Satellite Imagery & Analysis with PACE OCI

This project demonstrates how to load, explore, and visualize NASA PACE/OCI Level-1C data in Python/Jupyter. You’ll open a NetCDF file, browse metadata, plot single-band images and RGB composites, inspect pixel spectra, project imagery on a map, and compute simple indices.

📓 Open Notebook

---

Features

• Open PACE/OCI NetCDF (.nc) files with netCDF4
• Inspect dataset groups/variables and wavelengths
• Plot per-band images and a 3-band RGB composite
• Extract and plot spectra from selected pixels
• Map projection with Cartopy (lat/lon)

---

Process

1) Get the data

Place a PACE/OCI Level-1C file in (or adjust path inside the notebook):

../data/PACE_OCI.20240918T171411.L1C.V3.5km.nc
nc

2) Environment setup

pip

# Windows (Command Prompt or PowerShell)
py -m pip install --upgrade pip
pip install jupyterlab numpy matplotlib netCDF4 cartopy pyproj shapely

3) Launch Jupyter

# from your project folder
py -m jupyter lab

Open SatelliteImagery_PACE-OCI_Projects.ipynb.

---

How the Notebook is Organized

1. Imports / Setup Uses: numpy, matplotlib, netCDF4, cartopy (+ ccrs), plus small utility functions:

   • reflectance(rad, sza, f0) — converts radiance to reflectance:

     $$ \text{reflectance} = \frac{\pi \cdot \text{radiance}}{\cos(\text{sza}) \cdot F_0} $$

   • color3_make(ch1, ch2, ch3) — builds 3-band color composites.

2. Loading an Image Opens the NetCDF dataset and explores groups/variables, e.g.:

   • /geolocation_data/longitude, /geolocation_data/latitude
   • /sensor_views_bands/intensity_wavelength
   • radiance arrays and attributes

3. Wavelengths & Radiances Lists available wavelengths, selects bands, and sets display ranges.

   

4. Plotting

   • Single-band quicklooks
   • RGB composite (choose 3 bands → visualize)

5. Pixel Spectra Select pixel(s) by (x, y) and plot spectrum across wavelengths.

6. Map Projection Uses Cartopy with lat/lon to render georeferenced imagery

---

Helpful insights

• Pick bands thoughtfully for RGB (e.g., a red-ish, green-ish, and blue-ish channel).
• If imagery looks blown out, adjust radiance/reflectance scales in the plotting cells.
• For pixel spectra, verify (row, col) order and tilt/indexing match your array shapes.
• For map plots, ensure your latitude/longitude arrays have the same spatial shape as your imagery.

---

---

Final Presentation of Wildfires Spectra Reflectance Analysis

📑 View the full presentation (PDF)

---

Acknowledgments

• I would like to thank Brent Mcbride, Post Doc at UMBC for helping me with the code and analysis.
• Open-source Python ecosystem: NumPy, Matplotlib, netCDF4, Cartopy