ITD-UAV: Aerial Image Dataset for Maglev Track Intrusion Detection

ITD-UAV is a 3.5 GB dataset of aerial images for maglev track intrusion detection, released by Jiajing Xu. It is the first publicly available dataset for this specific task. The annotations and supplementary information are organized in the COCO format.

Use Cases

• Train object detection models for identifying intrusions on maglev tracks based on aerial imagery.
• Benchmark model performance for a specialized infrastructure monitoring task using the provided COCO annotations.
• Develop and evaluate safety systems for high-speed rail networks based on the described intrusion detection focus.

Strengths

• Dataset size is 3.5 GB, indicating a medium-scale collection of images.
• It is the first publicly released aerial-image-based dataset for maglev track intrusion detection.
• Annotations are provided in the widely supported COCO format, facilitating use with standard tools.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count (number of images/annotations) is unknown, which may limit suitability assessment.
• The geographic and temporal coverage of the aerial imagery is not specified in the provided metadata.

Provenance

Source

Jiajing Xu

Collection Method

Likely gathered via unmanned aerial vehicle (UAV) photography, as suggested by the dataset name and description.

Freshness

Last updated 2026-06-03 04:41:51; freshness should be verified.