Each season the Earth Observation Team of Agriculture and Agri-Food Canada (AAFC) maps crop types of individual fields across Canada. They developed an operational in-house software system to decipher crop types using satellite observations and trustworthy ground information. 

Multi-temporal optical & SAR data are collected over the agricultural growing regions of Canada. These images are combined with provincial crop insurance data (the most accurate, detailed and complete sources of information for crop types in Canada). However where insurance data cannot be used or accessed, point observation ground-truth information is collected by AAFC staff or other provincial partners via windshield surveys. Each year, AAFC staff and their supporters collect tens of thousands of points identifying crops across the country. Both of these point sources are combined and used as training and reference data for their classification system called AgriCarta. Each season the final map product is freely available to download on the government’s open data platform. 

This talk will look at how a small research project at AAFC in the early 2000’s has evolved in to an operational system that maps over 3 million km² of Canada's landcover each season, and how will continually evolve and innovate in to the future.  A wide range of topics will be discussed ranging from sensor platforms, data collection and handling, to algorithm and programming developments.