project - Research and innovation

Need to identify plants or trees around you? Use your smartphone!
Need to identify plants or trees around you? Use your smartphone!

Ongoing | 2009 - 2013 France, European Union, World wide
Ongoing | 2009 - 2013 France, European Union, World wide
Derzeit wird der Seiteninhalt nach Möglichkeit in der Muttersprache angezeigt

Objectives

N/A

Objectives

Recognising wild plants is essential and useful in many fields. For example in agriculture it helps to detect invasive plants or pests and it serves conservation or recreational purposes in natural areas. But farmers, many professionals or citizens are not botanists, and such knowledge is partially available, and frequently incomplete for ecosystems that possess the highest plant diversity, i.e. the Mediterranean and tropical regions.

This is why French researchers have developed the Pl@ntNet application. It allows any user to identify plants by using their smartphone. This collaborative work between computer scientists and botanists makes botany more accessible.

Activities

The Pl@ntNet research project, initially supported by Agropolis Fondation, was developed till 2013 by a consortium of French research institutions such as INRA, CIRAD and INRIA. It continues today supported by the “Floris’Tic” initiative and has been awarded in 2016 by the prestigious French Académie des Sciences.

The application is based on a collaborative approach: users wishing to identify a plant send photos via the Pl@ntNet app. After analysing the images, the system gives the user a list of possible species, with illustrations. This approach serves to supplement the image bank daily, enabling researchers to use it to monitor plant biodiversity on a very large scale.

Additional information

Pl@ntNet App is very easy to use. After the installation, the user will be requested to indicate the database to be consulted. In principle the criterion for this choice would be the geographical area, but in case of ornamental (not wild plants) then the "Useful plants" theme is recommended. The second step would be to take and/or upload a picture of a plant, preferably focused on a single organ (e.g. leaf, flower, stem or fruit). After comparing the picture with the database, the app will come up with the result (s) of the identification. If the match is not good, the user may add more pictures in order to refine the search. Of course plant species that are well illustrated in the botanical reference database will be recognised more easily.



The number of species and the number of images used by the application is constantly growing, thanks to the users` contributions. Thus if you correctly identify a plant, you can participate in the project by submitting your observation (pictures) with the "contribution" button. After being validated, your pictures will be added to the database, helping others to idenity the same plant in the future.

The Pl@ntNet app is available for free. It has had more than 2.4 million users since its launch in 2013 and is available in French, German, English, Spanish, Italian and Portuguese. A web version is also available.

It currently covers Western Europe, the Indian Ocean, South America and North Africa, where the app can recognise up to 7000 wild plant species. Major upgrades are planned in the medium term, including the addition of more geographical areas and the extension of the initiative to exotic or ornamental plants.

Project details
Main funding source
Other public (national, regional) research funds
Project acronym
Pl@ntnet
Derzeit wird der Seiteninhalt nach Möglichkeit in der Muttersprache angezeigt

1 Practice Abstracts

NA

Derzeit wird der Seiteninhalt nach Möglichkeit in der Muttersprache angezeigt

Contacts

Project coordinator

  • Project coordinator

Project partners

  • CIRAD: French Agricultural Research Centre for International Development

    Project partner

  • INRA: French National Institute for Agricultural Research

    Project partner

  • INRIA: French National Institute for Research in Computer Science and Control

    Project partner

  • IRD: Institut de Recherche pour le Développement

    Project partner