“Sensors”, GSW 2016

Jun Chen1,2 1Chief Scientist, National Geomatics Center 2President, ISPRS 25 May, 2016 , Rotterdam, The Netherlands Contents

Background

Methodology

Organization

Outlook Validation of Land Cover Maps

1 2 3

Image pre- classification processing Verification

Validation LC product

Documentation of the accuracy of land cover maps allows users to evaluate the utility of these maps for their particular applications GlobeLand30 - the World’s first 30m Earth Land Cover Map

www.globeland30.org

Chen , et.al., 2014, : Open access to Earth land-cover map, Nature, 514:434, 23 Oct. 2014 POK based Operational Mapping

30m imagery NDVI time-series Reference data

PerPer--classclass PixelPixel-based-based classification

extractionextraction Spectral Textural Multi-temporal Water features features information

Wetland Supervised classifiers/Decision trees

Snow & Ice ObjectObject-based-based Identificationidentification Artificial cover Multi-scale Automatic Object-based segmentation labeling result Cropland

KnowledgeKnowledge-based verification Forest Knowledge Manual Quality control Shrubland integration interaction & and product Grassland Barren land platform verification integration

Tundra Global land cover mapping at 30m resolution

Chen et.al., 2015. Global Land Cover Mapping at 30m Resolution: a POK-based Operational Approach, ISPRS Journal of Photogramme try and Remote Sensing , 103 (2015): 7-27

5 Geographical Distribution of Users

111 countries and 6000 users Geographical Distribution of Users

Download Frequency Preliminary Validation of GlobeLand30

Sample accurac Investigators sources points y and 5 Chinese Full class 154,070 83% institutions

(Prof. Y-F GIM, Dec., Water ,N. 91% Ban) 2014 Europe Overall accuracy,%

IIASA Forest Report JAXA forest map 85

Hansen's TC 87

GlobeLand30 88

Hybrid map 90

82 84 86 88 90 92 Validation at individual countries

Country Areas Accur. Sources (Investigator) Whole Germeny 92% Int. J Digital Earth, (Dr. J. J. Arsanjan) (CORINE, UA, 2016(on line) OSM, and ATKIS) Greece Thessaly Region 91% Land, 2015, 4,1-18 (Dr. Ioannis Manakos) Iran 6 study sites 77.9% Habitat International (Dr. J. J. Arsanjan) ,2016,1-7 8 areas >80% Remote Sensing, (Prof Maria Broveli) 2015(7), 2107-2122 Towards a Comprehensive Validation

 Separate and independent accuracy assessments of some 30-m products being made

 A comprehensive validation of multiple 30-m GLC maps would allow comparative assessments of uncertainty among different maps Output from an GEO/ISPRS Workshop (6-7 June 2015, Shanghai)

The workshop "Trust in Spatial Data and Validation of GLC Products" discussed the issues related to the validation of 30m GLC products with suggestions:

 this international validation activity be led by GEO and ISPRS, co-organized by some other organs (i.e., UN- GGIM,…).

 all the existing 30m GLC products (including GlobeLand30) be validated with the same approach.

 be conducted from Aug. 2015 to the end of 2016. Contents

Background

Methodology

Organization

Outlook Major tasks

• Formulate a technical specification suitable for the validation of 30-m GLC data , including the sampling design, response design and analysis protocols

• Develop advanced tools for facilitating the validation processes • Mobilize GEO members and participating organizations, including Integrating all available validation resources and expertise • Organize collaborative validation and report to GEO plenary Major Methodology

The Existing The Top-level Designing of Verification Global Land Cover products Technology in 30 m Verification Conceptual design

Estimation of Sample Integration of Layout of Sample Online Check Accuracy Assessment · Confusion matrix  Size · Regular grid Resources · Interactive Identify or develop · · · Area-weighted overall Two-ranking sampling · Landscape index Semantic mapping interpretation · Multiple stratified · Scale conversion accuracy · Spatial correlation · Credibility sampling · Phase integration · Kappa suitable methods for 5 key items

