jameelkhalidawan/Plant_Detection_Classification

Plant Species Classification Dataset

A comprehensive dataset containing 64 different plant species with high-quality images for machine learning and computer vision applications. Model Trainig code and trainned model with detailed performance analysis is present on the Github https://github.com/jameelkhalidawan/Plant-Detection-Model-using-Yolo

Dataset Overview

This dataset is designed for plant species classification tasks and contains images of various plant species organized in a structured format suitable for training deep learning models.

Key Statistics

• Total Classes: 64 plant species
• Total Images: 152,042 images
• Image Format: JPG
• Dataset Split:
  • Training: 106,395 images
  • Validation: 22,779 images
  • Test: 22,868 images

Dataset Structure

Plants_Datadet/
├── train/                    # Training data (106,395 images)
│   ├── [Plant Species 1]/    # Each folder contains images of one species
│   ├── [Plant Species 2]/
│   └── ...
├── val/                      # Validation data (22,779 images)
│   ├── [Plant Species 1]/
│   ├── [Plant Species 2]/
│   └── ...
├── test/                     # Test data (22,868 images)
│   ├── [Plant Species 1]/
│   ├── [Plant Species 2]/
│   └── ...
├── train_split/              # Additional split for training (generated by training code)
│   ├── train/
│   ├── val/
│   ├── train.cache
│   └── val.cache
└── image_counts.xlsx         # Detailed image count statistics

Plant Species Included

The dataset contains 64 diverse plant species including:

Trees and Shrubs

• Acacia dealbata Link
• Liriodendron tulipifera L
• Nandina domestica Thunb
• Pyracantha coccinea M.Roem
• Schefflera arboricola (Hayata) Merr
• Smilax aspera L
• Trachelospermum jasminoides (Lindl.) Lem
• Zamioculcas zamiifolia (Lodd.) Engl

Herbs and Wildflowers

• Aegopodium podagraria L
• Anemone alpina L
• Anemone hepatica L
• Anemone hupehensis (Lemoine) Lemoine
• Anemone nemorosa L
• Angelica sylvestris L
• Barbarea vulgaris R.Br
• Cirsium arvense (L.) Scop
• Cirsium vulgare (Savi) Ten
• Cymbalaria muralis P.Gaertn., B.Mey. & Scherb
• Dryopteris filix-mas (L.) Schott
• Epipactis helleborine (L.) Crantz
• Fragaria vesca L
• Helminthotheca echioides (L.) Holub
• Humulus lupulus L
• Hypericum androsaemum L
• Hypericum calycinum L
• Kniphofia uvaria (L.) Hook
• Lactuca serriola L
• Lamium album L
• Lamium galeobdolon (L.) L
• Lamium maculatum (L.) L
• Lamium purpureum L
• Lapsana communis L
• Lupinus polyphyllus Lindl
• Melilotus albus Medik
• Mercurialis annua L
• Nymphaea alba L
• Ophrys apifera Huds
• Pancratium maritimum L
• Papaver rhoeas L
• Papaver somniferum L
• Perovskia atriplicifolia Benth
• Trifolium incarnatum L

Succulents and Sedums

• Sedum acre L
• Sedum album L
• Sedum rupestre L
• Sedum sediforme (Jacq.) Pau

Ornamental Plants

• Anthurium andraeanum Linden ex André
• Fittonia albivenis (Lindl. ex Veitch) Brummitt
• Lavandula angustifolia Mill
• Lavandula stoechas L
• Pelargonium graveolens L'Hér
• Pelargonium inquinans (L.) Aiton
• Pelargonium zonale (L.) L'Hér
• Pelargonium zonale (L.) L'Hér. ex Aiton
• Punica granatum L
• Tagetes erecta L
• Tagetes patula L

Vegetables and Fruits

• Cucurbita maxima Duchesne
• Cucurbita pepo L

Tradescantia Varieties

• Tradescantia fluminensis Vell
• Tradescantia pallida (Rose) D.R.Hunt
• Tradescantia spathacea Sw
• Tradescantia virginiana L
• Tradescantia zebrina Bosse

Image Characteristics

• Format: JPG
• Naming Convention: Images use hash-based filenames (e.g., 0038b49e0352646885a8899be350813d927f34a5.jpg)
• Quality: High-quality images suitable for detailed plant identification
• Content: Various parts of plants including flowers, leaves, stems, and full plant views

Usage

This dataset is suitable for:

1. Plant Species Classification: Train models to identify and classify different plant species
2. Computer Vision Research: Develop and test image classification algorithms
3. Botanical Studies: Analyze plant characteristics and features
4. Educational Applications: Create learning tools for plant identification
5. Agricultural Applications: Assist in crop and weed identification

Dataset Splits

The dataset is pre-split into training, validation, and test sets to ensure proper model evaluation:

• Training Set: Used for model training and parameter optimization
• Validation Set: Used for hyperparameter tuning and model selection
• Test Set: Used for final model evaluation and performance assessment

Additional Files

• image_counts.xlsx: Contains detailed statistics about the number of images per class
• train_split/: Contains additional splits generated during the training process with cache files for faster data loading

Citation

If you use this dataset in your research or projects, please cite it appropriately and acknowledge the contributors.

License

Please check the license terms before using this dataset for commercial purposes.

---

This dataset provides a comprehensive collection of plant species images suitable for various machine learning and computer vision applications in botany, agriculture, and environmental studies.