How Data Can Help Feed the World

The world’s agricultural sector faces a big challenge: How to feed a growing global population while climate change is making the conditions for cultivating crops less favorable, with escalating droughts, heatwaves and floods – and how to do that sustainably.

Data can help. Akkodis has partnered with a major client to bring better data analysis to the agricultural sector, using a combination of artificial intelligence, machine learning, sensors, Internet of Things technology and robotics to create an integrated agriculture data platform that can provide insights into how crops grow and help farmers achieve better harvests.

That’s a vital challenge at a time when rising temperatures caused by climate change are negatively affecting crop yields around the world – and when global agriculture will need to produce more food in the next 50 years than it has in the previous 10,000 in order to sustain the growing world population.

Sustainability Mission

“The overall mission is sustainability,” said Anthony Nguyen, Technical Delivery Manager at Akkodis. “Agriculture is still very archaic when it comes to the tech that is used. It is about finding out how to use machine learning and AI to make that data more available to the regular farmer.”

Bringing together different datasets on one platform can give famers vital information. Analyzing images of pests on crops can help them decide when to apply pesticides and which ones to use. Information on the size of leaves on plants in the field can help farmers adjust the start date for harvesting, improving the quality of their yields.

“The quality of farming is important. Given the climate situation, it can help to understand how much their yields are going to be and what they can do to improve the yield,” said Nguyen.

“It’s a large, scalable operation across various fields and various scenarios,” said Kishore Raj, Vice President & Partner: Technology and Engineering Consulting and Services at Akkodis.

An important part of the project is the development of high-tech prototype rovers which trundle through fields, taking high-quality images of the crops. These are then combined with other data sources such as weather or soil information and satellite imagery.

“Data is really critical,” said Kishore. “There’s a large scale of data collection to be done to build the analytical capability and the artificial intelligence capability around it.”

Designing machines that can work efficiently on sometimes inaccessible farmland and collecting data from different sized crops without damaging them is no easy task. The rovers are now in use in hundreds of farms across North America and next steps will include rolling them out to other regions too. The bigger the scale of the project, the more data can be collected and the greater the accuracy of the findings, explains Nguyen.

Building Up the Knowledge Base

The project makes use of Internet of Things technology. “To be more real-time it does a lot of data processing within the machine itself, comparing it with the database in the cloud,” said Kishore. “We’re constantly building a lot of knowledge.”

Achieving scale was a key reason behind entrusting the data collection part of the project to Akkodis, according to Kishore. “Our client was a small start-up working at a high velocity. They didn’t have that scale and wanted to bring in a supplier that could help them scale up,” said Nguyen.

Akkodis’ collaboration with the client began with just a few experts working on the project, but more than 20 people are now working on it in high season and the client team expects help on the technology side with development and process improvements.

In addition to giving farmers the tools to improve yields, the platform can help improve the quality assessment process as the government tightens certification standards for some food products.

A particular focus at the moment is berries.

Being able to supply data about the berries instantly to certifying authorities from the connected farm makes it easier for the grower to obtain the certification needed and charge a better price for their fruit. Having access to the information on the database while the crops are still growing can also help reduce waste, a key way to improve the sustainability of the broader agricultural sector.

The team first builds up the knowledge base, scanning images of different qualities of berries into the database using a handheld device. “The more data we collect, the more scenarios come in, the more knowledge is available,” said Nguyen. The machine deployed in the field compares its real-world findings with what is already on the database.

That quality certification mission is likely to be broadened into other types of crops soon.

Efficient Data Analysis

“We train the rover and once it’s trained it knows what to look for. Instead of guessing you’re now using machines that are far more efficient when it comes to analyzing the data. The mission is building up knowledge about farming – there will be a commercial model later,” Nguyen says.

The progress made during the project feeds into a broader trend within the agricultural sector, Kishore says, as large farming equipment manufacturers aim to make their harvesters more intelligent, with a view eventually to having them run autonomously. “They are going big on digitalizing those machines, so they need to be fed with data, with knowledge.”

One development as agricultural data collection technology matures may be to reduce the size of the equipment responsible for gathering the data in the field. “There is a move away from the specific rover, moving to a smaller data collection platform, with the specific aim of being able to scale this out to smaller farms,” said Nguyen. “There are cost and infrastructure limitations for a big rover. We’re migrating towards something more the size of a desktop computer attached to a current piece of apparatus such as a tractor or harvester.”

“Sustainable farming is very important for all of us and technology adoption is very fast in this sector,” Kishore adds. “There is a high level of investment both on the public sector side and private companies investing in this space. The technologies are established but it’s about making it mature.”