A client needed to understand how much sugar was produced in Thailand ahead of the market. We built an entirely new methodology – providing unique insights for our client.
Thailand is a challenge due to the sheer scale and diversity of the country. It has small fields, a fraction of the size of those in Brazil, and is often covered by clouds.
Therefore to accurately predict the area of sugarcane being grown requires the following:
- High-resolution satellite data.
- This means 10 m resolution, not 250 or 60 m.
- High-frequency data:
- Collecting the data every five days, not once a month
- High volumes of ground truth
- This means 2000 to 4000 fields visits and not 50 to 100 fields
The data set resulted in trillions of data points, which required highly sophisticated methods to collate and process.
We built a unique solution to achieve this, at a fraction of the normal costs – using entirely new methods of collecting and processing data on this scale.
This resulted in several global firsts and the ability to predict the crop area for Thailand far ahead of the market.
HSAT provided accurate predicts of the crop swing and total area 5 months ahead of the market. HSAT achieved this by using hundreds of terabytes of data and visiting over 3,000 fields to build models with very high levels of accuracy.
A client needed to understand how much sugar was produced in Thailand ahead of the market.
Thailand is a challenge due to the sheer scale and diversity of the country. It has small fields, a fraction of the size of those in Bazil, and is often covered by cloud.
Therefore to accurately predict the area of sugarcane being grown requires the following:
- High-resolution satellite data.
- This means 10 m resolution, not 250 or 60 m.
- High-frequency data:
- Collecting the data every five day, not once a month
- High volumes of ground truth
- This means 2000 to 4000 fields visits and not 50 to 100 fields
The data set resulted in trillions of data points, which required highly sophisticated methods to collate and process.
We built a unique solution to achieve this, at a fraction of the normal costs – using entirely new methods of collecting and processing data on this scale.
This resulted in several global firsts and the ability to predict the crop area for Thailand far ahead of the market.
HSAT provided accurate predicts of the crop swing and total area 5 months ahead of the market. HSAT achieved this by using hundreds of terabytes of data, and visiting over 3,000 fields to build models with very high levels of accuracy.