This repository contains a complete implementation of a state-of-the-art plant disease detection system using YOLOv11x trained on 23,689 images across 116 disease classes covering major food crops. The model combines two complementary datasets — controlled laboratory images and real-world field images — to achieve strong generalization across domains.
1 # Clone repository
2 git clone https://github.com/SIV-TK/A-YOLOv11x-Benchmark-on-116-Classes-Using-Combined-Laboratory-and-Field-Images.git
3 cd A-YOLOv11x-Benchmark-on-116-Classes-Using-Combined-Laboratory-and-Field-Images
4
5 # Install dependencies
6 pip install ultralytics opencv-python matplotlib seaborn scikit-learn pandas pyyaml tqdm kagglehub
7
8 # Run the training notebook
9 jupyter notebook "YOLO(2).ipynb"
1 from ultralytics import YOLO
2
3 # Load model
4 model = YOLO ( 'best.pt' ) # or load from https://huggingface.co/JK-TK/PlantDiseaseDetection
5
6 # Run inference
7 results = model ( 'path/to/leaf_image.jpg' , conf = 0.25 )
8
9 # Display results
10 results [ 0 ] . show ( )
11
12 # Print detections
13 for box in results [ 0 ] . boxes :
14 class_id = int ( box . cls [ 0 ] )
15 confidence = float ( box . conf [ 0 ] )
16 class_name = model . names [ class_id ]
17 print ( f"Detected: { class_name } ( { confidence : .2% } )" )
The notebook automatically saves checkpoints. To resume an interrupted training run:
1 from ultralytics import YOLO
2
3 model = YOLO ( '/content/drive/MyDrive/PlantDisease_TrainingBackup_*/training_run/weights/last.pt' )
4 model . train ( data = 'dataset/data.yaml' , epochs = 100 , resume = True )
1 from google . colab import files
2 from ultralytics import YOLO
3
4 model = YOLO ( 'best.pt' )
5 uploaded = files . upload ( )
6
7 for img_name in uploaded :
8 results = model ( img_name , conf = 0.25 )
9 results [ 0 ] . show ( )
├── YOLO(2).ipynb # Complete training notebook
├── README.md # This file
├── LICENSE # MIT License
├── best.pt # Pre-trained weights (69.8% mAP50)
├── dataset/
│ ├── train/images/ # 16,582 training images
│ ├── train/labels/
│ ├── val/images/ # 4,737 validation images
│ ├── val/labels/
│ ├── test/images/ # 2,370 test images
│ ├── test/labels/
│ └── data.yaml # Dataset configuration
└── publication/
├── comparison_bar_chart.png
├── training_curves.png
├── confusion_matrix.png
├── dataset_source_pie.png
├── dataset_split_bar.png
├── overall_performance.csv
└── comparison_table.csv1 @software{kariuki_2026_plant_disease,
2 author = {Kariuki, James Kariuki},
3 title = {Multi-Domain Plant Disease Detection: A YOLOv11x Benchmark on 116 Classes},
4 year = {2026},
5 url = {https://github.com/SIV-TK/A-YOLOv11x-Benchmark-on-116-Classes-Using-Combined-Laboratory-and-Field-Images/tree/main/YOLO.ipynb},
6 note = {mAP50: 69.8\%, 116 disease classes, 23,689 images}
7 }
This project is licensed under the
MIT License .
For questions, issues, or collaboration requests, please
open an issue on GitHub.
