How an agri-tech algorithm helped G’s Growers predict demand for lettuces

Workers harvesting iceberg lettuce for G's Group.

Workers harvesting iceberg lettuce for G's Group. - Credit: G's Group

Collaboration with mathematicians has helped a Fenland salad grower to forecast customer demand 'better than the retailers' – a move which could save millions in wasted produce.

John Shropshire, director of G'S Growers.

John Shropshire, director of G'S Growers. - Credit: G's Group

G's Growers, based near Ely, is one of the largest producers of salad and vegetables in Europe. Crops such as iceberg lettuces are subject to all the usual farming variables of soil type, weather and water – but, unlike other agricultural crops, they cannot be stored or frozen.

Salad consumption is also dependant on weather, as a 'barbecue weekend' will boost demand, while a wet one depresses it. So the ability to predict both production supply and consumer demand is crucial to the business, as it aims to get crops ready for retail at the precise moment they are needed, with minimum wastage.

Director John Shropshire said a discussion with Microsoft at an Agri-Tech East meeting led to a contact at Cambridge University who put the company in touch with mathematicians at the Smith Institute in Surrey.

Together, they have developed a sophisticated monitoring system to measure the growth of iceberg lettuces and collate data on temperature and humidity within specific microclimates. Along with historical weather data and forecasts, it is fed into the IceCAM (Iceberg Crop Adaptive Model) algorithm which enables the firm to identify key growth stages and amend production schedules to mitigate against potential shortfalls.

A vegetation map created by aerial imagery from the Agri Eye system pioneered by Fenland-based growe

A vegetation map created by aerial imagery from the Agri Eye system pioneered by Fenland-based grower G's. - Credit: G's Group


You may also want to watch:


'We are now able to predict demand better than the retailers,' said Mr Shropshire. 'We are basically building a model that enables us to forecast how much lettuce we will have to harvest, before we have even sown it. That gives us a more accurate selling plan, which is important because we cannot store our product.

'We are growing lettuce in four countries in Europe, so potentially this could save us millions of Euros. Until now, we have been over-growing them by 30pc to make sure we always have enough. If we can get that down to 20pc, at the scale we are working, that is quite a big saving, both in money and environmental costs.

Most Read

'Agri-Tech East put us in touch with the best brains for working out the algorithms to give us a more sophisticated and accurate system. Before this, we had someone working it all out on a calculator.'

AERIAL DATA

The IceCAM algorithm at G's also uses aerial survey data collated by the firm's Agri Eye project, which uses light aircraft equipped with specialised cameras.

Both projects are run by precision farming manager Jacob Kirwan, who said: 'Agri Eye started separately from IceCAM but we are joining it together now, so we are getting vegetation maps with a colour scale showing how well crops are growing in different areas of the field, as well as counting and sizing functions which tell us how many plants will be viable at harvest.'

Mr Kirwan said his interest in food security developed during his biology studies at Leeds university. 'I think universities need to offer more guidance on what agriculture can offer to engineers and scientists, rather than just broad agricultural studies,' he said. 'I think eyes need to be opened on both sides, and we need to meet in the middle to make agriculture a more attractive job.'

Become a Supporter

This newspaper has been a central part of community life for many years. Our industry faces testing times, which is why we're asking for your support. Every contribution will help us continue to produce local journalism that makes a measurable difference to our community.

Become a Supporter
Comments powered by Disqus