Soda Can Asta: Underwater Object Detection Test Set with BlueROV2 Footage

A first-person underwater video test set captured with a BlueROV2 Heavy ROV. The footage was recorded across 15 distinct motion sequences and is annotated in YOLO format. The dataset was created by author rifqijuli and last updated on 2026-05 25.

Use Cases

• Benchmarking model generalization based on real-world ROV footage mentioned in the description
• Evaluating object detection performance across different motion sequences like forward, yaw, and ascend
• Testing cross-dataset compatibility based on annotations relabelled to match four established underwater datasets

Strengths

• Contains footage from 15 distinct motion sequences
• Annotations are in YOLO format and relabelled for compatibility with four established datasets

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment

Provenance

Source

huggingface

Collection Method

Footage captured with a BlueROV2 Heavy ROV and annotated in YOLO format.

Time Range

null

Freshness

Last updated 2026-05-25 21:39:12; freshness should be verified

Geography

null