Road Static Obstacle Image Dataset

This dataset provides a comprehensive collection of static road obstacle images, generated using high-fidelity synthetic simulation environments. It covers common roadside and lane-based traffic facilities such as water barriers, traffic cones, no-parking signs, toll gates, and other fixed infrastructure elements. Each object type is accurately represented with realistic dimensions, placement, and visual characteristics to closely match real-world traffic scenarios.

The dataset is designed for autonomous driving perception systems, High-Definition (HD) map construction, and road safety monitoring. It supports static obstacle detection tasks for Advanced Driver Assistance Systems (ADAS), enabling improved scene understanding, route planning, and hazard recognition in diverse road environments.

All samples are annotated with object detection bounding boxes, provided in JSON or COCO-compatible formats. Label categories include water barriers, traffic cones, no-parking signs, toll gates, and additional traffic-related fixtures. Annotation ensures balanced representation of each category to achieve high recognition accuracy across varied operational conditions.

The synthetic generation process allows controlled variation in environmental conditions such as different lighting, object density, and background complexity. It also replicates multiple road types, including urban streets, industrial park roads, and facility-dense environments, to ensure model robustness and adaptability to real-world deployments.

By combining photorealistic rendering, human annotation, and diverse simulated scenarios, this dataset supports the development of accurate static obstacle detection models, enhances HD map alignment, and strengthens road scene interpretation capabilities for intelligent transportation systems.

Keywords: road obstacle detection traffic facility detection barrier cone signage tollgate infrastructure

Sector: Intelligent Transportation / Autonomous Driving / HD Map Construction / Traffic Safety Monitoring

Metrics: Object detection accuracy, category-wise recall rate

Entities: Water barriers, traffic cones, no-parking signs, toll gates, traffic infrastructure

Geographies: China (urban roads, industrial parks, and facility-dense environments)

Price Range (USD): 4000–20000

Source type: Synthetic data generation from simulation environment, human annotation