Adrishyam 🖼️

A Python package for image dehazing using the Dark Channel Prior algorithm.

Package Link: Pypi

💡 Key Features

• Implements the Dark Channel Prior algorithm for effective image dehazing.
• Supports configurable parameters for advanced users.
• Outputs intermediate steps for better visualization of the dehazing process.

---

📦 Installation

Install Adrishyam via pip:

pip install adrishyam

---

🚀 Usage

🔧 Basic Usage

Dehaze an image by providing input and output paths:

from adrishyam import dehaze_image

dehaze_image(
    input_path="path/to/hazy/image.jpg",
    output_dir="path/to/output/directory"
)

⚙️ Advanced Usage

Customize dehazing parameters for fine-tuned results:

dehaze_image(
    input_path="path/to/hazy/image.jpg",
    output_dir="path/to/output/directory",
    t_min=0.1,  # Minimum transmission value (default: 0.1)
    patch_size=15,  # Size of the local patch (default: 15)
    omega=0.95,  # Dehazing strength (default: 0.95)
    radius=60,  # Filter radius for guided filter (default: 60)
    eps=0.01,  # Regularization parameter (default: 0.01)
    show_results=False  # Whether to display results (default: False)
)

---

📂 Output

Adrishyam generates step-by-step outputs in your specified output_dir:

• original.png ➡️ Original hazy image.
• dark_channel.png ➡️ Dark channel visualization.
• transmission.png ➡️ Estimated transmission map.
• refined_transmission.png ➡️ Refined transmission map.
• dehazed.png ➡️ Final dehazed image.
• result.png ➡️ Combined visualization of all processing steps.

---

🔍 Results

Example Outputs from Adrishyam:

Original

Dehazed

---

📜 License

Adrishyam is licensed under the MIT License. 📝
Feel free to use and contribute!