Space Data Management System (SDMS) - ESA SEOMseom.esa.int/S2forScience2014/files/06_S2forScience-ForestryII... · Space Data Management System (SDMS) Acquire, ... Satellite Land

Space Data Management System (SDMS)

Acquire, Query, Process and Deliver Earth Observation Data and Forest Information

Products to Developing Countries Pilot & Project

I.Jonckheere, E. Lindquist & A. Pekkarinen FAO Forestry Department

Sentinel-2 for Science Symposium Frascati, IT, May 2014

National Forest Monitoring Systems for REDD+ FAO Approach

Satellite Land Monitoring System

National Forest Inventory

GHG Inventory

NATIONAL FOREST MONITORING SYSTEM

MRV MONITORING

Web Interface

Community Monitoring

Other Forest-Related Monitoring Systems

For monitoring and reporting on

REDD+ activities

For assessing national area change over

time (Activity Data)

For assessing national carbon stock changes

(Emission Factors)

Problem • Measurement and monitoring of land cover / use

at national scale in developing countries is required for REDD+

• Poor access to satellite data, limited expertise and computing power are bottlenecks

• Particularly challenging to turn data into useful information / country statistics

Problem – Current Situation • Data is free but still disaggregated and can be

non-intuitive to search, query and download • Software is free but version control is

challenging, dependencies are a pain, use can be complicated, disjointed

• Computers are cheaper but infrastructure required for high performance processing is challenging for developing nations

• Internet is everywhere but many ministry departments have challenges to access

Solution • Easy(er) access to space data; • Easy(er) access to processing tools; • More intuitive / streamlined / automatic (pre)

processing; • Reduce burden of computing infrastructure; • Reduce data dimensionality for download; • Cloud-based or desktop functionality; • National ‘ownership’ / ‘control’ of data and

products; • Fully open-source…

SDMS Objectives Improve data access and delivery of (pre)processed satellite data and forest information products to enable developing countries to build the autonomous capacity to monitor their forest-related REDD+ activities.

Data Data Data

Users

Information

SDMS Objectives

Project Focus and Outputs

• Monitoring using (pre)processed satellite data

• Capacity building • Delivery of (pre)processed satellite data

• Development of open-source software tools

• REDD+ compliant

• Managed server + secure storage

• Internet-based + desktop functionality

• Simple user interface

• Metadata query and data retrieval as needed from existing repositories (Landsat, MODIS, Sentinel, WELD, etc…)

SDMS structure

• Open-source image processing and analysis (OpenForis, R, GDAL, etc.)

• Automation of key processing steps

• Compliant with GFOI Methods and Guidance http://gfoi.org/methods-guidance-documentation

• Consistently updated

SDMS structure

Cloud computing structure

(SDMS)

FAO Forestry

Open source image processing Linux / Mac / Cygwin compatible Command line Pixel or object-based image analysis Updated regularly Wiki and example exercises http://km.fao.org/OFwiki/index.php/Main_Page

http://www.fao.org/forestry/fma/openforis/en/

SDMS processing

Advanced image interpretation, open source platform

FAO Forestry

OpenFORIS

Image Segmentation and Classification

Time 1 Time 2

Advanced image interpretation, open source platform

FAO Forestry

Multi-date Image Segmentation

OpenFORIS

Advanced image interpretation, open source platform

FAO Forestry

Forest time 1

OpenFORIS

z

EXAMPLES

FAO Forestry

OpenFORIS Toolkit

http://km.fao.org/OFwiki/index.php/Main_Page

Change in forest cover T 1 - T 2 Automatic change detection

z

Advanced image interpretation, open source platform

FAO Forestry

OpenFORIS Toolkit

http://km.fao.org/OFwiki/index.php/Main_Page

Automatic image normalisation

Time 1 Time 2 Time 2 Normalized

Open Foris Collect Earth: a new feature sampling with open source software and free Google internet resources

Open Foris Collect Earth Papua New Guinea: support to 1st phase NFI and LULUCF REDD+ activity data assessment

Countries preliminary priority Africa (4) DRC, Zambia, Tanzania, Uganda

South America (7) Colombia, Peru, Paraguay,

Ecuador, Nicaragua, Guyana, Mexico

Asia (2) Vietnam, Nepal

• Target priority countries based on actual capacity, FAO actually working in-country

Why is this different from other initiatives? • Access to open and commercial data (HR); • Training for data processing; • National control; • Hosted by FAO: representing the countries' interest; • Similar cloud-based and desktop functionality;

SDMS Partners

• Government of Norway • Kongsberg Satellite Services (KSAT) • NASA • Analytical Mechanics and Associates (AMA) • Global Forest Observation Initiative (GFOI)

SDMS Conclusion

• Improve connection between data / users / information products for REDD+

• Increase production speed of products required for MRV

• Open, flexible system for rapid and standardized image processing

• Building national capacity for autonomous creation of national statistics

• Counting on Sentinel-2 data

SDMS processing • 4 things make a SDMS feasible today:

• Free data • Free, sophisticated open-source tools • Increased processing power of computers • Increased speed of internet connection

SDMS Pilot SDMS Pilot to demonstrate a prototype cloud-based architecture • Timing: January to June • Test countries: three countries (Ecuador, Uganda and Tanzania)

already initiated to test the concept • Actual status:

– Set-up of platform with NASA/AMA actually ongoing – Set-up of high band-width internet connection in countries

• Goal: 1. Interactive selection and storage of optical land imaging scenes, 2. Storage of historical image coverage archive, and 3. Interactive and automated processing of stored imagery using an

existing set of algorithms/tools, 4. Distribution of data & forest information products to the 3 predefined

pilot countries

Project

• 3 year project • 4 countries in 2014 • 4-6 new countries in 2015 • 4-6 new countries in 2016 • Selection of 13 countries in which FAO will work anyway • FAO mechanism is expected to rationalize the work that

FAO already performs, thereby enhancing its capacity allowing to assist more developing forest countries in a more efficient way using the cloud-based architecture

Thank you for your attention! [email protected]