🛰️ Advanced Deep Learning for Aerial & Satellite Image Object Detection (AeroVision-AI)

This project focuses on applying advanced deep learning techniques to automatically detect objects from aerial and satellite imagery. It leverages powerful Convolutional Neural Network (CNN) architectures such as ManualNet, Xception, and DenseNet for accurate image classification and detection.

---

🚀 Features

• 📷 Automated object detection from aerial/satellite images
• 🧠 Implementation of multiple deep learning models:
  • ManualNet (custom CNN)
  • Xception
  • DenseNet
• 📊 Model training and evaluation using Jupyter Notebooks
• 🌐 Web integration using Django framework
• 💾 Pre-trained model support (.h5 file included)
• 📁 Organized project structure for scalability

---

🛠️ Tech Stack

• Python 🐍
• TensorFlow / Keras
• OpenCV
• Django (Web Framework)
• NumPy, Pandas, Matplotlib

---

🧠 Models Used

🔹 ManualNet

A custom-built CNN architecture designed specifically for this dataset.

🔹 Xception

A deep CNN architecture based on depthwise separable convolutions for high performance.

🔹 DenseNet

A densely connected CNN that improves feature propagation and reduces vanishing gradient problems.

---

📊 Workflow

1. Data Collection & Preprocessing
2. Model Training (ManualNet, Xception, DenseNet)
3. Model Evaluation
4. Model Deployment using Django
5. User uploads image → Model predicts object

---

---

▶️ How to Run

1. Clone the repository

git https://github.com/deekshwar/ADVANCED-DEEP-LEARNING-TECHNIQUES-FOR-AUTOMATED-DETECTION-OF-AERIAL-AND-SATELLITE-IMAGERY-OBJECTS/tree/main
 cd ADVANCED-DEEP-LEARNING-TECHNIQUES-FOR-AUTOMATED-DETECTION-OF-AERIAL-AND-SATELLITE-IMAGERY-object