AgriVision Pro is a desktop-based machine learning application designed to automatically detect and classify plant diseases from leaf images. The system uses a hybrid approach combining unsupervised image segmentation and supervised classification.
-
Image Preprocessing
- Resize images to 256×256 pixels
-
Segmentation
- K-Means clustering (k=2) to isolate diseased regions
-
Feature Extraction (10 Features)
- Color: Mean & Std of RGB (6)
- Texture: GLCM Contrast, Correlation, Energy (3)
- Shape: Diseased area ratio (1)
-
Classification
- Random Forest Classifier (100 trees)
-
GUI
- Desktop application built using CustomTkinter
- Algorithm: Random Forest Classifier
- Training/Test Split: 80/20
- Dataset: PlantVillage
- Evaluation Metrics: Accuracy, Precision, Recall, F1-score
pip install -r requirements.txt

