X
Innovation

CSIRO using artificial intelligence to map 1.7m Australian grain paddocks

It developed ePaddocks for the agriculture sector to better understand the boundaries of grain paddocks across the country.
Written by Asha Barbaschow, Contributor

The Commonwealth Scientific and Industrial Research Organisation (CSIRO) has been using artificial intelligence (AI) to map Australia's paddocks from space.

Through the use of AI, scientists have identified the boundaries of around 1.7 million individual paddocks in Australia's grain growing region.

CSIRO has developed the AI-based technology into a new product, ePaddocks, which it touts as saving time for farmers and those in the agricultural sector when using digital services for farm analytics and insights.

ePaddocks is available as a dataset shapefile containing the boundaries of crop paddocks at national spatial resolution.

CSIRO said the product can be used for paddock-level monitoring, crop identification, and rural intelligence and portfolio analysis.

"The ePaddocks algorithm analyses satellite images of mainland Australia's agricultural areas to locate the boundary between the paddock and other paddocks and farm management areas," CSIRO explained. "We also teach it to recognise paddock boundaries regardless of their orientation and level of zoom (a process we call data augmentation)."

See also: Farmbot delivering remote watering solutions to Aussie farmers

CSIRO said ePaddocks can identify paddock boundaries from season to season but doesn't identify a particular property, landowner, or what paddock belongs to whom.

Prior to ePaddocks, users of farm management software were required to manually draw paddock boundaries for every service they used, such as satellite-assisted fertiliser application or crop growth monitoring. Users were also required to manually update this information every growing season.

"The satellite images we use, although publicly available, are cumbersome to download, store, and analyse by the average person," CSIRO remote sensing specialist Dr Franz Waldner said.

"So we apply our deep neural network and algorithms to produce the paddock boundaries based on vegetation signatures and land features.

"Our method only needs one satellite image taken at any point in the growing season to distinguish the boundaries. It relies on data driven processes and decisions rather than assumptions about what's on the ground."

ePaddocks is available from CSIRO's Ag Climate Data Shop, which the organisation touts as the place to go to access ag-related weather and climate data for Australia.  

In addition to ePaddocks, the shop also includes Your Seasonal Forecast, which takes output from the Bureau of Meteorology's (BoM) latest season climate model and transforms it into an application-ready data feed and the Point Weather Forecast. The Point Weather Forecast is also based on BoM data and provides a seven-day weather forecast for all locations in Australia that is updated twice daily.

MORE FROM CSIRO

Editorial standards