Scouting your crop for weeds, disease incidence or leaf damage has just become easier with the launch of a new, high-tech mobile identification tool from Bayer.
It’s called xarvio™ SCOUTING, and it’s a free phone or tablet-based app from Bayer’s Digital Farming Solutions group that builds on the technology of previous generation scouting apps with a simple-touse interface and much larger database of reference points.
“In the past, most weed identification tools were catalogue based,” says Warren Bills, business development with Bayer’s Digital Farming Solutions group. “You would take a picture of the weed and compare it to a database of pictures to see what matched. Our new xarvio SCOUTING app takes the photo and quickly identifies it using a deep learning algorithm comparing it to an extensive image database. It can also now identify disease and quantify leaf damage.”
Two years ago, Bayer’s global Digital Farming Solutions group developed its WeedScout app to provide a quicker, easier way to identify common weeds. It used the type of leading edge image capture technology (currently being used for facial recognition in other industries) and applied it to weeds. Users were encouraged to take as many photos of weeds as possible in order to populate the app with a wide variety of images.
“Once we started collecting a volume of several species of weeds at all different growth stages, the technology evolved to offer even more precise identification,” says Bills. “Each picture is analyzed by the algorithm and we now have thousands of reference points to help identify weeds. The algorithm intelligence in this software improves when more information is uploaded, providing greater accuracy in identifying weeds.”
Last year, Bayer Canada contracted with universities to start capturing as many images of various weeds as they could to help improve the algorithm. It also worked with different research groups for in-field weed collection. Experts validate the images to make sure the algorithm is identifying them correctly. While these groups are helping improve the user experience, the key audiences for this technology are expected to be growers, agronomists and retail outlets.
There are different versions of the app available for global users, and the latest version of the Canadian app can now reliably identify approximately 10 weed species.
Bills says it’s an accelerated start to the priority weed species that the app will eventually be able to accurately distinguish from a single photo.
In addition to weeds, the new xarvio SCOUTING app will also be able to identify disease and quantify leaf damage. Because most crop disease is ideally treated preventatively, some of the disease information will be most useful for identifying symptoms or for making disease management decisions the following year.
“This app will supplement other scouting tools and software, rather than replace them entirely,” says Bills. “The xarvio SCOUTING app is not an extensive record keeping software and, in its current version, does not recommend a crop protection treatment for the weed, disease or pest,” he explains. “We expect the app will evolve to offer more in the future. In some cases, such as with disease, we hope eventually the app will be able to send alerts, show pest intensity heat maps or predict pest movement across geographies.”
Bills says that the aggregate data collected by users will help Bayer determine the direction of the app in the future. “Crowd sourcing info will help make this tool more valuable and will also help us determine where we need to take it into the future,” he says. “Right now xarvio SCOUTING is easy to use, free and will help extend our knowledge of weeds, disease and insect pressure across Canada.”
Bayer’s xarvio SCOUTING app is now available on the Apple App Store or on Google Play.