[PhD’s Corner] Léo Pichon: Methods for assessing quality of data coming from high spatial and temporal resolution observations :  the case of decision making in viticulture

 Léo Pichon is one of the #DigitAg co-funded PhDs

 

 

Methods for assessing quality of data coming from high spatial and temporal resolution observations :  the case of decision making in viticulture

  • Start Date: November 2017
  • University: MUSE Montpellier University of Excellence / Montpellier SupAgro
  • PhD School: GAIA, Montpellier
  • Field(s): ICT for Agriculture, Agronomie
  • Doctoral Thesis Advisor: Véronique Bellon-Maurel (Irstea Dept Ecotechnologies)
  • Co-supervisors : Bruno Tisseyre (Montpellier SupAgro ITAP), James Taylor (Newcastle University, UK / Irstea ITAP)
  • Funding: Montpellier SupAgro
  • #DigitAg: Labelled PhD – Axis 3, 4 & 6 – Challenge 1 & 6

 

Keywords: Precision Viticulture, Decision making, Data quality, Crowdsourcing, High resolution data

Abstract:

Development of new technologies (UAV, Sentinel satellite, smartphones, etc.) has fostered the rise of new sources of information in agriculture. These technologies were mostly designed for other business sectors but still represent potential opportunities for being a support to expertise and decision making in agriculture. However, the data coming from these new sources of information have varied characteristics in nature, quality, and spatial, temporal or spectral resolution and extent. These data characteristics are not necessarily suited to the use as a support for expertise and decision-making. In that respect, evaluating the interest of a new source of observation considering its specific characteristics is a strong operational issue raising methodological scientific questions. This issue should be reinforced in the coming years with the emergence of new technologies with their own characteristics (hyperspectral imagery, nano-satellites, Terahertz, etc.).

In precision agriculture, many papers study the characteristics of measurement systems and optimal acquisition conditions for a dedicated and specific application (characterization of the variability of yield, vigor, water state, etc.). However, to our knowledge, there is no existing method to evaluate the interest of a new source of observation with given characteristics. The thesis will answer this challenge by proposing methods allowing i) to define the interest and the optimal mode of representation of a new source of spatial observation, ii) to evaluate the quality of a spatial data used as a decision support (iii) to improve the quality of spatial data in the particular case of a new source of observations:  crowdsourcing.

 

Contact:  leo.pichon [AT] supagro.fr

Networks: ResearchGate  LinkedIn