Satellite Data Acquisition

Python package for downloading satellite imagery from multiple sources with a standardized API.

Full Documentation on GitHub

Features

• Multiple satellites: Sentinel-2, Sentinel-1, Landsat 8/9, HLS, Copernicus DEM
• Flexible providers: Element84, Microsoft Planetary Computer
• Smart search: By coordinates or GeoJSON geometries
• Multi-temporal: Download imagery across date ranges
• Cloud filtering: Configurable cloud coverage thresholds
• Storage options: Local disk or AWS S3
• Type-safe API: Pydantic validation for all parameters

Installation

pip install sat-data-acquisition

Quick Start

from sat_data_acquisition import SatDataClient, ProcessingParams

# Define area (Copenhagen example)
geometry = {
    "type": "Polygon",
    "coordinates": [[
        [12.5464, 55.6761], [12.5864, 55.6761],
        [12.5864, 55.7061], [12.5464, 55.7061],
        [12.5464, 55.6761]
    ]]
}

# Configure search
processing_params = ProcessingParams(
    satellite='S2MPC',
    bands=['red', 'green', 'blue', 'nir'],
    start_date='2024-06-01',
    end_date='2024-08-31',
    cloud_coverage=20,
)

# Download imagery
client = SatDataClient()
dataset = client.search_and_create_image(
    geometry=geometry,
    processing_params=processing_params,
)

print(f"Downloaded {len(dataset.time)} images")
# Returns xarray.Dataset with shape: {'time': 12, 'y': 344, 'x': 291}

Visualization

import matplotlib.pyplot as plt
import numpy as np

# Get first image
image = dataset.isel(time=0)
rgb = np.dstack([image['red'], image['green'], image['blue']])

# Display
plt.imshow(rgb / 3000)
plt.title(f"Sentinel-2 - {str(image.time.values)[:10]}")
plt.show()

Save to Disk

from sat_data_acquisition import SaveParams
from sat_data_acquisition.processing import save_data

save_params = SaveParams(
    output_path='./data/images',
    save_to_local=True,
    save_as_geotiff=True,
    merge_bands=True,  # Single multi-band file vs separate files per band
)

for time_val in dataset.time.values:
    save_data(
        image=dataset.sel(time=time_val),
        identifier='copenhagen',
        datetime=str(time_val),
        satellite='S2MPC',
        provider='MPC',
        save_params=save_params,
    )

Available Satellites

Common Use Cases

Temporal Analysis

# Download monthly imagery for growing season
processing_params = ProcessingParams(
    satellite='S2MPC',
    bands=['red', 'green', 'blue', 'nir'],
    start_date='2024-04-01',
    end_date='2024-10-01',
    cloud_coverage=15,
)

dataset = client.search_and_create_image(geometry, processing_params)
# Analyze vegetation changes over time with xarray

SAR for All-Weather Monitoring

# Sentinel-1 works through clouds
processing_params = ProcessingParams(
    satellite='S1MPC',
    bands=['vv', 'vh'],  # Radar polarizations
    start_date='2024-01-01',
    end_date='2024-12-31',
)

S3 Storage

save_params = SaveParams(
    output_path='./temp',
    save_to_local=False,
    save_to_s3=True,
    s3_bucket='my-satellite-data',
    s3_path='projects/monitoring',
)

Documentation

Visit the GitHub repository for:

• Comprehensive guides: Detailed satellite specifications, processing parameters, and save options
• Working examples: Jupyter notebooks for each satellite with visualization
• Architecture docs: System design and data flow diagrams

Requirements

• Python 3.12+
• Core dependencies: xarray, rasterio, geopandas, odc-stac, pystac-client

STAC-Native Philosophy

Unlike legacy systems that download entire scenes, this package is STAC-native. It queries metadata first and leverages lazy-loading to stream only the specific pixels required for your geometry, significantly reducing bandwidth and storage requirements.

License

MIT License - see LICENSE

Acknowledgments

• Element84 for Earth Search STAC API
• Microsoft Planetary Computer for open data access
• ESA for Sentinel satellite programs
• NASA/USGS for Landsat data

Support

• Issues: GitHub Issues
• Documentation: GitHub Repository