What is the Underwater Bounding Box Dataset?

The Underwater Bounding Box Dataset is a curated collection of underwater images with bounding box annotations for marine objects such as fish, crabs, coral reefs, jellyfish, trash, and human divers. It supports AI training and marine research applications.

What are the applications of the Underwater Bounding Box Dataset?

This dataset can be used for underwater object detection, marine ecosystem monitoring, waste detection, autonomous underwater vehicles, and AI research in environmental conservation.

What annotation format does the dataset use?

The dataset provides bounding box annotations compatible with popular deep learning frameworks like YOLO, TensorFlow, and PyTorch.

Who can benefit from using this dataset?

Researchers, environmental scientists, data scientists, and AI developers working on marine biology, underwater robotics, and waste detection systems can benefit from this dataset.

Underwater Bounding Box Dataset

Home » Dataset Download » Underwater Bounding Box Dataset

Underwater Bounding Box Dataset

Datasets

Underwater Bounding Dataset

File

Underwater Bounding Dataset

Use Case

Underwater Bounding Dataset

Description

Explore the Underwater Bounding Dataset featuring high-quality images with bounding box annotations for Fish, Crab, Human, Trash, Jellyfish, Coral Reef, and Fish Group.

Description:

The Underwater Bounding Dataset is designed for accurate object localization in underwater environments. It features high-quality images with bounding box annotations for seven classes: Fish, Crab, Human, Trash, Jellyfish, Coral Reef, and Fish Group. The dataset is split into 70% for training, 10% for validation, and 20% for testing.

Download Dataset

Annotation and Pre-processing

Images are annotated using Roboflow with the following pre-processing steps:

Auto-Orient
Resize to 640×640
Contrast Stretching

Sources

Brackish Dataset (Aalborg University)
Trash Debris Dataset (University of Minnesota)

Purpose

This dataset aims to advance research in underwater object detection and localization, supporting applications in marine ecology, pollution detection, and underwater robotics. The dataset includes a diverse range of underwater scenes captured in various conditions, such as different water depths, lighting conditions, and levels of turbidity. It contains high-resolution images and videos annotated with precise bounding boxes and segmentation masks for various objects of interest, including marine life, coral reefs, submerged debris, and man-made structures.

This dataset is sourced from Kaggle.

Contact Us

Let's Discuss your Data collection Requirement With Us

To get a detailed estimation of requirements please reach us.

Underwater Bounding Box Dataset