Road Anomaly Detection

The dataset for this study was compiled from various sources to ensure a diverse and representative collection of road anomalies. Primary data were gathered by recording videos and capturing images using mobile cameras and surveillance systems in different cities across Pakistan. Additionally, footages were obtained through collaboration with friends, who provided videos from social media platforms and other online sources, including contributions from different countries. This comprehensive approach allowed for the creation of a dataset that encompasses a wide range of scenarios and conditions pertinent to the study's focus on road anomalies.

本研究所用数据集源自多类数据源，旨在构建覆盖全面且具有代表性的道路异常样本集合。核心数据通过在巴基斯坦境内不同城市使用移动摄像头与监控系统录制视频、采集图像的方式获取。此外，研究团队还通过与友人合作获取素材，后者提供了来自社交媒体平台及其他在线渠道的视频内容，涵盖不同国家的相关样本。这种多源整合的全面采集方案，最终构建出涵盖本研究聚焦的道路异常相关的多样化场景与工况的数据集。