Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

Image: New York Times

Google Earth Engine


Noel Gorelick
Google Switzerland
University of Zurich
Presented by: Eric Bullock

Bill Silva, at USFS Geospatial Technology and


Applications Center (GTAC)

Presentation: http://bit.ly/33CsbKq
About me
● PhD in Geography from Boston
University
● Currently Post-Doctoral
Researcher
● Remote sensing of land cover and
land cover change
● Focus on monitoring disturbance
in tropics
● Not a programmer, and do not
represent Google
● Bogor -> Boston: 16000 km

Presentation: http://bit.ly/33CsbKq
Tentative Workshop Outline
Today
Google Earth Engine Overview
Data exploration on the Earth Engine
Basic land cover classification
No programming

Tomorrow
Outputting data
Change detection

Friday
Accuracy Assessment and area estimation
Degradation and deforestation monitoring
using CODED

Presentation: http://bit.ly/33CsbKq
Google Earth Engine
Data analysis and visualization platform
Designed for scientists, not software engineers
Goals
Make it easy
Scalable and shareable workflows
Enable non-traditional users
Focused on society's biggest challenges
Deforestation Disaster Climate Change
Drought Disease Sustainability

Presentation: http://bit.ly/33CsbKq
Why is a powerful tool
Public data catalog: vast amounts of available data (you
don’t need to store data)

Processing power (computation engine)

Interactive development platforms

Save and share work routines

Presentation: http://bit.ly/33CsbKq
25PB Data Catalog
Imagery
Landsat 1-8 Optical & Thermal, 30m
MODIS 250m Daily Global
Sentinel-1 10m SAR
Sentinel-2 12 bands, 20m
Geophysical
Digital Elevation
Land Cover
Surface Temperature
Weather Forecasts, Climate Models
+500 other datasets
Upload your own data
Load your own data
• Fusion Tables

Polygons in GEE come


in Fusion Tables.
These are in a
spreadsheet format
but can contain
geospatial information
that can be viewed in
GEE

Presentation: http://bit.ly/33CsbKq
Make your own data in the map
You can digitize vectors datasets (poligons) in the map window
These records can be converted to code

Presentation: http://bit.ly/33CsbKq
Why is a powerful tool
Public data catalog: vast amounts of available data (you
don’t need to store data)

Processing power (computation engine)

Interactive development platforms

Save and share work routines

Presentation: http://bit.ly/33CsbKq
Computation Engine

● Unprecedented speed: reduce


processing times by orders of magnitude
by using the distributed, cloud-based
computing power

● Ease of use and lower costs: online


platform with easy access to data,
scientific algorithms, computational
power.

Presentation: http://bit.ly/33CsbKq
Why is a powerful tool
Public data catalog: vast amounts of available data (you
don’t need to store data)

Processing power (computation engine)

Interactive development platforms

Save and share work routines

Presentation: http://bit.ly/33CsbKq
Two platforms

Graphical User Interface (Explorer)


•https://explorer.earthengine.google.com/#workspace
•User friendly way to begin exploring and analyzing data

Application Program Interface (Code Editor)


•https://code.earthengine.google.com/
•Powerful geospatial tool to create complex custom analysis
•Requires some programming knowledge
•Supports both JavaScript and Python
•Tool can be turned into Earth Engine Applications

Presentation: http://bit.ly/33CsbKq
Explorer Platform

A point and click platform:

•Audience: non-programmers
•Basic analysis functionality
•Per pixel math (~ESRI’s raster calculator),
•Neighborhood algorithms,
•Terrain algorithms, etc
•Save and share workspace
•Explore and export data

Presentation: http://bit.ly/33CsbKq
Code Editor
Applications
No account required

Cannot export or save


results

No programming
required
Why is a powerful tool
Public data catalog: vast amounts of available data (you don’t need to
store data)

Processing power (computation engine)

Interactive development platforms

Save and share work routines

Presentation: http://bit.ly/33CsbKq
Status

Free for academic and research use


More than 100,000 users in 205 countries
900+ composable "algorithms"
Large "freemium" quota
350 / 600 Earth Engine based papers in 2016 / 2017

Your IP is your IP
Customer owns all Intellectual Property Rights in
Customer Data, Customer Code, and Application(s).

Available for commercial users in 2019

Presentation: http://bit.ly/33CsbKq
Image Processing Tools

● Data exploration and visualization


● Image classification
● Change detection
● Segmentation
● Machine learning and “Deep learning”
● Vector operations
● Time series analysis (such as CODED)
● Accuracy and area estimation

Presentation: http://bit.ly/33CsbKq
Personal Motivation

● Previously: Self-developed
open source software
● Data download necessary
● Impossible to do large
area analysis without
high-performance
computing system

Presentation: http://bit.ly/33CsbKq
Personal Motivation

● Now: All analysis performed


on GEE
● No downloading required
● No system requirements
● Large user community
● Google engineers do most of
programming, not us
scientists
● Currently being used as
primary platform by UNFCCC
reporting countries
Presentation: http://bit.ly/33CsbKq
● CODED: Continuous
Degradation Detection
● Amazon: ~150,000
Landsat scenes
● ~75 years on one
high-performance
computer
● GEE: 5 days

Presentation: http://bit.ly/33CsbKq
Sign up for an account

https://signup.earthengine.google.com/

Thanks!
Eric Bullock
bullocke@bu.edu

Presentation: http://bit.ly/33CsbKq
Getting started with Earth Engine Apps
● Visualize Landsat composites

○ https://bullocke.users.earthengine.app/view/landsat-slider

● Classifying IPCC land cover classes

○ https://bullocke.users/earthengine.app/view/ipcc-classification

● Earth Engine Explorer (Account required)

○ https://explorer.earthengine.google.com/#workspace

Presentation: http://bit.ly/33CsbKq
Optional Extra Apps
● Animate Landsat data over time:

○ https://emaprlab.users.earthengine.app/view/lt-gee-time-series-animator
● Landcover mapping using REMAP1

○ https://remap-app.org
● Land cover change using LandTrendR2:

○ https://emaprlab.users.earthengine.app/view/lt-gee-change-mapper

● CODED results in Indonesia

○ https://bullocke.users.earthengine.app/view/coded-indonesia

1. Kennedy, R.E., Yang, Z., Gorelick, N., Braaten, J., Cavalcante, L., Cohen, W.B., Healey, S. (2018). Implementation of the LandTrendr Algorithm on
Google Earth Engine. Remote Sensing. 10, 691.
2. Murray, N. J., Keith, D. A., Simpson, D. , Wilshire, J. H. and Lucas, R. M. (2018), Remap: An online remote sensing application for land cover
classification and monitoring. Methods Ecol Evol. Accepted Author Manuscript. . doi:10.1111/2041-210X.13043

You might also like