GreenCropTracker: Deriving crop information from digital photography
This article was originally written and submitted as part of a Canada 150 Project, the Innovation Storybook, to crowdsource stories of Canadian innovation with partners across Canada. The content has since been migrated to Ingenium’s Channel, a digital hub featuring curated content related to science, technology and innovation.
Agriculture and Agri-Food Canada scientists have developed a tool to help create better models for simulating crop growth and obtaining physical descriptors of agricultural crops. GreenCropTracker is a software tool that can be used to derive crop information from a simple digital photo taken over a crop canopy in agricultural fields. The resulting model predictions help growers understand what happens in the field so they can manage production more efficiently.
Drs. Elizabeth Pattey and Jiangui Liu at the Ottawa Research and Development Centre thought that a time series of photos taken over a crop canopy for documenting the growth advancement would contain quantitative information that could be evaluated. They developed a histogram-based threshold method to differentiate gaps from plant tissues and then incorporated this image analysis method into GreenCropTracker.
GreenCropTracker can be used to study crop production in experimental plots and fields, to evaluate the impact of climate variation on crops and to provide ground data for deriving crop descriptors from remote sensing data. It also verifies the output of crop growth models and integrates surface conditions with regional meteorological models. The software, which has already been tested over corn, wheat, and soybean crops, is used by the Food and Agriculture Organization of the United Nations and has been requested by nearly 50 countries because it provides fast and inexpensive crop information.
Software distribution site: http://www.flintbox.com/public/project/5470/