From data to information

An introduction to using remote sensing data and open-source tools for analysis of foreshore characteristics

Edward P. Morris (UCA-CEIMAR) and Bas Oteman (NIOZ)

The capacity of light to carry and convey information is perhaps its most important, and remarkable, characteristic.

Ian Walmsley in 'Light : A very short introduction', Oxford University Press.

https://youtu.be/gF9UlTYR5d0

https://youtu.be/28hImDyJXu0?list=PLbyvawxScNbs9ET50PFwZTworudkNsnH_

Remote sensing 101

Platform, sensor and resolution

Satellite

platform-sensor

Platform

Landsat 8, Sentinel 1A, Sentinel 2A

Sensor or Instrument

OLI + TIRS, C-SAR, MSI

Resolution

Spatial: Pixel size, extent
Temporal: Revisit time
Spectral: Band width, Number of bands, Sensitivity

Electromagnetic radiation

Has both wave and particle properties at the same time (wave-particle duality).

Waves can be described by their amplitude and wavelength
Polarisation: waves oscillate in more than one orientation
Photons are the elementary particles of eletromagnetic radiation
Photons can be interact with matter by scattering and absorption
Reflectance is scattering in a defined direction

"Electromagneticwave3D" by Lookang many thanks to Fu-Kwun Hwang and author of Easy Java Simulation = Francisco Esquembre - Own work. Licensed under CC BY-SA 3.0 via Commons.

Electromagnetic wave spectrum

"EM spectrum". Licensed under CC BY-SA 3.0 via Commons.

Atmospheric electromagnetic opacity

"Atmospheric electromagnetic opacity" by NASA (original); SVG by Mysid. - Vectorized by User:Mysid in Inkscape, original NASA image from File:Atmospheric electromagnetic transmittance or opacity.jpg.. Licensed under Public Domain via Commons.

Passive visible and infrared

platform-sensor Ocean Optics Web Book • All contents 2015 Creative Commons Attribution license.

Measures (sun) light

day-time only*
radiance (light) passes through atmosphere
- affected by scattering and absorption
- affected by clouds
- requires atmospheric correction procedures

Passive visible and infrared

The surfaces of substances have specific scattering and absorption properties

allows identification of surfaces (i.e, the blue surface)
allows quantification of substances (i.e., the number of blue 'rocks')

Red and green light is absorbed, blue reflected.

"Simple reflectance" by User:Phidauex - Own work. Licensed under Public Domain via Commons.

Passive visible and infrared

Water strongly absorbs light giving it a distinctive reflectance spectrum that can be summarised by spectral indices (ratios), e.g., the 'Normalised Difference Water Index' (NDWI, McFeeters et al. 1996).

"Absorption spectrum of liquid water" by Kebes at English Wikipedia. Licensed under CC BY-SA 3.0 via Commons.

NDWI = (NIR - Green)/(NIR + Green)

Passive visible and infrared

All photosynthetic organisms contain chlorophyll giving them a distinctive reflectance spectrum that can be summarised by spectral indices (ratios), e.g., the 'Normalised Difference Vegetation Index' (NDVI).

"Vegetation spectral response" by wiki.landscapetoolbox.org. Licensed under Public Domain via Commons.

NDVI = (NIR - Red)/(NIR + Red)

Passive visible and infrared

The different pigments of photosynthetic organisms result in different 'colours', potentially allowing identification of foreshore vegetation types.

wikipedia.org/wiki/Chloroplast.

Passive visible and infrared

Colorvision Cadiz Bay

Passive visible and infrared

Spectral ratios, such as NDWI and NDVI can be used to quantify properties of foreshores.

"Fig. 6 Relationship between NDVI and chlorophyll..." by Kromkamp et al. 2006.

"Time-ensemble (2000-2015) mean modified normalised difference water index (MNDWI) of Cadiz Bay derived from the Landsat (USGS) archive" by FAST project.

Active radio

Synthetic Aperture Radar (SAR)

active radio pulses sent and collected
effective antena size much larger using movement of satellite and clever processing (Synthetic Aperture).
measures polarised radio waves e.g., horizontal and vertical.
allows increased spatial resolution.

"A geometric model of a SAR system" by Zhou et al. 2009.

Active Radio

Synthetic Aperture Radar (SAR)

most common bands are L, C and X
different band frequencies are suited to different applications
- penetrating canopies: L and P, very high resolution: X, multi-purpose: C

"Commonly used frequency bands..." by Moreira et al. 2013.

Active Radio

"Surface and volume scattering of a SAR beam for trees " by (Fernandez-Ordonez et al. 2009.).

Surface interactions

still water tends to be a specular reflector of radio waves i.e, usually low backscatter.
complex interactions with vegetation result in different return signals.
buildings often have strong backscatter signal.
interactions vary with beam polarization i.e., different polarizations provide more information.

Active Radio

"Speckle occurs in SAR images due to the coherent sum of many elemental scatterers within a resolution cell..." by Moreira et al. 2013.

