Read the current Progressive Forage digital edition

Tools for weed identification

Lynn Marie Sosnoskie for Progressive Forage Published on 30 October 2019
A screenshot of the Pl@ntNet app

The U.S. loves smartphones; the statistics portal estimates 81% of all U.S. adults will own at least one device in 2019.

Mobile applications (apps) account for over 90% of users’ smartphone internet time, with most of it focused on socializing, gaming and other forms of entertainment. But did you know there are applications which have been purposely developed to meet agricultural and horticultural needs, including tools to assist with weed identification?

And why, exactly, is weed identification so important? Simply stated, because the weed community present in any given field, pasture, orchard, vineyard or rangeland system is likely to be comprised of species which are differentially sensitive to a variety of herbicide modes of action, as well as physical and cultural control practices. Consequently, to develop the most effective weed-control program, managers must be aware of the species present and how susceptible they are to control.

People have asked if there are any available apps which can scan pictures of weeds and tell you exactly what you are looking at. As it turns out, yes, there are. One popular tool is the Pl@ntNet app, which allows users to upload one or more pictures of leaves, flowers, fruits or bark and then compares the photos to a sizable image bank using visual recognition software. The app then returns a list of possible matches and provides the user with links to the Wikipedia entries for each of the species in case more information is desired.

Another app is iNaturalist. Like Pl@ntNet, the iNaturalist app compares your image against those in a database and provides users with a list of suggestions, links to information about the species and maps of occurrences worldwide. Both Pl@ntNet and iNaturalist allow you to share your images with their communities. Important note: Some apps may not be available for all smartphone operating systems.

The ability to achieve a successful identification is highly dependent on the quality of the pictures submitted. Images should be focused on individual plants or plant parts. Care should be taken to minimize the number of background features which could affect the analysis and reduce your chances of attaining a proper ID. Images should be taken at a distance close enough to highlight a morphological feature, but not so close to make a structure indistinguishable. It is not reasonable to expect any app to be 100% effective; however, these programs can be very useful when trying to narrow the spectrum of possible weeds. It is always good practice to double-check if you are not certain about an identification. Good physical guidebooks include:

  • Weeds of the Northeast (Uva et al., 1997. ISBN-13: 978-0801483349)

  • Weeds of the South (Bryson et al., 2009. ISBN-13: 978-0820330464)

  • Weeds of the Midwestern United States and Central Canada (Bryson et al., 2010. ISBN-13: 978-0820335063)

  • Weeds of California and Other Western States (DiTomaso, 2007. ISBN-13: 978-1879906693)

Local extension advisors can also help with the identification of troublesome species. Identifications can be made either from live specimens or from digital images. If submitting a fresh sample, collect several plants representative of the infestation. Collect as much of each individual plant as possible (including roots, flowers and seed pods).

Place the samples in a sealed Ziploc bag with a few damp paper towels and deliver (or ship) your specimens to your county office (the faster, the better, as fresh samples are easier to identify). Be sure to include your name and contact information, the location of where the weeds were collected (GPS coordinates, address and description of the habitat, i.e., roadside, orchard, alfalfa field, etc.), as well as any other relevant information.

If emailing digital images, be sure to capture pictures of the plant’s leaves, stems, flowers, roots, and seedpods, as well as a photo of the specimen in its environment. Try to remove excess vegetation from around the sample so its growth habit is apparent. Take photos of any unique structures (such as spines) which could be a diagnostic tool. Include the same contact information (name, address, etc.) as listed above in your message.  end mark

PHOTO: A screenshot of the Pl@ntNet app shows kochia weed and gives a brief description of it. Getty Image

Lynn Marie Sosnoskie is an assistant professor at Cornell University. Email Lynn Marie Sosnoskie.