1 of 17

EO4GEO - Shaping the future workforce of the space geospatial sector in Poland

Part 1: Tackling the needs of the Polish business community: an overview of EO4GEO tools (21st October, Podkarpackie)

Part 2: EO/GI skills development and capacity building in the Polish education (22nd October, Mazovia)

https://www.nereus-regions.eu/2021/06/22/registration-is-open-eo4geo-workshop-21-22-october-2021-shaping-the-future-workforce-of-the-space-geospatial-sector-in-poland/

2 of 17

�

Online: https://www.nereus-regions.eu/wp-content/uploads/2021/06/Draft-Agenda_Poland-1.pdf

Draft �agenda

3 of 17

BoK tools presented by Rob Lemmens (University of Twente), Aida Monfort Muriach (Universitat Jaume I), Florian Albrecht (University of Salzburg)
EO tools presented by Hande Erdem (VITO)

Agenda – Day 2 – EO4GEO Hands-on session

4 of 17

The EO4GEO Body of Knowledge – annotated workflows

University of Salzburg,

Department for Geoinformatics – Z_GIS

Florian Albrecht

EO4GEO Hands-on session

Users of EO

Forestry

Agriculture

Local & regional planning

Emergency & social protection

5 of 17

Development

Builds on previous initiatives from GI
Addition of detailed EO part
Addition of market oriented perspective

Current status

381 concepts from GI
520 concepts from EO, of which 71 describe �EO services and their user communities
901 concepts in total (285 fully described so far)

Access via https://bok.eo4geo.eu/, more information �on http://www.eo4geo.eu/bok/

The EO4GEO Body of Knowledge

Version 5 of the EO4GEO BoK

5

6 of 17

Example BoK concept “Normalized Difference Vegetation Index (NDVI)”

Unique ID	IP3-1-2-3	https://bok.eo4geo.eu/IP3-1-2-3
Description	what the concept is about	“... The reflection of plants in red and near infrared parts of the electromagnetic spectrum indicate plant maturity or plant stress. ... The NDVI is calculated as: �NDVI = (NIR-R) / (NIR+R) …”
External resources	referenced literature, further reading	Campbell, J.B., & Wynne, R.H. (2011). Introduction to remote sensing. Guilford Press Lillesand, T. and Kiefer, R. (2003). Remote sensing and image interpretation (5th Ed.), New York: Wiley. [...]
Skills	things you can know, learn, apply	- Analyse vegetation growth & development on a land parcel with an NDVI time series derived from EO - [...]
[...]

IP3-1-2-3 Normalized Difference Vegetation Index (NDVI)

6

7 of 17

Skills represent tasks (and sub-tasks) of a workflow

IP3-1-2-3 Normalized Difference Vegetation Index (NDVI)

NDVI time series

NDVI map

H. Erdem, VITO, 2021.

Analyse vegetation growth & development on a land parcel with an NDVI time series derived from EO data

Create NDVI time series �diagramme

Integrate optical and SAR data to �enhance NDVI time series

Analyse diagramme for... �[insert application domain]

1

2

3

7

8 of 17

Skills forming a workflow for vegetation �growth analysis on a cloud platform for EO

Download

Raw EO data

Prepare

Upload

Imagery & parameters

Select optical imagery from a satellite sensor with spectral bands suitable for vegetation mapping (e.g. Sentinel-2)

PS1-7 Optical spectrometers

IP3-1�Band math

Apply script that calculates NDVI using band math on the selected data cube of images

Apply advanced script that fills cloud gaps with �NDVI-like information derived from SAR data

IP3-11 Time series analysis

IP3-1-2-3 Normalized Difference Vegetation Index (NDVI)

PP1-3-4 Spectral Signature of Vegetation, Water, Soil

Select a land parcel and create its time series diagram of SAR-enriched NDVI

Analyse diagram for ...

IP5-1 Data cubes

PS1-3-1-1 Imaging Radar

[PP2-2] Interaction of microwaves with matter

Access & use

Cloud platform

& define time period and area of interest

*

1

8

Workflow that focuses on providing EO-based information products for Agriculture and Forestry that are based on data cubes for time series analysis and the fusion of S1 and S2 data products

Wenns um den Vergleich des Workflows mit dem allgemeinen Fernerkundungsablauf geht, dann würde ich mit Sen2Cube wohl am ehesten zeigen können, dass die zunehmende Automatisierung der Bildverarbeitung den Fokus der manuellen Auswertung auf die Entwicklung von Abfragen verschiebt. Es würde trotzdem voraussetzen, dass sich der Anwender ein Grundverständnis der enthaltenen Verarbeitungsschritte erarbeitet um die Qualität seiner Abfrageergebnisse zu beurteilen.

Hello Hande,

I am preparing the example case for our Poland workshops where I relate a workflow to the concepts of the EO4GEO BoK.

I need some clarifications from you about your Azores-Presentation (slide 13).

My objective is to point out applications for agriculture, forestry and local/regional planning. I think this is the most valuable for the Polish regions.

Last meeting we discussed to include fusion of S1 and S2. I remember your Azores-Presentation (slide 13) included the CropSAR FAPAR that does this fusion (explained on the Vito blog; copernicus land on FAPAR).