Speckle

caused by the many scatterers in a 'pixel'.
coherence of the scattered signal results in strong fluctuations across the image.
'multi-looking' used to reduce speckle at cost of spatial resolution.
- non-coherent averaging of the intensity image.
adaptive filtering can also be used to reduce speckle.

Active Radio

"Interferometric Wide Swath Dual-Polarisation Sentinel-1 acquired on 2nd December 2014, band combination: VH-VV-VH (RGB) over the Sundarbans Mangrove, India and Bangladesh." © Copernicus (2014 / 2015) / ESA Image © Satellite Application Catapult Ltd, 2014.

Further resources

General:

Passive visible and infrared:

Ocean optics web book

Active radar (SAR):

Moreira, A., Prats-Iraola, P., Younis, M., Krieger, G., Hajnsek, I., & Papathanassiou, K. P. (2013). A tutorial on synthetic aperture radar. Geoscience and Remote Sensing Magazine, IEEE, 1(1), 6-43.
SAREDU; full open-access course on SAR by DLR

How to access remote sensing data

Data availability

Some earth observation satellites with open access data, based on wiki.landscapetoolbox.org. *C=Color, E=Elevation, H=Hyperspectral, M=Multispectral, P=Panchromatic, R=Syntheric Aperature Radar
Name	Abbreviation	Resolution (m)	Availability	Return Interval	Type*	Platform
Landsat Thematic Mapper 5	L5-TM	30,60	1984 to 2013	16 days	M	Satellite
Landsat Enhanced Thematic Mapper 7	L7-ETM	15,30,60	1995 to 2003	16 days	P,M	Satellite
Moderate-resolution Imaging Spectroradiometer	MODIS	250,500,1000	2000 to present	Daily	M	Satellite
Medium Resolution Imaging Spectrometer	ENVISAT-MERIS	300	2002 to present	3 days	M	Satellite
Landsat Data Continuity Mission (Landsat 8)	L8-OLI	15,30,100	2013 to present	16 days	P,M	Satellite
Copernicus Sentinel-1A	S-1A	4,10,25,40	2015 to present	12 (6) days	R	Satellite
Copernicus Sentinel-2A	S-2A	10,20,60	2015 to present	12 (6) days	M	Satellite

Data type and format

Landsat Missions Timeline

http://landsat.usgs.gov//about_mission_history.php.

Data type and format

Landsat product types

Generally surface reflectance products are prefered, however see L8-SR user notes and Landsat Higher Level Science Data Products:

not completely validated
not all scenes can be processed: Landsat 8 Pre-WRS-2 scenes (before April 11, 2013), scenes with a solar zenith angle greater than 76°.
Efficancy of L8SR correction will be likely reduced in: Hyper arid or snow covered regions, Low sun angle conditions, Coastal regions where land area is small relative to adjacent water, Areas with extensive cloud contamination.

High latitudes (> 65º) may not be valid.

http://landsat.usgs.gov//landsat_level_1_standard_data_products.php.

Data format and type

Landsat Standard L1 products

Standard Terrain Correction (Level 1T -precision and terrain correction) if possible.
Cubic Convolution (CC) resampling method.
30-meter (TM, ETM+, OLI) and 60-meter (MSS) pixel size (reflective bands).
Universal Transverse Mercator (UTM) map projection (Polar Stereographic projection for scenes with a center latitude greater than or equal to -63.0 degrees).
World Geodetic System (WGS) 84 datum.
MAP (North-up) image orientation.
GeoTIFF per band, quality assement + metadata (.MTL)

Data format and type

Landsat Higher Level Science Data products

surface reflectance
- includes basic classification (water, cloud) and per pixel quality mask
- metadata is .xml (schema: http://espa.cr.usgs.gov/v1.2)
surface reflectance spectral indices
transport format as standard GeoTIFF or ENVI binary or HDF-EOS2, see guide

Data type and format

Sentinels Timeline

ESA, http://www.unoosa.org/pdf/pres/stsc2015/tech-53E.pdf.

Data type and format

Sentinel 1 product types

https://sentinel.esa.int/web/sentinel/missions/sentinel-1/data-products.

Data type and format

Sentinel 1 product format

https://earth.esa.int/web/sentinel/user-guides/sentinel-1-sar/data-formats/sar-formats.

The SENTINEL-SAFE format wraps a folder containing image data in a binary data format and product metadata in XML.

Data type and format

Sentinel 2 product types

https://sentinel.esa.int/web/sentinel/missions/sentinel-2/data-products.

Data type and format

Sentinel 2 product format

https://sentinel.esa.int/web/sentinel/user-guides/sentinel-2-msi/data-formats.

The SENTINEL-SAFE format wraps a folder containing image data in a binary data format and product metadata in XML.

Earth Explorer

official bulk dissemination system for the Landsat (and other) products.
requires registration (quick and easy); email address used to send download links!
no restrictions on data use but requested to acknowledge source:
- USGS Products: 'Data available from the U.S. Geological Survey.'
- NASA LP DAAC Products: 'These data are distributed by the Land Processes Distributed Active Archive Center (LP DAAC), located at USGS/EROS, Sioux Falls, SD. http://lpdaac.usgs.gov'

surface reflectance products generated 'on demand'.

http://earthexplorer.usgs.gov/.