Technical Specification · R&D and Testing Online Sampling scheme  · Sample check Formulate a technical Verification System · Accuracy assessment · Online geo-tagging specification (guidelines) Develop online tools Global Overall Regional Accuracy Verification Assessment A Landscape-Index Based sampling approach

Countries with different spatial heterogeneities

Sample Country rLSI Sample size density  SWS-Shannon index-based Mauritania 2978 29 cluster sampling

Spain 1461 29  SRS-stratified random SRS Bangladesh 426 29 sampling Mauritania 1910 19 SWS 2115 42 More reasonable sample Bangladesh 460 31 densities and sample sizes Mauritania 1.29 1579 15 derived using the LSI approach LSI Spain 2.68 1627 32 Bangladesh 3.46 617 42 On line Validation System Collaborative Validation

Global Validation

Coordination

Region based Country based Continent Based Feature based validation validation validation validation

Sample plan Sampling Validation Resources Assessment

· Sample plan for · Collaborative · GLC products · Assessment report for region collection of · HR images region · Sample plan for samples from · Aerial photographs · Assessment report for country world · Geo-tagged pictures country · Sample plan for · Distribution of · Geo-wiki/DCP · Assessment report for continent samples to world · In-situ measures continent · Sample plan for · Assessment report for feature feature · Global assessment report

Technical Specification Validation support tools Contents

Background

Methodology

Organization

Outlook Key Players

 NGCC

 GOFC/GOLD

 IIASA

 Tongji Uni.

 KTH

 CO-DATA

 ISPRS

 …

Collaborative Validation

Initiativing Validation A GEO-led Validation

 Establish a GEO-led collaborative framework for the validation of GLC data products

 Complete the validation (assessment) of all available multiple 30-m GLC maps 40 countries and PO joined

 25个GEO members and 15个others( UN-GGIM、CO-DATA) Countries Participated

序号 国家/组织 序号 国家/组织 序号 国家/组织

1 14 UAE 27 Mongolia 2 15 28 Thailand 3 China 16 Sweden 29 EU- JRC 4 17 Rwanda 30 FAO 5 Ecuador 18 Uganda 31 ISPRS 6 Laos 19 Indonesia 32 IIASA 7 Greece 20 33 CODATA 8 Italy 21 Mexico 34 GOFC-GOLD

9 Japan 22 Argentina 35 PAIGH 10 Portugal 23 Shirilanka 36 RCMRD 11 24 Tanzania 37 UNGGIM

12 South Africa 25 Columbia 38 UNESCAP 13 Ukraine 26 Jordan 39 ECA Workshop on Validation of GLC Datasets June 28 – 30, Tongji University, Shanghai

http://celiang.tongji.edu.cn/GLCVal2016/Home.html

 , Generic approaches for validation of land cover datasets  Validation schemes for GLC datasets at various spatial resolutions  Allocation of global samples in the case of heterogeneity and correlation  Accuracy assessment in remote sensing-derived GLC datasets  Development of online tools for validation of GLC datasets  Identification of scientific/technological/application challenges related to GLC validation  Trust and degree of trust in validation of global land cover datasets  Others

A Special Issue on ISPRS J Geoinformation

Validation on Global Land Cover Datasets

 present the new concepts and outputs of GLC datasets validation, including algorithms and models of validation schemes, sample collection and accuracy assessment, as well as on-line validation and tagging tools.

 not limited to the GEO led international validation campaign.. http://celiang.tongji.edu.cn/GLCVal2016/Home.html Contents

Background

Methodology

Organization

Outlook Outlook

 Establish a GEO-led collaborative framework for the validation of GLC data products

 Report both the validation results and the collaborative validation framework to GEO community

 Lead to a coordinated regular validation practices

Thanks

Thank your attention! www.globeLand30.org  [email protected]