What can I get out of the time series on slide 13? It shows a field that had a winter crop, then the field was tilled and from early June, a new crop started to grow which was then harvested in late September. Am I right?

- For agriculture this allows monitoring vegetation health and detecting harvesting time.

- For forestry, this allows to monitor forest health? (Check the reference of: Lechner, A. M., Foody, G. M., & Boyd, D. S. (2020). Applications in Remote Sensing to Forest Ecology and Management. One Earth, 2(5), 405–412. https://doi.org/10.1016/J.ONEEAR.2020.05.001 )

- For local/regional planning, this allows to monitor the progress of soil sealing through land planning, i.e. land use change from green areas to settlement areas. (Guidelines to best practices to limit, mitigate or compensate soil sealing - European Commussion)

Did I understand your slide 13 correctly? Do my purposes make sense for which the CropSAR FAPAR can be applied? Maybe you have some ideas that make it more specific what a farmer, a forester or a regional planner can do with the data?

Kind regards,

Florian

9 of 17

Advanced skills for fusion of optical and SAR �data for supporting vegetation growth analysis

SAR Pre- Processing

Sentinel-1�GRD SAR

Upload

Cloud platform

PP2-3-6 Principles of Synthetic Aperture Radar (SAR)

Holtgrave et al. 2020

Prepare RVI script

Download

Apply band math to calculate the Radar Vegetation Index (RVI) from backscatter pulse of horizontal and vertical polarisations for all selected SAR images

Create the new script “Fill cloud gaps in NDVI” on the cloud platform that does the following EO processing steps:

Use regression to fill cloud gaps in optical NDVI maps with RVI-derived NDVI estimates

Perform quality check for script “Fill cloud gaps in NDVI” and publish it on the cloud platform for public use

Select SAR imagery (e.g. from Sentinel-1) for a given time period and area of interest

Apply CNN analysis to estimate the relationship between RVI and optical NDVI maps

IP3-1�Band math

IP3-4-6-1 Convolutional neural networks (CNN)

Vito-Blog on CropSAR 2019

PP2-1-4 Polarisation

IP3-4-6 �Deep learning

PP2-2 Interaction of microwaves with matter

2

9

10 of 17

Applications for time series analysis of vegetation parameters

… identifying new impervious areas to monitor soil sealing policies

… identifying meadow harvesting times to ensure conformity to CAP regulations

… identifying insect-induced forest damage for forest health monitoring

TA11-5-2�Users�in local &

regional planning

TA11-1-3�Users�in forestry

TA11-1-1�Users in agriculture

Analyse diagramme for ...

H. Erdem, VITO, 2021.

Common Agricultural Policy (CAP)

Mountain Pine beetle, CSFS

Barta et al. 2021

Google Maps

Augustin et al. 2019

3

10

Agriculture: diagramme presents plant growth and changes to bare ground

Forestry: diagramme presents how the NDVI decreases when insects start to affect trees

Soil sealing: diagramme presents locations where vegetation growth ended and did not return again

Case 4 - Crisis management:

[TA11-6-2] Users in emergency & social protection
Near Real-Time Wildfire Progression Monitoring with Sentinel-1 SAR Time Series and Deep Learning https://www.nature.com/articles/s41598-019-56967-x
… identifying current state of wildfire progression for forest fire fighting
challenge: near-real time requires

high frequency monitoring (up to one hour)
fast (i.e. automatic) processing
mobile availability with quick and easy access (and very low delay times)
easy-to-understand map visualisations

Option 2 for Case 4:

Fully Automated Countrywide Monitoring of Fuel Break Maintenance Operations
https://www.mdpi.com/2072-4292/12/18/2879

Agriculture:

Hande-Erdem_Terrascope-EO-tools.pdf slides 12 or 13
https://blog.vito.be/remotesensing/cropsar2019

Forestry:

https://www.sciencedirect.com/science/article/pii/S0303243421000428

Local & regional planning:

[sen2cube.at]

11 of 17

http://www.eo4geo.eu/tools/

11

12 of 17

EO4GEO BoK - Living Textbook - Hands-on instructions

12

13 of 17

Define educational offers based on the BoK

Academia, training providers

Common vocabulary for knowledge and skills ( BoK )

Explore most offered Knowledge and Skills

http://www.eo4geo.eu/tools/cdt user: poland@eo4geo.com pwd: Poland

13

14 of 17

Define job offers based on the BoK

HR offices from private companies, public institutions, or academia

Job and Training offers based on the BoK

Explore most demanded Knowledge and Skills

http://www.eo4geo.eu/tools/jot user: poland@eo4geo.com pwd: Poland

14

15 of 17

Annotate any document with BoK concepts

All types of audiences

Annotate CVs, papers… with the Body of Knowledge

Use this annotated resources in the other eo4geo tools

http://www.eo4geo.eu/tools/bat

15

16 of 17

Compare any BoK annotated resource

All types of audiences

Find best matching job offers to an annotated CV

Rank educational offers according to an occupational profile

Many more!

http://www.eo4geo.eu/tools/bmt

16

17 of 17

Tools info:

http://www.eo4geo.eu/tools

Test user: poland@eo4geo.com

pwd: Poland

Thanks!

17