Sentinel Scientific Data Hub

official bulk dissemination system for the Sentinels products
requires registration (quick and easy)
use of data implies acceptance of Terms and Conditions of the use and distribution of Sentinel data
no limitations but must attribute 'Copernicus Sentinel data [Year]' or 'Contains modified Copernicus Sentinel data [Year]'

https://scihub.esa.int.

Other examples of how to get the data

General:

Google Earth Engine (GEE); Landsat archive, Sentinel-1 (available late 2015?), MODIS, MERRIS, ect. Sentinel-2 (planned).
- Google Storage also mirrors much of these archives. You can install their “gsutil” (free) and then list this directory, which is indexed by sensor (for example, L8 is Landsat 8) and path/row: gsutil ls gs://earthengine-public/landsat/
- Note Sentinel-1 has undergone further pre-processing in GEE

Sentinels:

Sentinel Data Access Service (SEDAS); useful portal by UK Space Agency

Other examples of how to get the data

Landsat:

USGS Earth Resources Observation and Science (EROS) Center Science Processing Architecture (ESPA) On Demand Interface for higher level products and processing options
Landsat Look for quick looks and standard data
GloVis an alternative to EarthExplorer
Python Landsat-Download
Python Landsat-util

Via commandline using curl:

# download a specific sentinel product
S:\>curl -u User:Password -JO "https://scihub.esa.int/dhus/odata/v1/\
Products('b94d872a-ef4f-4117-8a6b-1a90466ae785')/$value"

How to extract information

Basic concepts

Georectification: Mapping the data to real world coordinates.

Calibration: Converting the data to real world units

at-instrument; instrument-specific digital numbers to units , i.e., radiance
top-of-atmosphere; accounting for platform to target geometry.
surface; removing atmospheric affects and corrections for topography

Morphological operations: Filtering, changing resolution, ect.

Course classification: Assigning different classes (e.g., clouds, shadows, water) using segmentation or clustering.

Open source image processing (some options)

with GUI:

Sentinel Application Platform (SNAP); official software for processing sentinel data, integrates NEST and BEAM.
Quantum GIS (QGIS); extensible GIS system with many plugins linking to image processing packages (BEAM plugin).
BEAM
SeaDas
Monteverdi2
GRASS

Packages and frameworks:

Google Earth Engine; JavaScript and Python
Orpheo Toolbox (OTB); C
RSGISLib; Python
OpenCV; C, Python
raster; R

Sentinel Application Platform (SNAP)

A common architecture for all Sentinel Toolboxes

open-source (Java) available at GitHub or precompiled at STEP.
successor of the mature BEAM + NEST software.
efficient parralization of calculations resulting in good performance on modern PCs.
Python bindings available soon?
Sensor specific toolboxes (Sentinels, but also many other sensors).
Developed by Array Systems and DLR (S-1), C-S (S-2) and Brockmann Consult (S-3) for ESA.

Sentinel Application Platform (SNAP)

http://step.esa.int/main/toolboxes/snap/.

Getting help

Detailed in built help
Toolbox tutorials, also consider NEST and BEAM tutorials.
SNAP forum; a shared user experience
SNAP development; how to contribute

The power of the Graph Builder and Graph Processing Tool (GPT)

graphical building of workflows [graphs]
save workflows in XML and share
apply to single image or batch process many images
run from within the GUI or the command line (GPT)

ESA Research and Service Support (RSS)

'RSS has the mission to support the Earth Observation (EO) comunity in exploiting EO data'

supports institutions, scientists and developers.
supports data provisioning, access and processing.
help to scale up workflows
computation on demand and cloud computing service free for 'non-profit' organisations.
- contact rss_team@esa.int to check conditions and egibility.

Acknowledgents

Flag_of_Europe.svg" The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 607131. All views presented are those of the author’s, the European Union is not liable for any use that may be made of the information contained therein.

The FAST project team for support and contributions.
Landsat data courtesy of USGS.
This presentation includes Copernicus Sentinel data 2015.
Videos from NASA and ESA YouTube channels.
Comic design and production; The Pilot Dog
Slide deck made using slidify

Find out more about FAST

Get this talk:

slides: http://edwardpmorris.github.io/dsd_fast_teaching
source: https://github.com/edwardpmorris/dsd_fast_teaching

Foreshore Assessment using Space Technology project comic strip (EN)

Now your turn...

Using Sentinel 1 toolbox (S1TBX) to get information about coastal systems

Introduction to the Sentinel 1 toolbox (S1TBX)

This workshop will cover the basics of working with S1TBX to process Sentinel-1 data:

open S1A-IW-GRDH product
basic pre-processing
export data

Using the Graph Builder in S1TBX

use the graph builder to save your workflow.
take a look at the graph XML format.
edit the XML so as apply different types of speckle filter and multilooking, compare results.
set up thresholding of the image using band-math and extract 'water'.
run a basic classification and compare to the threshold result.