Case study Identity Card

Long title: Pl@ntNet platform, a citizen science initiative dedicated to plant identification and biodiversity monitoring.
Geographical coverage of the case: Worldwide ambition (South America, Europe, Africa, Asia). Up to date, concrete realizations in Europe and Indian Ocean
Case study team: Pierre Bonnet (CIRAD, Amap), Timothée Morin (Cerdi, intern), Agathe Devaux-Spatarakis (CIRAD, innovation), Benoit Bertrand (CIRAD, Ipme), A. Joly (Inria, Zenith), S. Dufour-Kowalski (Inra, Amap), J. Carré (Tela Botanica), J.-f. Molino (Ird, Amap), N. Boujemaa (Inria Saclay), D. Barthélémy (CIRAD, Bios),
Years covered: 2003-…

Case study description

The innovation under study: Pl@ntNet, botany in the digital age

Describing, identifying and naming plants is no simple matter for amateurs hoping to venture into the challenging, complex world of botany. To help them, researchers at CIRAD and their colleagues have spent several years developing digital tools to make this discipline accessible to all. Their project is particularly innovative: plant identification is based on a system that automatically analyses photos of plants. With almost four million users in 150 different countries since its launch in 2013, their mobile application, Pl@ntNet, is a success. It is also driving the science forward, since its users can add photos to the database. The story is just beginning.

Pl@ntNet was born from a meeting between a digital science researcher and a botanist. Their idea was to make botany accessible to all, thanks to digital tools and information sharing. This meant consolidating the traditional approach, which could potentially discourage amateurs, and developing a visual identification process based on easily observable criteria.

Their project slowly took shape between 2007 and 2008, and was officially launched in 2009. Financed by Agropolis Fondation, it involves research organisations – CIRAD, IRD, INRA and INRIA – and the Tela Botanica network of French-speaking botanists. It is based on expertise, methodologies and databases developed over the last 15 years by these different teams. In 2013, it culminated in the Pl@ntNet mobile application.

Making botany accessible to all

The principle is simple: users take a photo with their smartphone of part of the plant they wish to identify (leaf, flower, fruit, stalk or bark) and upload it to the Pl@ntNet application. The system then compares it to the 500 000 or more images in the database and suggests names of the most visually similar species. Users can then identify the species from the results found.

Citizen science to achieve a scale change

But the application does not stop at that. It is also a tool for citizen science, as users can add photos of plants they have identified with certainty to the database. This new data is then collaboratively checked before being included in the database. More than 800 000 observations are currently being analysed.

Almost four million users

Pl@ntNet is used by all sorts of people, from amateurs to senior researchers, in around 150 countries. In total, almost four million people have adopted this new means of accessing knowledge about botany and have contributed to its development.

Flora from France and elsewhere

Today, the application covers almost 6 500 wild species in France, in other words more than 80% of all flora, based on several hundred thousand images. It also extends to North Africa, the Eastern Mediterranean, the Indian Ocean (Reunion Island and Mauritius), North America, the French West Indies, French Guiana, the Tropical Andes and the Hawaiian Archipelago.

With its translation into Spanish, Portuguese, Italian, English, German, Dutch, Czech, Slovak, Arabic and Chinese, the application is expanding internationally and could eventually include new flora from South America, tropical Africa and Asia. It is also widely used in Brazil and Australia, despite the fact that the species found there differ greatly from those currently listed.

Defining expectations and developing new uses

This research platform is relatively recent, and it is therefore too soon to assess all of its impacts. But it is already possible to identify their first effects and anticipate their future benefits.

The researchers have thus posted an online questionnaire to identify the fields in which the application is used and the changes it has prompted for users. They have also interviewed people connected to the project in different ways to seek their views on its current and potential impacts. These impacts have been highlighted in two key fields. The first concerns the management of natural or cultivated ecosystems, contributing in particular to enhancing existing species distribution maps. The second relates to education, with the development of new learning mechanisms: current partnerships with environmental education associations clearly illustrate the potential impacts identified through this questionnaire.

The researchers have also incidentally been able to identify users’ expectations and to gather specific information to steer the evolution of the application towards new geographical or thematic contexts.

Changing practices and democratising science

The Pl@ntNet application results in practice changes for all of its users, especially scientists, who now, for the first time, have access to a shared platform for exchanging and sharing their data.

It is also altering the general public’s perception of its natural environment. With the botanical skills and knowledge acquired, users can take a fresh look at their surroundings and get to know them in a fast, fun way. This fun side to the application may also encourage teachers to use it with their pupils and to change their teaching methods.

A variety of potential products

Among the potential products derived from this application, the survey highlights botanical guides, dedicated to a geographical area, such as a natural park, a national forest or a specific ecosystem. One possibility is that Pl@ntNet could be used as an autonomous, offline application, with the option of downloading a database according to the user’s geographical location.

Another anticipated innovation is the extension of this visual recognition system to other entities or organisms, such as plant diseases or insects.

Climate and agriculture: two fields of application to be explored

Other changes have been identified that could lead to a range of impacts. The development of real-time mapping of plants, with the study of their phenology (the influence of climate variations on plant flowering, fruiting, etc.), is one of these. It would make it possible to study how species evolve over the year, taking into account all geolocated botanical data and adding a temporal dimension to this data, with the added possibility of long-term monitoring of the effects of climate change on vegetation.

Pl@ntNet could also have numerous applications in agriculture: by adapting it to machines, such as drones or autonomous robots, to identify weeds at an early stage, for example.

Credibility and capacity building

Pl@ntNet has received considerable media attention, with the result that this broad dissemination has made the project partners more visible and more credible in the eyes of the general public. Moreover, it has helped to increase the popularity of biodiversity informatics: an important impact for a relatively recent, little-known discipline.

The beginning of a long story

This information will all help to shape the evolution of the system and to establish an appropriate monitoring process. New ways of questioning users are being developed, aimed at better understanding how they use Pl@ntNet and more effectively meeting their expectations.

Pl@ntNet is continuing to develop, with better error detection, scientific assessment of its visual search engine, management of bias and more in-depth exploration of big data. Eventually, the goal is to develop a participatory botanical observatory, by digitising and collating as much botanical data as possible.

Stakeholders

Stakeholders involved in the innovation process

Research stakeholders : CIRAD, Inria, Inra, Ird
Civil society stakeholders : Tela Botanica members
Institutional stakeholders :  Agropolis Fondation

Stakeholders impacted by the innovation process

Research stakeholders : LifeCLEF consortium, International computer scientists groups using pl@ntnet data.
Civil society stakeholders : Naturalists, Foresters, Educators, Students, Farmers, Guides, Hickers.
Institutional stakeholders :  Agropolis Fondation