[Defended thesis] Baptiste Oger: Adaptative Constrained Optimization for Spatial Sampling in Precision Agriculture

Adaptative Constrained Optimization for Spatial Sampling in Precision Agriculture

My name is Baptiste Oger and I'm an agricultural engineer specializing in applied statistics. I have a double master's/engineer's degree and dual skills in agronomy/statistics, as I chose to specialize in statistics in my final year, via a Montpellier SupAgro course with AgroCampus Ouest and the University of Rennes 2. This interdisciplinary thesis matches my profile in agronomy, statistics and computer science. The computer science part allows me to continue my training.

My subject is yield estimation in viticulture. This estimation, based on measurements taken by the winegrower on his plot, takes place a few days before the harvest. Knowing the quantities that are going to be harvested enables them to better manage the quality of the product (respecting quotas) as well as to calibrate labor and equipment (number of vats, etc.) according to actual needs, in order to minimize operating costs. In concrete terms, I propose to develop a method that tells the winegrower where to take these measurements in order to obtain the best possible yield estimate, while respecting his constraints (time, travel).

• Starting date: October 2017
• University: MUSE Montpellier Université d’Excellence / Institut Agro
• PhD school:  GAIA Montpellier
• Scientific field:  IT, precision agriculture
• Thesis management:  Bruno Tisseyre (Institut Agro, ITAP)
• Thesis supervisors:  Philippe Vismara (Institut Agro, MISTEA)
• Funding: #DigitAg – Institut Agro
•  #DigitAg : Cofunded PhD – Axe 3 (Capteurs et acquisition et gestion de données)

Keywords: Precision agriculture, yield optimization, spatial sampling

Abstract: The wine industry needs to know the yield of each vine field precisely to optimize quality management and limit the costs of harvest operations. Yield estimation is usually based on random vine sampling. The resulting estimations are often not precise enough because of the high variability within vineyard fields. Recent works showed the interest of using vegetative index (i.e. NDVI, GLCV, etc.) derived from high spatial resolution airborne/satellite images to optimize sampling. These works showed it was possible to improve yield estimation by 15 % depending on the considered vine field and the strength of the correlation between vegetative index and yield components. Other recent research has proposed a unique original approach, based on the consideration of spatial and operational constraints, to optimize the operation of within-field machine operation in viticulture based on high spatial resolution information derived from airborne images and experts zoning (application to optimize selective harvest). The originality of the PhD project is to propose an interdisciplinary approach that takes into account both these results to optimize the spatial sampling carried out by an operator. The work will aim at developing a methodology which considers i) high resolution information describing the within-field spatial variability ii) operational as well as spatial constraints (time required to perform an observation, time required to walk from a site of measurement to another, spatial organization of the cultivation like rows etc.) and iii) specificity of the field under consideration (like spatial organization of the variability as well as the strength of the correlation between high resolution data and the agronomic information under study, etc.). The research will provide with the wine industry an adaptative Constrained Optimization for Spatial Sampling in Precision Agriculture. In the short term, the methods may be embedded in a mobile platform like a smartphone with localization facilities. In the long term, these results will be quite usable to optimize the sampling carried out by mobile platforms such as robots.

Contact : baptiste.oger@supagro.fr

Social network: LinkedIn

Communications & Papers

Download the thesis manuscript

Baptiste Oger, 30 novembre 2017 : Présentation du sujet à Stéphane Tavert, Ministre de l’Agriculture et de l’Alimentation

Baptiste Oger, #DigitAg PhD introduces his studies to @StTRAVERT : stats et agriculture de précision pour estimer le rendement et aider le viticulteur à améliorer la gestion de sa vendange #AgTech pic.twitter.com/dMhlQKp1oU

— #DigitAg (@DigitAgLab) 30 novembre 2017

Papers in international conferences

• Oger B., Vismara P., Tisseyre B.(2019), Combining target sampling with route-optimization to optimise yield estimation in viticulture, Precision Agriculture 2019 – Papers Presented at the 12th European Conference on Precision Agriculture – https://hal.archives-ouvertes.fr/hal-02609782/

• Oger, B., Vismara, P., Tisseyre, B.(2019), Échantillonnage sous contraintes en viticulture de précision, Proc. 12th European Conference on Precision Agriculture (ECPA 2019)- https://hal-lirmm.ccsd.cnrs.fr/lirmm-01924365/file/oger_JFPC2018.pdf

• Oger, B., Laurent, C., Vismara, P., Tisseyre, B. (2021), Is the optimal strategy to decide on sampling route always the same from field to field using the same sampling method to estimate yield?, Oeno One – https://doi.org/10.20870/oeno-one.2021.55.1.3334