I. Objective  

To share our experience in collecting and using ground truth data for generation of agriculture statistics, and exploring opportunities to improve ground truth data collection, use, and exchange  

II. Participants 

Experts invited include GIS and statistics experts from national governments, academia, the private sector, and international organizations.  

III. Background 

Over the last few years, there has been a shift towards increased use of Earth Observation (EO) data to produce official statistics, particularly in the agriculture sector. This is in response to the increased costs of traditional questionnaire-based surveys, declining survey response rates, and increased availability of open source, high-resolution satellite imagery. As a result, national statistics offices throughout the world, including in Asia and the Pacific, are expanding their use of EO data to produce agricultural statistics, such as crop areas, yield estimates, and forest cover.  

As the costs of image acquisition fall and the technical capacity to process images increases, one of the most significant costs comes from the collection of ground-truth data. Ground-truth data are in-situ data that verify – typically using sampling techniques – the commodities and land areas identified or estimated by the image processing algorithms. This ground truth data is used both to improve the algorithms, and to validate their results. 

Yet despite the importance of ground truth data, they are rarely disseminated as a data product in themselves. This results in a plethora of study-specific ground truth data sitting in individual computers and servers.  

At the same time, traditional agriculture censuses and surveys may collect data that could be used as ground-truth data, sometimes with marginal changes in data collection, but remain unused for this purpose.  

This results in an untapped opportunity to share ground-truth data as a global public good, and improved use of survey and census data as a source of ground truth data. 

It is also noted that many countries through their agricultural statistics systems regularly collect agricultural data that may be relevant for remote sensing specialists to help inform their estimates. 

This, however, requires further collaboration to better understand the needs of each and how the data already being collected in official statistics could better be collected and disseminated for the benefit of the remote sensing community and agricultural statisticians alike. 

As many countries in the region are gaining experience with remote sensing, an expert group meeting focused on ground-truthing data collection and dissemination would serve as a platform to enhance collaboration in this area. Participants will be able to share their experiences with ground-truth data collection and open a dialogue for establishing ground-truth data dissemination standards and methodologies. The expert group meeting intends to explore ground-truthing practices in the region and to realize how ground-truthing data can be made more accessible.  

IV. Format  

Two to three half-day virtual meetings to discuss the following issues:  

• Data Collection (format, sources, tools, content)
• Data Use (minimum. variables)
• Privacy considerations / data publication  

Interested experts are requested to present short papers on one of the above topics. The best of  these papers may be proposed for a session in the 9th International Conference on Agriculture  Statistics to be held in the United States in 2023 (conference dates will be updated). 

V. Duration 

The virtual meeting is tentatively scheduled for three half-days from 13-15 September 2022.  

VI. Expected Outcomes 

• Exchange knowledge
• Begin discussions on how to share and exchange ground truth data subject to privacy  protection
• Produce a regional publication on the shared experiences, 
• Propose a session for the next International Conference on Agriculture Statistics  

VII. Participating in the Expert Group Meeting 

If you are interested in contributing a paper or participating, please complete the following registration form at: https://forms.gle/8TMDW3AQ6iuV9r3GA  

VIII. Contact 

For more information please contact: 

Sangita Dubey: [email protected] (Regional Statistician for Asia-Pacific, FAO) 

Kavinda Gunasekara: [email protected] (Associate Director, Geoinformatics Center of AIT)