Motorcycle Safety Violations in Heterogeneous Traffic from Kozhikode, India

1.4 GB of annotated traffic surveillance frames targeting helmet non-compliance and multi-passenger overloading in high-density urban scenes. The dataset was created by Arun Natarajan with permission from the Kozhikode District Police and last updated on 2026-05-24. Data is formatted for YOLO object detection frameworks, with number plates blurred for privacy.

Use Cases

• Training object detection models to identify helmet non-compliance based on the described safety anomalies.
• Developing models for detecting multi-passenger overloading ('triple riding') in dense traffic scenes.
• Benchmarking computer vision algorithms under challenging conditions like severe occlusions and diverse illumination mentioned in the description.
• Researching class distribution and model performance in long-tail scenarios typical of real-world urban traffic.

Strengths

• Data is 1.4 GB in size, indicating a medium-scale collection of images.
• Annotations were manually bounded and verified for high precision using the CVAT platform.
• Dataset is structured in the standard YOLO Object Detection format for seamless integration with deep learning frameworks.
• Captures real-world challenges such as multi-object occlusions and diverse illumination profiles.

Limitations

• Row count is unknown, which may limit suitability assessment for specific model training needs.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data reflects geographic bias inherent to its source, being captured solely in Kozhikode District, Kerala, India.

Provenance

Source

Raw keyframes captured from high-angled urban surveillance and traffic monitoring ANPR systems.

Collection Method

Frames were captured across multiple primary junctions within Kozhikode District, Kerala, India, and annotated using CVAT.

Time Range

null

Freshness

Last updated 2026-05-24 11:29:09; freshness should be verified.

Geography

Kozhikode District, Kerala, India