A Real-World Road Dataset For Manhole and Speed Breaker Detection in Bangladesh

This dataset contains real-world road images from Bangladesh, focusing on the detection of manholes and speed breakers. The images were captured under different lighting and road conditions to reflect real traffic environments. The dataset is intended for training and testing computer vision and machine learning models for road safety and infrastructure detection. The dataset is divided into five distinct folders, each representing the dominant road object appearing in the images, such as manholes and speed breakers. 1. Uncovered_Manhole: Road images showing uncovered or open manholes without protective covers. 2. Square_Manhole: Images containing square-shaped manholes on the road surface. 3. Speed_Breaker: Road images showing speed breakers or road humps. 4. Good_Manhole: Images of properly covered and intact manholes in good condition. 5. Broken_Manhole: Road images showing damaged or broken manholes.本数据集收录了来自孟加拉国的真实道路影像，聚焦于检查井（manhole）与减速带（speed breaker）的检测任务。所有影像均在多样化光照与道路环境下采集，以还原真实的交通场景，专为训练与测试计算机视觉及机器学习模型、服务于道路安全与基础设施检测领域而打造。 本数据集共分为5个独立文件夹，每个文件夹对应影像中的主导道路目标，具体分类如下： 1. 无盖检查井（Uncovered_Manhole）：展示无防护盖板、裸露或敞开状态的检查井的道路影像。 2. 方形检查井（Square_Manhole）：收录路面带有方形检查井的道路影像。 3. 减速带（Speed_Breaker）：展示减速带或道路隆起带的道路影像。 4. 完好检查井（Good_Manhole）：收录盖板完好、状态完整的正常检查井道路影像。 5. 破损检查井（Broken_Manhole）：展示存在损坏或破损情况的检查井的道路影像。