Driving up accuracy of maize classification to exceed 99%

Alcis is a geographic information services company that works to improve the lives of the world’s most vulnerable people through innovative and world-class Geographic Information Systems. It specialises in supporting fragile and conflict affected states.

The company was founded in 2004 to support UK government efforts around crop production in Afghanistan, where, through the use of remotely sensed imagery and other geographic data, it was able to provide unique insights on what was happening on the ground.

The company is still very active in this country, primarily working alongside agricultural development programmes to gather relevant, accurate and timely data that helps international donors plan and implement effective interventions to provide maximum assistance to the beneficiaries.

Collaboration delivers geospatial data analysis 

Alcis was looking to improve the understanding of agricultural cycles and practices for a maize mapping and sustainable water use project in rural Afghanistan. To execute this project, Alcis was looking for a partner that could provide geospatial data processing expertise for high quality analysis of satellite images.

By collaborating with the University of Surrey, funded by the national SPRINT business support programme, Alcis developed a method for mapping maize crops in past crop cycles, without the need for ground truth data, using ESA’s Sentinel 2 imagery, and knowledge of the nature of maize growth and climate variables.

The project also developed a new spatial product, derived from globally available Earth Observation data, that could inform on the climate change impacts on water resources in Afghanistan, the water consumed by changing agricultural practices and the sustainability of this consumption in the context of climate change.

Using remotely sensed satellite data, the project developed signatures for the maize crop based on its phenology (natural growth cycle) and the local meteorological measurements. These signatures have been used to identify this crop at scale across the north of Afghanistan. Following successful trials, the project then sought to develop this data across the entire country.

Benefits achieved as a result of the SPRINT project 

• Successful development of novel spatial technology that can be scaled up for other global projects
• Has enabled accuracy of maize classification results of over 99 per cent
• Follow up project planned for over 100 locations in Uganda to map and monitor maize production that will inform global food programmes operating in the country

Scaling up technology for Ugandan project

Alcis is now scaling up the Afghanistan SPRINT project by applying the same technique developed with the University of Surrey in a new project in Uganda. The company will work with local farmers to map and monitor maize production in around 100 locations in northern Uganda.

The environment in Uganda poses some issues – although there’s a lot more water and it’s easier to get field data, there is more cloud cover for the satellite images. However, the value of the crop data collected will not only potentially help external bodies such as Unicef, government departments and the UN’s world food programme, but it could also reduce the risk of damage to crops from the likes of locusts or army worms.

A springboard into new markets

Tim Buckley, Chief Operating Officer at Alcis said: “Alcis’ clients work in fragile, conflict-affected areas, including rural Afghanistan, where getting good quality data is a major challenge. Our driving force is using cutting-edge technology for community-based applications to change people’s lives for the better. Longer-term, we’re expanding our offerings with other crops in other countries, increasing our footprint geographically, and we believe that this SPRINT project will be a springboard into new markets for us.

“The University of Surrey has such a strong reputation for space research, with a rigorous approach to data analysis and a willingness to experiment. The knowledge, information and theory that is the bedrock of academia can be of vital use in real-world applications, particularly in the space sector. Our relationship with Surrey has enabled us to accelerate and develop our processes to make them even more efficient.

“We’ve achieved some really good results from our SPRINT activity, and Belen and her team at Surrey have been just fantastic with their expertise, time and support. Without the collaboration with Surrey, this whole model might just have stayed in the R&D pile but their engagement has allowed us to stay focused.

“The new technique will be based on Sentinel imagery collected every ten days so it really is a novel technology, as no other data collection exists from other commercial vendors. Consequently, there is real potential to take this as yet undeveloped Ugandan model and put it into other countries. And that’s thanks, in major part, to SPRINT and the University of Surrey team.”

Bringing big data processing capabilities