What progress has been made in research on precision agriculture? A review of ECPA 2019

With the evolution of new technologies, precision agriculture has continued to develop since its emergence in the 1980s. It supports farmers in their decision-making and helps to address environmental challenges such as climate change and resource depletion. In July this year, the European Conference on Precision Agriculture, ECPA 2019, was held in Montpellier, bringing together 380 scientists from 37 different countries. We review the key scientific elements of the event with its organiser, Bruno Tisseyre (Montpellier SupAgro, UMR ITAP).

From its inception in the 1980s, how is precision agriculture defined in 2019?

[Bruno Tisseyre]: Opinions vary depending on the field of application and the technological developments considered, and the universally accepted definition of precision agriculture is thus very recent. It was published during ECPA 2019 by the International Society of Precision Agriculture (ISPA) after consultation with 45 scientists from all over the world who were asked to discuss it and to reach a consensus. The definition adopted by ISPA places the emphasis on the use of data and expertise to address agricultural, economic and environmental challenges, rather than on technologies:

Precision Agriculture is a management strategy that gathers, processes and analyses temporal, spatial and individual data and combines it with other information to guide site, plant or animal specific management decisions to improve resource efficiency, productivity, quality, profitability and sustainability of agricultural production.

Are there areas of the conference where progress is more obvious?

[B.T.]: Yes! There are five sectors in which the number of papers is particularly high.

First, temporal aspects (real-time or near real-time) are a noteworthy research front. The proportion of articles received that deal with temporal as well as spatial aspects is a real development. For remote sensing, time series are represented in numerous sessions of the conference, in particular for irrigation, but also for nitrogen management and many other sectors. These are essentially time series of images. The arrival of the European Sentinel-2 satellites and their comprehensive, updated data is a game changer. Obtaining low cost time series of images is unquestionably an important factor in these developments. There are also several communications on the dynamic assessment of plant parameters to integrate them into development models, crop models, etc., as well as on the assessment of plant characteristics, in particular indices to improve evapotranspiration models for irrigation management for example, or the assessment of plant water requirements. This is all made possible by the availability of time series of images. Time series of yield data are also present. For example, the (#DigitAg labelled) thesis work by Cécile Laurent raises methodological questions: how can knowledge be extracted from these series?

The second new field is on-farm experimentation, with a whole session devoted to it this year. We are just starting out, with feedback communications and case studies. In particular, the social and economic aspects were addressed: how can we associate farmers and advisers? How will they organise themselves? Few methodological aspects were examined. For example, managing the uncertainty of data was not discussed. Questions will emerge, as it is still early days. A satellite workshop at the conference was also dedicated to “On-farm experimentation”. The COFE international consortium on experience sharing with farmers is led by Simon Cook, Professor at Curtin University in Australia.

The third point is the emergence of artificial intelligence and deep learning in the majority of sessions, including the traditional sessions at a conference such as ECPA, namely the management of water, weeds and nitrogen. We held a special session on image processing using AI, for example in viticulture for the detection of grapes, or the estimation of parameters using drone images.

The fourth important element is the growth of emerging issues linked to localisation (GNSS positioning system) and robotics. We are entering a new phase, another level. The Chinese-British communication on selective tomato harvesting is one example.

The final original aspect is that we attended no fewer than seven presentations on the adoption of technologies and methodologies in precision agriculture, spread over two sessions, one of which was dedicated to its economic issues. Nina Lachia thus presented a summary of studies by the Digital Agriculture Observatory (AgroTIC – #DigitAg).

Finally, we note that there are fewer communications on drones. Why is this? Less social need? Variable regulations depending on the country? “Competition” from the availability of Sentinel-2 data? We can put forward several theories.

Inspirational speakers

The presentations by the four keynote speakers were particularly appreciated by all, because of the range of subjects addressed and the quality of the speakers. Jean-Baptiste Féret examined the post-NDVI process[A1] ; Benjamin Addum took a pragmatic look at successes and failures in African contexts and environments, which can provide lessons here; Sofía Cilla gave us a number of keys and resources for making better use of European services (EGNOS/GNSS) for the correction of GPS signals; and, finally, Alex Escolà took stock of 3D measurement in agriculture with a state of the art.

The next ECPA meeting is in Hungary. Gábor Milics of Széchenyi István University will be taking over for Budapest 2021.

 

Contact : bruno.tisseyre [AT] supagro.fr

#DigitAg Publications in Precision Agriculture ’19 Conference proceedings

  1. Production gap analysis – an operational approach to yield gap analysis using historical high-resolution yield data sets
    C. Leroux, J. Taylor, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_8
  2. Investigating the harmonization of highly noisy heterogeneous datasets hand-collected over the same study domain
    L. Pichon, C. Leroux, V. Geraudie, J. Taylor, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_91
  3. A collective framework to assess the adoption of precision agriculture in France: description and preliminary results after two years
    N. Lachia, L. Pichon, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_105 
  4. A multispectral processing chain for chlorophyll content assessment in banana fields by UAV imagery
    G. Rabatel, J. Lamour, D. Moura, O. Naud – https://doi.org/10.3920/978-90-8686-888-9
  5. Evaluation of a functional Bayesian method to analyse time series data in precision viticulture
    C. Laurent, M. Baragatti, J. Taylor, T. Scholasch  A. Metay, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_7
  6. An iterative region growing algorithm to generate fuzzy management zones within fields
    C. Leroux, H. Jones, B. Tisseyre
  7. Combining target sampling with route-optimization to optimise yield estimation in viticulture
    B. Oger, P. Vismara, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_20
  8. Disentangling the sources of chlorophyll-content variability in banana fields
    J. Lamour, C. Leroux, G. Le Moguédec, O. Naud, M. Léchaudel, B. Tisseyre – https://doi.org/10.3920/978-90-8686-888-9_37
  9. Investigation on LiDAR-based indicators for predicting agrochemical deposition within a vine field
    A. Cheraiet, M. Carra, A. Lienard, S. Codis, A. Vergès, X. Delpuech, O. Naud – https://doi.org/10.3920/978-90-8686-888-9_18

TALK: Jean-Baptiste Féret, Linking Remote Sensing with physical modelling of vegetation to characterize vegetation status and development: methods, challenges and perspectives