Garbage Detection Dataset
This dataset contains annotated images for the detection of objects related to waste sorting, classified into six categories (biodegradable, cardboard, glass, metal, metal, paper, plastic). Annotations are compatible with YoLov5 and make it easy to train vision models in real time.
Around 20,900 files (JPG images + TXT labels), organized by train/valid/test
Attribution 4.0 International (CC BY 4.0)
Description
Garbage Detection Dataset includes annotated images for the classification and detection of waste-related objects in six categories. The data is organized into files for training, validation and testing, with annotations in TXT format compatible with YoLov5.
What is this dataset for?
- Train object detection models to automatically sort waste.
- Develop intelligent sorting and recycling systems in real time.
- Test and improve YOLO algorithms and other computer vision methods.
Can it be enriched or improved?
Yes, this dataset can be enriched with new images representing more types of waste, in various environments, or with additional annotations (e.g. segmentation, depth).
🔎 In summary
🧠 Recommended for
- Machine vision projects
- Environmental AI
- Automated sorting systems
🔧 Compatible tools
- YoloV5
- PyTorch
- TensorFlow
- OpenCV
💡 Tip
Consider using image augmentation techniques to improve the robustness of models in different contexts.
Frequently Asked Questions
What waste categories are included in this dataset?
Six categories: biodegradable, cardboard, glass, glass, metal, paper and plastic.
Are annotations compatible with YoLov5?
Yes, annotations are provided in the standard TXT format used by YoLov5.
Is this dataset suitable for a commercial project?
Yes, the CC BY 4.0 license allows commercial use with attribution.
