Cryomalus: Apple Orchard Management System

Cryomalus is a comprehensive system for monitoring and managing apple orchards using drones and computer vision. It delivers real-time insights into tree health, nutrient levels, pest infestations, and production estimates, combining advanced data collection and analysis to optimize orchard management.

Key Features

• Thermal Anomaly Detection: Utilizes a custom Convolutional Neural Network (CNN) trained in PyTorch to detect thermal anomalies in apple trees, predicting the coordinates of anomalies in thermal images.
• Nutrient Monitoring: Tracks and displays nutrient levels such as nitrogen, phosphorus, and potassium, providing detailed insights into soil conditions and tree health.
• Pest Control: Monitors and reports pest counts, including aphids, mites, and worms, helping manage pest-related issues effectively.
• Production Estimation: Estimates the orchard's yield based on data collected from drone surveys, offering accurate predictions for planning and resource allocation.

Project Modules

1. Thermal Anomaly Detection with CNN

• Files: neural_network_files/train_model.py, thermal_cnn.py
• Technologies: PyTorch, pandas, numpy
• Highlights:
  • CNN Architecture: Defines a robust CNN model for thermal anomaly detection.
  • Training & Inference: Implements a training script and inference pipeline to detect and predict anomaly locations in thermal images.

2. Dashboard for Real-Time Orchard Monitoring

• Files: pages/index.js, components/Dashboard.js
• Technologies: React, Next.js, Mapbox, TailwindCSS
• Highlights:
  • Map Visualization: Integrates Mapbox for a geographic representation of the orchard.
  • Data Tables: Displays tree health, pest counts, and nutrient levels in a user-friendly interface.
  • Production Estimates: Shows current yield projections, aiding in efficient orchard management.

3. Apple-Classifier - Computer Vision Module

• Files: main.py, apple_classifier.py
• Technologies: PyTorch, Transformers, Pillow (PIL), Matplotlib
• Highlights:
  • Early Disease Detection: Employs machine learning to identify diseases from drone-captured images.
  • Bounding Box Visualization: Highlights regions of interest in images for detailed analysis.
  • Dense Region Captioning: Provides comprehensive descriptions of specific regions, enhancing understanding and decision-making.

Other Features

• Drone Autopilot Integration: Implement automated flight paths for efficient data collection.
• Real-Time Notifications: Develop a notification system for critical issues detected in the orchard.
• Model Optimization: Optimize models for faster inference on edge devices, improving overall system efficiency.

Cryomalus represents a significant advancement in orchard management, leveraging cutting-edge technology to deliver actionable insights and enhance decision-making for improved yie