Annual Cropped Area Maps
Testimonials
FEWS NET had a very good experience during our Kenya August maize crop assessment with the data you provided and we were able to rapidly validate it within the counties we visited. It formed the basis for computation of cropped acreage adjustments in Kenya. Very useful.
Gideon Galu, USGS FEWS NET
6th Grain's satellite imagery allows me to have a single source of dependable data. It’s more accurate. At the 10-meter resolution, it’s 10 times or 100 times more accurate than if you just ask a couple of farmers and make an extrapolation to the whole country. These days, with cloud processing technologies, we can produce this amount of data in good time, to produce good, reliable statistics that people can work with. So 6th Grain’s data is absolutely key for me.
Mario Kunz, Syngenta
How Our Crop Mapping Works
The 6th Grain Crop Mapping system utilizes advanced machine learning to precisely present cropped area maps. This machine learning system is underpinned by four key steps:
- Comprehensive training data is created using expert knowledge and ground information such as household surveys and photos.
- Sentinel2a/b 10m satellite data is downloaded and composited to support the labeling team and to prepare for running the classification model.
- The classification model is implemented to identify small fields and commercial farms by crop.
- We then implement the evaluation to iterate on the map, resulting in high quality 10m annual crop classification map with error estimates.
Crop persistency analysis
The system can recognize and highlight areas where local farmers do not rotate their crops, for example if they have been planting the same crop for several seasons in a row. By identifying these farmers through our crop persistency analysis tool, organizations can develop and implement targeted agriculture advice and sales strategies.
Field BOUNDARY DETECTION
Together with crop area identification, the Crop Mapping system automatically recognizes large, mid and small size fields, helping the user accurately form strategies when approaching the area. Each country has a different field segmentation size, so we change the ‘medium’ field size according to the agroecological system and country being mapped.
Additional layers of data
The Crop Mapping system functions extend well beyond visualizing areas with crops. The system allows the end user to overlay crop data with precipitation, climatology, elevation and soil data. The user can also upload locations of important agricultural infrastructure (e.g. retailers, distributors, demonstration plots, warehouses and depots, etc.) to overlay on the map for easy interpretation and analysis.
Rich visualization
Results of the system are available in a web-based user interface, which contains comprehensive data visualization features to analyze the cropped area data. The data is visualized over the map of the world, allowing the end user to zoom in and out to see details on a plain or satellite background with field boundaries visible form space. The user can switch between different crops, animate data across time, overlay it with infrastructure objects, precipitation, climatology, soil and elevation indicators, compare seasons with each other, etc.
The system also contains rich reporting functionality: the user can build a detailed report for a defined area: whether it is a selected province or a municipality of a country or even extract cropped area statistics using a drawn circle or rectangle.