Technology to predict crop types.
Crops in every field farmed in Scotland have been recorded by satellite imagery to create a new interactive map to helping to breakdown agricultural land use.
The Scottish Crop Map uses data from 2019 to predict the crop types using radar images from the European Space Agency (ESA) Copernicus Satellite Programme and to recognise the crops growing in nearly 400,000 fields in Scotland.
The map has been developed by the Scottish Government’s Rural & Environmental Science and Analytical Services (RESAS), working in collaboration with EDINA at the University of Edinburgh, who built a data ingest process for the satellite data and provided expertise and support for modifying the data pipeline, and the Joint Nature Conservation Committee (JNCC) to utilise new statistical and computing science techniques.
Rural Affairs Secretary Mairi Gougeon said:
“This map is a very exciting development and, along with other digital innovations, will provide greater insight into rural land use. Future iterations of the map, combined with other datasets, could have a significantly broader scope of use, providing real time data that would help with community and farm level land-use planning and management.
“The statistics we have on Scotland’s agricultural sector typically rely on surveys being completed by a large number of farmers. Gathering data in this way would reduce the amount of time our farmers and crofters have to spend reporting this.
“In order to help improve the accuracy of the map, and before further developments are made to track minor crops, farmers and crofters are being asked to review the current map and report any inaccuracies in the crop identified. Engagement with farmers and crofters with their local knowledge is vital and will help develop future iterations of the map and enable us to extend and enhance what the map could do.”
Gavin McLachlan, Vice Principal and CIO University of Edinburgh said.
“EDINA has been working with the agricultural sector for many years, and we were pleased to collaborate on this important project with the Scottish Government and JNCC. As well as providing satellite data expertise to develop the project, EDINA is also a champion of ensuring data is understood by wider audiences. Developing the spatial presentation of the data, to provide maximum benefit to the farming community, was an important element of the project.”
The statistics from the production of this map should be used with caution while the methods and techniques are still under review. However analysis of the results predict that the model was working with an accuracy of 85% and above 90% for barley and wheat.
As this is a first iteration of the model, the interactive map and accompanying documentation will be released under the ‘Experimental Statistics’ banner to allow further collaborations and consultation with agricultural and academic experts to refine the model. This will allow RESAS to agree future requirements and agreements on methods and standards used in developing the map.
Based at the University of Edinburgh, EDINA is a word-class centre of excellence, working closely with customers in the private, public and education sectors in geospatial and satellite data, computational learning, text and data mining, e-preservation and machine learning.
With over twenty years’ experience, EDINA has a proven track-record of collaborating on innovative projects and service delivery across the full digital lifecycle; employing an agile, customer centric approach, to design, develop, test and deploy scalable digital solutions.
There is a problem
Thanks for your feedback