Photo: BayerScrolling through the app store it seems there is an almost infinite amount of both useful and useless applications you can add to your smart phone, but helpful horticultural ones tend to be in short supply.
Despite advances in technology, effective identification apps for plants, plant diseases and/or weeds seem particularly tricky to achieve. Some plant ID apps have the user send images to an expert who usually can classify the plant, but maybe not in the timely manner one would hope.
Scientists with Bayer Digital Farming are working on perfecting a software that can recognize weeds on their own from images. Because misidentification can lead to using the wrong herbicides and insufficient weed control, Bayer is striving to keep its system highly accurate.
The software system is fed thousands of images of different plant species and then trained how to recognize a specific weed based on its color, shape, morphology and other visible aspects.
According to Hans-Joachim Santel, a weed scientist with Bayer, the photos are collected from colleagues and botanists, as well as from the mobile app, which is in beta testing right now.
“When an expert takes a photo, he/she will also observe the physical plant – outside of just the visual characteristics – to ascertain additional clues like stem shape, leaf texture, root structures, scent, latex in the stem and leaf secretions that all can aid identification,” Santel said. “We need this kind of scrutiny to ensure correct documentation and ultimately, to provide correct education to the system. This is absolutely mandatory to beat the ‘garbage in, garbage out’ principle. Then, we monitor whether the system identifies the catalog of photos correctly.”
Bayer is working on training the system to be able to identify the weeds no matter the photo quality as users take photos on all kinds of mobile devices. This way no special equipment will be required and individuals will receive a reliable identification system.
This technology is currently available through Bayer’s app called WEEDSCOUT. It is free for both iPhones and Androids. Users are encouraged to take as many pictures of weeds as possible to help improve the software system’s recognition process.
Bayer say that the current results are not very precise for the beta version.
“The objective is for the final version of the app to deliver reliable results to the users,” Bayer said.
To download the app, click here.
