into two levels: level 0 and level 1, corresponding to the "Cls0" and "Cls1" catalogs, respectively. Level 1 includes cavity, detachment, water-rich, looseness, and background. Level 0 categories combine the four main disease types mentioned earlier into a single "diseases" category, with the other category being "background." This dataset was jointly published by Hebei University and the 519 Team of North China Geological Exploration Bureau. We support individuals or teams using the data for research purposes. We also welcome collaboration for commercial use. For commercial inquiries, please contact us for authorization.

本研究发布了两类城市道路病害数据集：用于道路病害检测的DURDD数据集，以及用于道路病害分类的CURDD数据集。DURDD涵盖四类主要的地下道路病害：空洞、脱空、富水与松散。该数据集同时提供三种标准标注格式的病害检测数据：COCO、Pascal VOC及YOLO格式。CURDD数据集分为两级：0级与1级，分别对应“Cls0”与“Cls1”目录。1级类别包含空洞、脱空、富水、松散及背景类。0级类别则将前述四类主要病害合并为单一的“病害”类，其余类别为“背景”。本数据集由河北大学与华北地质勘查局519队联合发布。我们支持个人或团队将该数据集用于科研用途，同时欢迎就商业使用开展合作。若需进行商业使用咨询，请联系我方获取授权。

搜集汇总

背景与挑战

背景概述

URDD是一个城市道路病害检测与分类的开放数据集，包含检测数据集DURDD（四种主要病害类型，支持COCO、Pascal VOC、YOLO格式）和分类数据集CURDD（两个分类级别）。该数据集由河北大学等机构联合发布，支持研究用途。