Tutorial — Acquiring Free Satellite Imagery

Goal: After this tutorial, you can find and download satellite imagery for any location on Earth for free, both manually and programmatically.

All intelligence analysis starts with data. The good news: terabytes of satellite imagery are freely available. Sentinel-2 optical, Sentinel-1 SAR, Landsat (50+ year archive), SRTM elevation — all free, all global coverage. You will never need to pay for data to learn GEOINT.

Step 1: Copernicus Data Space (Browser)

URL: https://dataspace.copernicus.eu

This is your primary source. ESA’s Copernicus program provides Sentinel-1 (SAR), Sentinel-2 (optical), Sentinel-3 (ocean/land), and Sentinel-5P (atmosphere) — all free.

API Update 2024

The Copernicus Data Space Ecosystem (CDSE) replaced the legacy Copernicus Open Access Hub (SciHub) in 2023. The new API uses OAuth2 authentication and OData queries. If you encounter old SciHub code examples, they will not work with the new system.

Create an Account

Go to https://dataspace.copernicus.eu
Click “Register” — free, requires email verification
This account works for both the browser interface and the API
Important: Save your credentials securely — you’ll need them for API access

API Authentication (New System)

The CDSE uses OAuth2 for authentication. You’ll need to:

Obtain an access token using your username/password
Include the token in API requests
Tokens expire and must be refreshed

# CDSE OAuth2 authentication example
import requests
 
def get_access_token(username, password):
    """Obtain OAuth2 access token for CDSE API."""
    token_url = "https://identity.dataspace.copernicus.eu/auth/realms/CDSE/protocol/openid-connect/token"
    
    data = {
        'grant_type': 'password',
        'username': username,
        'password': password,
        'client_id': 'cdse-public',
    }
    
    response = requests.post(token_url, data=data)
    response.raise_for_status()
    
    return response.json()['access_token']
 
# Use token in subsequent requests
headers = {'Authorization': f'Bearer {access_token}'}

Browser Interface

Open the Copernicus Browser (browser.dataspace.copernicus.eu)
Navigate to your area of interest (zoom/pan or search by place name)
Left panel: select data source
- Sentinel-2 L2A — atmospherically corrected surface reflectance (use this)
- Sentinel-2 L1C — top-of-atmosphere (raw, needs atmospheric correction)
- Sentinel-1 GRD — SAR amplitude (easier to use)
- Sentinel-1 SLC — SAR complex (for interferometry)
Set date range and maximum cloud cover (e.g., <20%)
Click “Search” — results appear as footprints on the map
Click a result to preview → “Download” to get the full product (ZIP file)

Product naming convention:

S2A_MSIL2A_20240315T100021_N0510_R122_T35VLF_20240315T134020.zip
│   │      │                │     │    │       │
│   │      │                │     │    │       └─ Processing timestamp
│   │      │                │     │    └─ Tile ID (MGRS grid)
│   │      │                │     └─ Relative orbit
│   │      │                └─ Processing baseline
│   │      └─ Acquisition date/time
│   └─ Product level (L2A = surface reflectance)
└─ Satellite (A or B)

Tile structure (Sentinel-2): Each product covers a 100x100 km tile in the MGRS grid. Estonia is covered primarily by tiles T35VLF (Tallinn area), T35VLE, T35VMF. A full tile at all bands is ~800MB compressed.

What’s Inside a Sentinel-2 Product

After unzipping, the structure is:

S2A_MSIL2A_.../
  GRANULE/
    L2A_T35VLF.../
      IMG_DATA/
        R10m/           # 10m resolution bands
          T35VLF_B02_10m.jp2   # Blue
          T35VLF_B03_10m.jp2   # Green
          T35VLF_B04_10m.jp2   # Red
          T35VLF_B08_10m.jp2   # NIR
          T35VLF_TCI_10m.jp2   # True Color Image (RGB composite)
        R20m/           # 20m resolution bands
          T35VLF_B05_20m.jp2   # Red Edge 1
          T35VLF_B06_20m.jp2   # ...
          T35VLF_B11_20m.jp2   # SWIR 1
          T35VLF_B12_20m.jp2   # SWIR 2
          T35VLF_SCL_20m.jp2   # Scene Classification (cloud mask)
        R60m/           # 60m resolution bands
      QI_DATA/          # Quality indicators

The SCL (Scene Classification Layer) is critical for analysis — it classifies each pixel as cloud, cloud shadow, water, vegetation, bare soil, etc. Use it to mask clouds before any analysis.

Step 2: USGS EarthExplorer

URL: https://earthexplorer.usgs.gov

Primary source for Landsat data (since 1972!) and other US datasets.

