PLANT-DISEASE-DETECTION

---

🌿 Plant Disease Detection using Custom CNN

This project presents a Convolutional Neural Network (CNN)-based approach to classify plant diseases from leaf images. It’s built using PyTorch and trained on a dataset consisting of healthy and diseased leaves of Apple, Corn, and Grape.

---

📁 Dataset

The dataset is divided into:

• Train Folder: Contains training images, categorized into subfolders based on class names.
• Test Folder: Contains unseen test images for evaluating model performance.

🏷️ Class Labels

['Apple___Apple_scab', 
 'Apple___Black_rot', 
 'Apple___Cedar_apple_rust', 
 'Apple___Healthy', 
 'Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot', 
 'Corn_(maize)___Northern_Leaf_Blight', 
 'Grape___Black_rot']

---

🧠 Model Architecture

A custom CNN model was designed with:

• 2 Convolutional Layers: Each followed by ReLU activation and MaxPooling.
• Dropout Layer: Applied after the first fully connected layer to prevent overfitting.
• 2 Fully Connected Layers: Final output mapped to the number of disease categories.

---

🔧 Hyperparameters

Parameter	Value
Image Size	150x150
Batch Size	32
Learning Rate	0.001
Epochs	5
Optimizer	Adam
Loss Function	CrossEntropy

---

🔄 Workflow

📦 Preprocessing

• Images resized to 150x150
• Normalized using ImageNet means and std
• Converted to tensors using torchvision transforms

🚂 Training

• Model trained for 5 epochs
• Tracked training loss every 100 batches

🧪 Evaluation

• Model tested on unseen images from the test set
• Final accuracy reported as:

  Accuracy of the network on the test images: 75%
  

---

📊 Visualization

Utility functions are included to:

• Display random images with predicted labels
• Plot dataset distribution

Example:

display_random_image(train_loader, class_names)

---

📦 Libraries Used

This project utilizes the following Python libraries:

• PyTorch – For building and training the custom CNN
• Torchvision – For loading datasets and applying image transformations
• NumPy – For numerical operations
• Matplotlib – For data visualization
• PIL (Pillow) – For image loading and processing
• os – For file path handling
• random – For random selection of images
• Google Colab (optional) – For GPU-accelerated training in the cloud

---

▶️ How to Run

1. Clone or download the repository
2. Upload dataset to the mentioned path (Google Drive or local directory)
3. Open PLANT_DISEASE_CLASSIFICATION .ipynb in Google Colab or Jupyter Notebook
4. Run cells sequentially

---

🏆 Results

• Test Accuracy Achieved: ~75%
• Dataset Size: 106 training, 106 testing images
• Balanced training observed with no severe overfitting

---

🙏 Acknowledgements

• PyTorch & Torchvision
• PlantVillage dataset (via Kaggle)
• Google Colab for training support

---