Multi-Spectral Band Visualization Before Cloud Detection

This project focuses on visualizing multi-spectral satellite imagery data and detecting clouds using a rule-based masking approach. The output includes RGB composites, NDVI maps, cloud masks, and cloud-masked RGB/NDVI images for five satellite scenes.

---

🛰️ Project Overview

• Goal: Understand how clouds appear in different spectral bands and apply a basic cloud detection algorithm.
• Tech Stack: Python, NumPy, Matplotlib, Google Colab
• Data Type: Multi-spectral satellite imagery (13 bands per scene)
• Scenes Used: 5 distinct satellite image scenes

---

📁 Dataset

• Five .npy files representing multi-spectral scenes.
• Each file contains a (13, H, W) array: 13 spectral bands for a given scene.
• Uploaded and processed in Google Colab.

---

🧪 Processing Steps

1. Load multi-spectral .npy files
2. Normalize key spectral bands (Blue, Green, Red, NIR, SWIR1)
3. Generate RGB composites
4. Calculate NDVI
5. Detect clouds using a simple threshold-based rule:

   cloud_mask = (
       (Blue > 0.25) &
       (Green > 0.25) &
       (Red > 0.25) &
       (NIR > 0.25) &
       (SWIR1 < 0.3)
   )

6. Visualize and save:
   • Original RGB
   • NDVI
   • Cloud masks
   • RGB & NDVI with cloud masking applied

---

🖼️ Output Files (per scene)

• sceneX_rgb.png: Original RGB
• sceneX_ndvi.png: Original NDVI
• sceneX_cloud_mask.png: Detected cloud mask
• sceneX_rgb_masked.png: RGB with clouds masked (gray)
• sceneX_ndvi_masked.png: NDVI with cloud areas set to NaN

---

💡 Learnings

• Clouds are bright in visible/NIR but not in SWIR
• Simple rule-based masking works surprisingly well for clear/cloudy scenes
• Visualizing different spectral bands is crucial for remote sensing analysis

---

📦 How to Run (Google Colab)

1. Upload all .npy files to a folder
2. Load them into a list using np.load()
3. Follow the notebook code to process each scene
4. Visualize and export images using matplotlib

---

📬 Contact

Vinit Singh Pathir
LinkedIn
Feel free to reach out if you want to collaborate on remote sensing, data science, or climate projects!

---

🌍 License

This project is open-source and free to use under the MIT License.

🚀 Thank you for visiting the project!