HSAT has now surveyed over 40,000 fields around the world and is building the world’s largest database of crops
Understanding food supply is critical to global markets and is becoming increasing complex due to the climate change. The traditional models for predicting food production are no longer working – with flooding in Pakistan, droughts in France, and heat waves Kansas
To resolve this a combination of satellite data and machine learning can be used to predict crops around the world.
Unfortunately, machine learning models require training data known as “ground truth“. The ground truth for crops and fields means physically visiting and surveying thousands of fields – and to do this at scale this means every crop in every country that predictions are required for.
Without this data the predictions will be inaccurate or impossible
Traditional Method – Very Slow and Very High Costs
Traditionally the process of ground truth collection was very expensive – over £50,000 per crop per country. This process also takes many months. These costs are so high and the time scales so low, this meant data was often not collected and the analysis could not be conducted.
HSAT Method – the “Uber Model” – Fast and Low Costs
HSAT has changed this , using an entire new model. We are now able survey thousands of fields a matter of days. As the speed increases and the costs decrease this means far more data can be collected and we can provide entirely new insights.
Like many ideas, AirBnB, Uber, etc – the idea is simple and can scale globally. However, they require technology and a lot of data engineering to put this in place. This is what HSAT is doing, working with local resources using local tech. This speed and scale of collection is a global first.
HSAT was awarded funding for this new method by the Norwegian Space Agency