Multiclass Weather Dataset
Meteorological image dataset containing 5 distinct categories and a mixed test file. The images are accompanied by a CSV file listing the labels of the test images. Adapted for multi-class visual recognition in the meteorological field.
1,531 images divided into 5 category folders + test folder, with CSV labels
CC0: Public Domain
Description
The “Multiclass Weather Dataset” dataset contains 1,531 images divided into 5 main classes, each in its own folder, as well as a test folder containing a mixture of categories with a CSV file for the labels. This dataset is designed for the classification of weather images into distinct categories.
What is this dataset for?
- Train image classification models for different types of weather conditions.
- Develop visual weather analysis systems based on satellite or terrestrial images.
- Testing multi-class recognition algorithms in computer vision.
Can it be enriched or improved?
The dataset can be enriched by adding additional images for each category, improving annotations with temporal or geographic metadata, or providing image segmentations for finer analyses.
🔎 In summary
🧠 Recommended for
- Computer vision students
- Researchers in applied meteorology
- AI developers
🔧 Compatible tools
- TensorFlow
- PyTorch
- Keras
- OpenCV
💡 Tip
Use increases in weather images (blur, contrast) to improve the robustness of the model.
Frequently Asked Questions
What are the weather categories present in this dataset?
The dataset includes 5 distinct categories, each represented by an image folder.
Can this dataset be used for real-time recognition?
Yes, it can be used to train lightweight models for real-time weather classification applications.
Are the images annotated with temporal or geographic information?
No, annotations are basic, only class labels per image, without temporal or geographic metadata.
