GeoVision

GeoVision is a computer vision-based geological analysis system developed using Python. The project analyzes rock and lithological images to detect cracks, perform grain-level segmentation, and classify rock properties using machine learning techniques.

---

🚀 Features

• 🔍 Crack / Fault Detection

  • Detects cracks using edge detection and contour analysis
  • Estimates real-world crack sizes from pixel measurements

• 🧱 Grain Segmentation

  • Uses watershed algorithm for separating grains
  • Extracts features such as area, orientation, diameter, and intensity

• 🧠 Rock Classification

  • CNN-based classifier for multi-class rock identification
  • Supports 50+ rock types

• ✨ Luster Classification

  • Texture-based classification using Local Binary Patterns (LBP) and SVM

• 🧪 Sandstone Segmentation

  • Pixel-wise classification using Random Forest with handcrafted features

• 🌐 Web Interface

  • Flask API for backend processing
  • Streamlit UI for user interaction

---

🏗️ Project Structure

geovision/
│
├── app/
│   ├── flask_api.py
│   └── streamlit_app.py
│
├── src/
│   ├── segmentation/
│   ├── faults/
│   ├── classification/
│   ├── analysis/
│
├── data/
├── experiments/
├── notebooks/
│
├── requirements.txt
├── README.md
└── .gitignore

---

⚙️ Installation

1. Clone the repository:

git clone https://github.com/git-authority/GeoVision.git
cd GeoVision

2. Install dependencies:

pip install -r requirements.txt

---

▶️ Usage

Run Streamlit UI

streamlit run app/streamlit_app.py

Run Flask API

python app/flask_api.py

---

📊 Dataset

This project uses:

• Sandstone tomography dataset for segmentation

• Custom datasets for:

  • Rock classification
  • Luster classification
  • Cementing material analysis

⚠️ Datasets are not included due to size constraints.

Place datasets inside the data/ directory.

---

🧠 Technologies Used

• Python
• OpenCV
• Scikit-learn
• TensorFlow / Keras
• Flask
• Streamlit
• NumPy / Pandas
• Scikit-image

---

📌 Notes

• Some scripts in experiments/ are exploratory and not part of the main pipeline
• File paths have been normalized for portability
• Models can be stored in the models/ directory

---

👤 Author

Developed as part of Smart India Hackathon (SIH) 2023 project.

---

⭐ Future Improvements

• Integrate mobile frontend (Flutter)
• Improve model accuracy and deployment
• Add real-time API integration
• Optimize performance for large-scale datasets

---