README.md

---
annotations_creators:
  - expert-generated
language:
  - en
license: cc-by-sa-4.0
multilinguality: monolingual
pretty_name: Moon Detection Dataset for YOLOv8
size_categories:
  - 1K<n<10K
source_datasets:
  - original
task_categories:
  - object-detection
---


# Moon Detection Dataset for YOLOv8
This dataset was developed as part of the CubeRT-02 project, a CubeSat mission aimed at testing AI-powered vision systems in aerospace contexts. It consists of 7500+ annotated images for object detection of the Moon, optimized for use with the YOLOv8 architecture.

## 📸 Dataset Collection & Annotation
**Initial set**: ~400 web-sourced images used for theoretical model exploration.

**Main dataset**: 7500+ images captured over 6 months using smartphones and amateur camera devices, reflecting real-world scales, perspectives, and conditions.

**Cleaning**: Manual filtering of blurry, low-quality, or irrelevant images.

**Annotation**: Performed via Roboflow platform with bounding boxes for YOLOv8.

**Augmentation**: Basic augmentations applied during preparation; none used during final training due to negative effects on performance.

## 📁 Dataset Structure

This dataset follows the YOLOv8 format. A Python script and a YAML configuration file are included to help you easily train or test the dataset using Ultralytics' YOLOv8 implementation.

## Authors

Luis Adrian Cabrera | https://github.com/LuisAdrian5519

Jesse Banda Chaidez | https://github.com/Jessebnda

## Contributors
José Alejandro Padilla Pérez | https://github.com/PadillaPepe777

Erick Blanco Nakashima

Juan Pablo Riojas Pesqueira

Guillermo Villegas

Juan Adrian Astorga

Julio Castañeda

Juan Pablo Aboytes

Maximo Millán Cabrera