GEO – SIGMA – Capacity - PowerPoint Presentation

Slide1

GEO – SIGMA – Capacity Building Tools for land cover validation and in-situ data collection

Steffen

Fritz

, Myroslava

Lesiv,Linda

See, Ian McCallum,

Juan-Carlos, Christoph

Perger, Dmitry

Schepaschenko

,

Anatoly

Shvidenko

,

Inian Moorthy, Carl

Salk, Martina

Duerauer

, Mathias

Karner

, Tobias Sturn, Christopher Dresel,

Dahlia

Domian, Antonia

Dunwoody, Olha Danylo, Juan-Carlos

Laso

Bayas

Earth Observations Group

Ecosystems Service and Management (ESM)

Slide2

Geo-Wiki Engagement Platform

Geo-Wiki is an open platform that provides citizens with the means to engage in environmental monitoring by providing feedback on existing spatial information overlaid on satellite imagery or by contributing entirely new data.

10,000+

Registered users

http://www.geo-wiki.org

/

Slide3

Land

Cover

Agriculture

Livestock

Cities

Risk

SIGMA

Biomass

Geo-Wiki Branches

Picture Pile

Cropland Capture

Data Collection Tools

Laco

-Wiki

Pixel Validator

Games for Visual Interpretation

Analytical Tools – Visual Interpretation

In-situ

Tools

Geo-Wiki

Pictures

FotoQuest

GeoODK

…

Slide4

LACO-WIKI - Land Cover Validation tool

Slide5

OverviewUploadSamplingValidationReporting

Slide6

Logging in / Languages

Slide7

Configurable Validation Sessions

Slide8

Step 1: Data Upload/Management

Slide9

Slide10

ETRS1989 LAEA Support

Slide11

Slide12

Step 2: Generate Samples

Slide13

Step 2: Generate Samples

Slide14

Stratified Vector and Raster Sampling

Slide15

Slide16

Geometry Validation & Comments

Slide17

Accuracy AssessmentRaw data / confusion matrixPossibility to weight confusion matrix by object areaOverall accuracy / errors of omission and commissionKappaAverage Mutual Information (AMI)

Quantitative and Allocation Disagreement (Pontius and

Millones

, 2011)We can add others as required by the community

Slide18

In-situ component – LACO-Wiki mobile Based on FotoQuest Austria

Slide19

Slide20

Steps to get to area estimationUse initial confusion matrix to estimate varianceDefine overall precision needed for the overall class area

Add a second sample (e.g. point sample) and depending on the variance add new sample points needed to reach precision

Legend

c

= Class of interest

h

= Stratum given class

A

= Areap

= Proportion/probability

n

= Sample

Area estimation when stratification occurs:

With Variance and Coefficient of Variation:

Slide21

Slide22

Slide23

Picture Pile

http://

geo-wiki.org/games/picturepile

Slide24

Picture Pile

Deforestation Pile

Pending:

Cropland

Human Impact

Oil palm identification

 New

scoring mechanism

More reference data based on

expert classifications

Slide25

Picture Pile

150+

Players

October, 2015

???, 2015

80,500+

Classifications

90%

< 5 seconds

11 000

Unique images

Deforestation Pile – 1 week

Time (s)

Classifications

Official

launch: October 28

Check it out!

Slide26

Slide27

FotoQuest GO

http://fotoquest.at

/

Slide28

Scaling to FotoQuest

Europe!

FotoQuest

Austria

July, 2015

September, 2015

12 weeks

200+

Players

2000+

Quests

12,000+

Photos

12%

Quests completed

Slide29

Geo-wiki pictures

Slide30

Managing Pictures in fotoquest – link to CIRAD observatory

Slide31

Fill out SurveysEdit Surveys

Send Survey to DB

View Survey on a Map

View OFFLINE Layers

GPS Location

Polygon Feature

GPS Trace

Offline- Maps-LayersLocal Storage

GeoODK

Collect:

Officially Released September 2014

Slide32

Mapping & Offline Layers

Open Street Map

MapBox

Tiles (.

mbtiles

)

Google KML

Local StorageOnline/Offline Device Location Functionality

Slide33

Offline Mapping of Data

Slide34

Thanks!fritz@iiasa.ac.at

Steffen

Fritz, Linda See, Ian McCallum, Christoph

Perger, Dmitry Schepaschenko, Myroslava Lesiv, Inian Moorthy,

Anatoly Shvidenko, Carl Salk, Martina Duerauer, Mathias

Karner, Tobias Sturn, Christopher Dresel, Dahlia Domian, Antonia Dunwoody, Olha

Danylo, Juan-Carlos LasoEarth Observations Group

Ecosystems Service and Management (ESM)