vincentsarago/landsat8_in_sentinel2

Find Landsat-8 in Sentinel-2 - Work In Progress

This repo is a response to this (:point_down:) blog post to see if it was possible to find a Sentinel-2 scene where we can spot Landsat 8 satellite.

I am prepared to be wrong on this new blogpost "When #Landsat8 (almost) meets #Sentinel2". https://t.co/5RXzAerKCu @USGSLandsat @CopernicusData #earthengine #EarthObservation #remotesensing #freedata pic.twitter.com/VjUCVCH92e

— Philipp Gärtner (@gartn001) July 6, 2018

Prerequisites

1. Install Python modules

pip install pandas geopandas shapely

Note: if ☝️ fails for pyproj, try: pip install git+https://github.com/jswhit/pyproj.git#egg=pyproj

2. Download Landsat and Sentinel indexes

   • Landsat: wget http://storage.googleapis.com/gcp-public-data-landsat/index.csv.gz && gunzip index.csv.gz > landsat.csv
   • Sentinel: wget http://storage.googleapis.com/gcp-public-data-sentinel-2/index.csv.gz && gunzip index.csv.gz > sentinel.csv

$ cat landsat.csv | wc -l
13 749 024

$ cat sentinel.csv | wc -l
6 741 652

Run

1. Find overlapping Sentinel-2 and Landsat-8 scenes

python find_landsat_in_sentinel.py > l8_overlap_s2.json

Come back in ~5 hours

The script should find ~5800 Sentinel-2 scenes which have at least one overlapping Landsat-8 scene acquired within 5 seconds.

Each output should be like:

{
  "type": "Feature",
  "properties": {
	// Sentinel-2 scene ID
    "S2_PRODUCT_ID": "S2A_MSIL1C_20170915T112701_N0205_R080_T35XNJ_20170915T112704",
	// Sentinel-2 acquisition date
    "S2_SENSING_TIME": "2017-09-15T11:27:04.460000",
	// List of overlapping Landsat-8 ID
    "L8_SCENE_ID": [ "LC82080022017258LGN00" ],
    "L8_PRODUCT_ID": [ "LC08_L1GT_208002_20170915_20170928_01_T2" ],
	// List of overlapping Landsat-8 acquisition date
    "L8_SENSING_TIME": [ "2017-09-15T11:27:01.014935" ]
  },
  "geometry": {
    "type": "Polygon",
    "coordinates": [
      [
        [ 32.737655971799995, 79.1376077611 ],
        [ 32.737655971799995, 80.132118355 ],
        [ 26.9985522679, 80.132118355 ],
        [ 26.9985522679, 79.1376077611 ],
        [ 32.737655971799995, 79.1376077611 ]
      ]
    ]
  }
}

[TODO]

2. Validate Scenes (S2/L8) geometries by reading scenes metadata

3. Check Landsat orbit parameters to confirm Sentinel-2 could see it