What’s Available

Landsat-9 (2021-present): 30m multispectral, 15m pan, 100m thermal
Landsat-8 (2013-present): same as above
Landsat-7 (1999-present): has scan-line corrector failure since 2003 (striped data)
Landsat-5 (1984-2013): Thematic Mapper, 30m
Landsat 1-4 (1972-1993): MSS, 60-80m — historical archive
ASTER: 15m visible, 30m SWIR, 90m thermal
SRTM DEM: 30m global elevation
Declassified satellite imagery: CORONA, GAMBIT, HEXAGON — cold war spy satellites, now unclassified

Download Process

Create free USGS account at https://ers.cr.usgs.gov/register
Define area of interest (map, coordinates, or place name)
Select dataset (e.g., Landsat 8-9 OLI/TIRS C2 L2)
Set date range, cloud cover
Results → click to preview → “Add to Bulk Download” or direct download
Use the Bulk Download Application for large orders

Declassified spy satellite imagery is a unique resource. CORONA (1960-1972) provides ~2m resolution imagery decades before commercial satellites existed. Invaluable for historical analysis of military installations, urban growth, environmental change.

Step 3: Copernicus Data Space API (Programmatic)

For automated workflows, search and download via API. The OData API provides full access.

Search for Sentinel-2 Products

import requests
from datetime import datetime, timedelta
 
def search_sentinel2(lat, lon, start_date, end_date, max_cloud=20, max_results=10):
    """
    Search for Sentinel-2 L2A products covering a location.
    Uses Copernicus Data Space OData API.
    """
    # OData API endpoint
    base_url = "https://catalogue.dataspace.copernicus.eu/odata/v1/Products"
 
    # Build filter
    # Geographic filter: intersects with point
    geo_filter = (
        f"OData.CSC.Intersects(area=geography'SRID=4326;POINT({lon} {lat})')"
    )
 
    # Collection and processing level
    collection_filter = (
        "Collection/Name eq 'SENTINEL-2' and "
        "Attributes/OData.CSC.StringAttribute/any(att:att/Name eq "
        "'productType' and att/OData.CSC.StringAttribute/Value eq 'S2MSI2A')"
    )
 
    # Cloud cover filter
    cloud_filter = (
        f"Attributes/OData.CSC.DoubleAttribute/any(att:att/Name eq "
        f"'cloudCover' and att/OData.CSC.DoubleAttribute/Value lt {max_cloud})"
    )
 
    # Date filter
    date_filter = (
        f"ContentDate/Start gt {start_date}T00:00:00.000Z and "
        f"ContentDate/Start lt {end_date}T23:59:59.000Z"
    )
 
    filter_str = f"{geo_filter} and {collection_filter} and {cloud_filter} and {date_filter}"
 
    params = {
        "$filter": filter_str,
        "$orderby": "ContentDate/Start desc",
        "$top": max_results,
    }
 
    response = requests.get(base_url, params=params)
    response.raise_for_status()
    data = response.json()
 
    results = []
    for product in data.get("value", []):
        results.append({
            "id": product["Id"],
            "name": product["Name"],
            "date": product["ContentDate"]["Start"],
            "size_mb": product.get("ContentLength", 0) / 1e6,
        })
    return results
 
 
# Search for Sentinel-2 over Tallinn in the last 30 days
results = search_sentinel2(
    lat=59.44, lon=24.75,
    start_date="2024-03-01",
    end_date="2024-03-31",
    max_cloud=30,
)
 
for r in results:
    print(f"{r['date'][:10]}  {r['name'][:60]}...  {r['size_mb']:.0f} MB")

Download a Product

import requests
import os
 
def get_access_token(username, password):
    """Get access token from Copernicus Data Space."""
    token_url = "https://identity.dataspace.copernicus.eu/auth/realms/CDSE/protocol/openid-connect/token"
    data = {
        "client_id": "cdse-public",
        "username": username,
        "password": password,
        "grant_type": "password",
    }
    response = requests.post(token_url, data=data)
    response.raise_for_status()
    return response.json()["access_token"]
 
 
def download_product(product_id, token, output_dir="."):
    """Download a product by its ID."""
    url = (
        f"https://zipper.dataspace.copernicus.eu/odata/v1/"
        f"Products({product_id})/$value"
    )
    headers = {"Authorization": f"Bearer {token}"}
 
    response = requests.get(url, headers=headers, stream=True)
    response.raise_for_status()
 
    # Get filename from content-disposition header
    cd = response.headers.get("content-disposition", "")
    filename = cd.split("filename=")[-1].strip('"') if "filename=" in cd else f"{product_id}.zip"
    filepath = os.path.join(output_dir, filename)
 
    total = int(response.headers.get("content-length", 0))
    downloaded = 0
 
    with open(filepath, "wb") as f:
        for chunk in response.iter_content(chunk_size=8192):
            f.write(chunk)
            downloaded += len(chunk)
            if total > 0:
                pct = downloaded / total * 100
                print(f"\r  Downloading: {pct:.1f}%  ({downloaded/1e6:.0f}/{total/1e6:.0f} MB)",
                      end="")
    print(f"\n  Saved: {filepath}")
    return filepath
 
 
# Usage:
# token = get_access_token("your_email", "your_password")
# download_product(results[0]["id"], token, output_dir="./data")

Alternative: sentinelhub Library

The sentinelhub Python library provides a higher-level API for Copernicus Data Space:

# pip install sentinelhub
from sentinelhub import (
    SHConfig, BBox, CRS, DataCollection,
    SentinelHubRequest, MimeType, bbox_to_dimensions,
)
import numpy as np
 
# Configure credentials (Copernicus Data Space)
config = SHConfig()
config.sh_client_id = "your_client_id"       # from dataspace.copernicus.eu
config.sh_client_secret = "your_client_secret"
config.sh_base_url = "https://sh.dataspace.copernicus.eu"
config.sh_token_url = (
    "https://identity.dataspace.copernicus.eu/auth/realms/"
    "CDSE/protocol/openid-connect/token"
)
 
# Define area: Tallinn harbor area
tallinn_bbox = BBox([24.70, 59.42, 24.82, 59.47], crs=CRS.WGS84)
resolution = 10  # meters
size = bbox_to_dimensions(tallinn_bbox, resolution=resolution)
 
# Request true color image
evalscript = """
//VERSION=3
function setup() {
    return {
        input: ["B04", "B03", "B02"],
        output: { bands: 3 }
    };
}
function evaluatePixel(sample) {
    return [3.5 * sample.B04, 3.5 * sample.B03, 3.5 * sample.B02];
}
"""
 
request = SentinelHubRequest(
    evalscript=evalscript,
    input_data=[
        SentinelHubRequest.input_data(
            data_collection=DataCollection.SENTINEL2_L2A,
            time_interval=("2024-06-01", "2024-06-30"),
            maxcc=0.2,
        )
    ],
    responses=[SentinelHubRequest.output_response("default", MimeType.TIFF)],
    bbox=tallinn_bbox,
    size=size,
    config=config,
)
 
images = request.get_data()
image = images[0]  # numpy array, shape (H, W, 3)
print(f"Image shape: {image.shape}")

Step 4: Google Earth Engine

URL: https://code.earthengine.google.com (JavaScript) or Python API

Google Earth Engine (GEE) is a cloud platform for processing planetary-scale geospatial data. Instead of downloading terabytes, you write analysis scripts that run on Google’s servers.

What’s Available (Free for Research/Education)

Sentinel-2 (2015-present)
Sentinel-1 (2014-present)
Landsat entire archive (1972-present)
MODIS, VIIRS
SRTM, Copernicus DEM
Night lights (DMSP, VIIRS)
Many more: 900+ public datasets

Python API Setup

# pip install earthengine-api
import ee
 
# Authenticate (first time: opens browser for Google account)
ee.Authenticate()
ee.Initialize(project="your-project-id")  # from console.cloud.google.com
 
# Get median Sentinel-2 image for Tallinn, summer 2024
tallinn = ee.Geometry.Point(24.75, 59.44)
s2 = (
    ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")
    .filterBounds(tallinn)
    .filterDate("2024-06-01", "2024-08-31")
    .filter(ee.Filter.lt("CLOUDY_PIXEL_PERCENTAGE", 20))
    .median()  # pixel-wise median reduces clouds
)
 
# Compute NDVI
ndvi = s2.normalizedDifference(["B8", "B4"]).rename("NDVI")
 
# Export to Drive (downloads to your Google Drive)
task = ee.batch.Export.image.toDrive(
    image=ndvi,
    description="tallinn_ndvi_summer2024",
    scale=10,
    region=tallinn.buffer(10000).bounds(),  # 10km radius
    maxPixels=1e9,
)
task.start()
print(f"Export started. Check status at: https://code.earthengine.google.com/tasks")

GEE JavaScript (Quick Analysis in Browser)

// Paste into code.earthengine.google.com
var tallinn = ee.Geometry.Point(24.75, 59.44);
 
var s2 = ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')
    .filterBounds(tallinn)
    .filterDate('2024-06-01', '2024-08-31')
    .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 20))
    .median();
 
// True color
Map.centerObject(tallinn, 12);
Map.addLayer(s2, {bands: ['B4', 'B3', 'B2'], min: 0, max: 3000}, 'True Color');
 
// NDVI
var ndvi = s2.normalizedDifference(['B8', 'B4']);
Map.addLayer(ndvi, {min: 0, max: 0.8, palette: ['red', 'yellow', 'green']}, 'NDVI');

Step 5: Other Free Sources

NASA Worldview

URL: https://worldview.earthdata.nasa.gov

Quick browse of global imagery (MODIS, VIIRS, Landsat, Sentinel)
Good for fire detection, smoke plumes, large-scale events
Animation feature: watch daily changes over time
Direct download of visible tiles

SRTM DEM (Elevation Data)

URL: https://dwtkns.com/srtm30m/ or via USGS EarthExplorer

30m resolution global elevation model (56S to 60N latitude)
Essential for terrain analysis, viewshed, slope
Alternative: Copernicus DEM (also 30m, better quality, global)

OpenAerialMap

URL: https://openaerialmap.org

Crowdsourced aerial and drone imagery
Varies in coverage and quality
Useful for high-res reference imagery in specific areas

Night Lights

VIIRS DNB (NASA): 750m, monthly composites
Applications: urbanization, power outage detection, military activity

Step 6: Data Formats

Format	Extension	Contains	Used By
GeoTIFF	.tif	Raster + georeference metadata + CRS	Everything — the standard
JPEG2000	.jp2	Compressed raster + georeference	Sentinel-2 raw products
NetCDF	.nc	Multi-dimensional arrays (time series, climate)	MODIS, climate data
Shapefile	.shp (+.dbf,.shx,.prj)	Vector geometry + attributes	Boundaries, AOIs
GeoJSON	.geojson	Vector (JSON format)	Web mapping, APIs
GeoPackage	.gpkg	Vector or raster (SQLite-based)	Modern replacement for shapefile
SAFE	.SAFE (directory)	ESA product structure	Sentinel-1/2 products

GeoTIFF is king. When in doubt, convert to GeoTIFF. All tools (rasterio, GDAL, QGIS) handle it natively. It stores georeference, CRS, and band data in a single file.

# Convert JP2 to GeoTIFF with GDAL
# gdal_translate input.jp2 output.tif -of GTiff -co COMPRESS=DEFLATE
 
# Or in Python with rasterio:
import rasterio
from rasterio.enums import Compression
 
with rasterio.open("sentinel2_B04.jp2") as src:
    profile = src.profile.copy()
    profile.update(driver="GTiff", compress="deflate")
 
    with rasterio.open("sentinel2_B04.tif", "w", **profile) as dst:
        dst.write(src.read())

Try This Next

Exercise 1: Download and View Your City

Go to Copernicus Browser, find a cloud-free Sentinel-2 L2A image of your city
Download the product, unzip it
Load B04 (Red), B03 (Green), B02 (Blue) with rasterio
Display as true color composite — can you identify major landmarks?
Compute NDVI using B08 and B04 — find the greenest park

Exercise 2: SAR of a Harbor

On Copernicus Browser, find a Sentinel-1 GRD-IW product covering a major harbor (Tallinn, Helsinki, or any port)
Download the VV polarization band
Load and display in dB scale: 10 * np.log10(amplitude)
Can you identify ships? What do they look like in SAR? (See SAR Fundamentals and Analysis)

Exercise 3: Historical Comparison

On USGS EarthExplorer, find Landsat-5 TM imagery of Tallinn from 1990
Find Landsat-9 of the same area from 2024
Compare: what changed in 34 years? New construction, urban expansion, coastline?
This is your first change detection exercise

Self-Test Questions

What is the difference between Sentinel-2 L1C and L2A products?
How do you filter Sentinel-2 results to get only cloud-free imagery?
What is the MGRS tile ID for Tallinn, Estonia?
Name three advantages of using Google Earth Engine over downloading data locally.
What data format should you use for most raster analysis workflows?

GEOINT Notes

Explorer

Tutorial - Acquiring Free Satellite Imagery

Tutorial — Acquiring Free Satellite Imagery

Step 1: Copernicus Data Space (Browser)

Create an Account

API Authentication (New System)

Browser Interface

What’s Inside a Sentinel-2 Product

Step 2: USGS EarthExplorer

What’s Available

Download Process

Step 3: Copernicus Data Space API (Programmatic)

Search for Sentinel-2 Products

Download a Product

Alternative: sentinelhub Library

Step 4: Google Earth Engine

What’s Available (Free for Research/Education)

Python API Setup

GEE JavaScript (Quick Analysis in Browser)

Step 5: Other Free Sources

NASA Worldview

SRTM DEM (Elevation Data)

OpenAerialMap

Night Lights

Step 6: Data Formats

Try This Next

Exercise 1: Download and View Your City

Exercise 2: SAR of a Harbor

Exercise 3: Historical Comparison

Self-Test Questions

Graph View

Table of Contents

Backlinks