Open Innovation Tackles Agricultural Challenges

Published Oct-22-18

A novel concept that applies gaming technologies to spot weeds.

Agri-Tech East, United Kingdom

The Story:

Open Innovation Tackles Agricultural Challenges Agriculture like many other industries has a lot of intractable problems. In an attempt to solve some of them a business focused cluster of organizations in the UK called Agri-Tech East hosted a hackathon in April 2018. This was the first time it had adopted this approach and hoped to bring a lot of diverse thinking to bear on three challenges. They were:

Challenge One: Supporting and enhancing traditional approaches to weed control.
Challenge Two: Making food production more accessible.
Challenge Three: Data integration to provide a more accurate assessment of growing conditions.

This open innovation event attracted participants from four continents and numerous disciplines. Coders, machine learning experts, farmers, plant scientists and more formed nine multidisciplinary teams and they spent a weekend on devising solutions.

The goals of the >sudo: grow hackathon were not only to solve problems by looking at them in new ways but also to speed up idea generation because innovation in the agri-food industry traditionally takes time to come to fruition.

Working on the Challenges

Two teams worked on the first challenge which was focused on new ways of supporting traditional approaches to weed control. One of the problems highlighted was that of black grass which infects more than one million hectares of arable crops in the UK. Over the years this weed has become more resistant to herbicides.

The winning solution was Weedsport which applies gaming technologies to artificial intelligence weed identification. It is a more sophisticated form of weed spotting, an essential practice that as the name implies spots weeds. If farmers don’t know where weeds are located it's more difficult to stop their spread. The WeedSpot innovation used a gaming engine to create 3D visualizations of weeds. To help train algorithms for weed classification synthetic visual data was generated supplemented by real-world data to validate.

Three teams took part in the second challenge to look at novel ways of making food production more accessible. Teams had to apply their thinking to vertical growing systems that use water and soluble nutrients instead of soil. While this allows for food production in unexpected places the system needs improved control systems to ensure optimum growing conditions for different crops. The winner was team Grow’s solution for an AI-controlled system with an easy to understand dashboard

The third challenge concentrated on how to tame and leverage big data as it applies to agriculture. Farmers capture a lot of data from numerous sources when monitoring and managing their farms. For example, crop yield information and data about soil moisture. However, the data is only really beneficial when it is specific enough for a particular location. An example given to participating teams was that of two weather stations that are six miles apart. One records rainfall and the other doesn't. So how do you decide which conditions are relevant for a particular field?

Four teams took part in this challenge and the winner was Durian Pi. Their solution combined existing data to identify new correlations and used interpolation techniques on these to predict weather conditions between 26 weather stations

Future Steps

The next steps are to turn winning concepts and prototypes into business opportunities. All have the potential to make a huge impact in agriculture.

Share on        
Next Story »

What Can we Solve for You?