Crop Disease Detection

Crop disease classification for indoor farming.

CLI

Command Line Interface used to generate the dataset and plot class distribution.

You have the possibility to choose between 4 types of labels:

• plants
• diseases
• general_diseases
• healthy (binary)

You can also remove the background in the images with different techniques:

• Watershed
  
• HSV filters & Morphological transforms
  
• K-Means Clustering ...
  

python cli/cli.py path_to_images_folder

Training framework

Framework to train different computer vision models (CNN & transformers) in Tensorflow for crop disease classification.
The classification task can either be multiclass or binary.

Training parameters are in the file configs/train_config.yml.
Evaluation parameters are in the file configs/infer_config.yml.
Available models for training are in the file resources/models_to_eval.json

Training script:

python run_training.py --config configs/train_config.yml

Inference script:

python run_infererence.py --configs configs/infer_config.yml

Gradio App (Demo):

python app.py

Report on: