# asteroid_detector

**Repository Path**: aster/asteroid_detector

## Basic Information

- **Project Name**: asteroid_detector
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-08-13
- **Last Updated**: 2025-08-14

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Quantum Asteroid Detector

This project uses quantum computing via Qiskit to detect asteroids in space images, demonstrating how quantum algorithms can be applied to image classification tasks.

## Key Components

1. **Image Generator**: Creates synthetic space images with and without asteroids
2. **Quantum Classifier**: Uses quantum computing principles to train a model for asteroid detection
3. **Detector**: Command-line tool to detect asteroids in new images using quantum similarity

## Requirements

- Python 3.7+
- Qiskit
- NumPy
- Matplotlib
- Pillow (PIL)
- scikit-image
- scikit-learn

Install dependencies:

```bash
pip install qiskit matplotlib numpy pillow scikit-image scikit-learn
```

## Usage

### 1. Run the Complete Pipeline

For convenience, you can run the entire pipeline with a single command:

```bash
python run_pipeline.py
```

This will generate the dataset, train the quantum model, and test it on a sample image.

### 2. Generate Training Images

Generate a dataset of synthetic space images:

```bash
python generate_images.py
```

This creates a `dataset` directory with:
- `with_asteroid`: Images containing asteroids
- `without_asteroid`: Images without asteroids

### 3. Train the Quantum Classifier

Train the quantum classifier on the generated dataset:

```bash
python quantum_asteroid_detector.py
```

This will:
- Process the dataset and reduce dimensions for quantum processing
- Train a quantum classifier using quantum similarity measurements
- Evaluate model performance and generate a confusion matrix
- Save the trained model as `quantum_asteroid_model.pkl`

### 4. Detect Asteroids in New Images

Use the trained model to detect asteroids in new images:

```bash
python detect_asteroid.py path/to/your/image.png
```

The detector will:
- Create a quantum circuit to encode the image features
- Use quantum similarity to compare with training examples
- Display the detection result with confidence score
- Show the quantum circuit used for detection

## Quantum Computing Approach

### Quantum Feature Encoding

- Images are processed and reduced to fit in a quantum system
- Features are encoded into quantum states using multiple rotation gates (RX, RY, RZ)
- Quantum superposition allows processing all feature combinations simultaneously

### Quantum Entanglement

- All-to-all connectivity between qubits creates a fully entangled quantum state
- Entanglement enables quantum advantage by creating correlations impossible in classical systems

### Quantum Similarity Measurement

- The quantum swap test algorithm measures similarity between quantum states
- This quantum algorithm provides a way to calculate inner products between quantum states
- Similarity scores determine the most similar training examples for classification

## Technical Details

- Feature dimension is determined by the number of qubits (2^n_qubits features)
- Multiple rotation gates create rich quantum state representations
- The quantum classifier uses k-nearest neighbors with quantum similarity metrics
- Confidence scores are calculated based on quantum similarity measurements

## Limitations and Future Work

- Limited number of qubits (default: 2) due to simulator constraints
- Image resolution is significantly reduced for quantum processing
- Future improvements could include:
  - Implementing variational quantum circuits for feature learning
  - Testing on real quantum hardware
  - Exploring quantum neural networks for image classification
  - Applying quantum phase estimation for enhanced feature extraction

## Visualization

The project generates several visualizations:
- `sample_images.png`: Examples of generated images
- `quantum_circuit.png`: Visualization of the quantum circuit
- `confusion_matrix.png`: Model performance evaluation
- `prediction_results.png`: Sample predictions
- `quantum_circuit_detection.png`: Circuit used for detection
- `asteroid_detection_result.png`: Detection results with confidence