Subject Time series analysis of hyperspectral images for early diagnosis of septoria symptoms in durum wheat

Recruited as a Research Officer at INRAE in Montpellier, my postdoctoral research is conducted within the UMR AGAP (Genetic Improvement and Adaptation of Mediterranean and Tropical Plants) and ITAP (Technologies and Methods for Tomorrow's Agriculture). My research project focuses on the analysis of time series of hyperspectral images for the early assessment of wheat septoria symptoms. I hold a DUT in Biology with a specialization in agronomy and a Licence’s degree in Agrosciences. I furthered my education with a Master's degree in Quality Management of Plant Productions, specializing in methods and strategies aimed at improving the quality and safety of agricultural products. Throughout my studies, I developed a keen interest in plant-pathogen interactions, particularly the mechanisms employed by the host to defend against pathogens. I then completed my thesis at CIRAD in Reunion Island. My research focused on studying various metabolomic and spectroscopic approaches with the primary objective of detecting biochemical signatures associated with pineapple fruitlet core rot disease. Using these advanced techniques, I was able to identify specific biomarkers, paving the way for new diagnostic and management methods for this fungal disease that significantly affects pineapple productions.
My postdoctoral research focuses on the study of hyperspectral image series to detect early symptoms of Septoria tritici Blotch (STB) on durum wheat, caused by the pathogenic fungus Zymoseptoria tritici. This disease leads to significant production losses due to the irreversible damage it causes to the plant foliage. Given the considerable impact of this disease on agricultural yields, it is crucial to develop non-destructive, rapid, and effective detection tools and methods. My work aims at using chemometric methods to assess early diagnosis of STB and develop phenotyping tools, thereby contributing to the protection and improvement of durum wheat crops.
Traditional methods used to manage the spread of the disease require significant economic, energy, and labor resources. The design and use of optical sensors coupled with chemometric tools offer a promising alternative to improve agricultural practices. These technologies allow for faster, larger-scale, and more cost-effective action against STB in durum wheat. By integrating these advancements, it would be possible to detect the disease symptoms early, optimizing interventions and reducing the economic and environmental impacts of traditional methods. Additionally, these technologies would accelerate the selection of the best durum wheat phenotypes by favoring the most resistant genotypes to STB.

• Starting date:  1st March 2024
• Scientific field: Chemometrics - Spatio-temporal kinetics
• Post-doc supervisors: Martin Ecarnot, UMR Agap, Inrae, Jean-Michel Roger and Ryad Bendoula, UMR Itap, Inrae
• #DigitAg:  Axe 3 : Capteurs et acquisition et gestion de données, Challenge 2 : Le phénotypage rapide, Challenge 3 : La protection des cultures

Keywords : Hyperspectral imaging, chemometrics, phytopahtology, phenotyping

Abstract: The development and implementation of new technologies among various agricultural stakeholders aim to promote innovation and establish reliable, fast, simple, and cost-effective methods. In my postdoctoral research, I will use visible-near infrared hyperspectral imaging and various chemometric methods to identify markers of wheat STB disease. The spatial and spectral dimensions will be analyzed daily to reveal the specific signatures of the disease over time. Classification and prediction methods will be developed, applied, and compared on the dataset to diagnose the early onset of STB symptoms on durum wheat leaves. Additionally, variable selection methods will be studied to identify the wavelengths of interest associated with the disease, potentially leading to the development of simplified sensors.

Contact: lorraine.latchoumane [AT] inrae.fr
Social network: ResearchGate