DI-UMONS : Dépôt institutionnel de l’université de Mons

Recherche transversale
(titres de publication, de périodique et noms de colloque inclus)
2020-07-02 - Article/Dans un journal avec peer-review - Anglais - 8 page(s)

Debauche Olivier , Mahmoudi Said , Mahmoudi Sidi , El Moulat Meryem, Manneback Pierre , Lebeau Frédéric, "Edge AI-IoT Pivot Irrigation, Plant Diseases and Pests Identification" in Procedia Computer Science, 177, 40-48

  • Edition : Elsevier, Amsterdam (Netherlands)
  • Codes CREF : Sciences agronomiques (DI3600), Informatique générale (DI1162)
  • Unités de recherche UMONS : Informatique, Logiciel et Intelligence artificielle (F114)
  • Instituts UMONS : Institut de Recherche en Technologies de l’Information et Sciences de l’Informatique (InforTech), Institut NUMEDIART pour les Technologies des Arts Numériques (Numédiart)
Texte intégral :

Abstract(s) :

(Anglais) Overcoming population growth dilemma with less resources of soil and water, the irrigated agriculture allows us to increase the yield and the production of several crops in order to meet the high requirements of demands of food and fibers. Efficiently, an irrigation system should correctly evaluate the amount of water and also the timing, when applying certain irrigation doses. Global warming of the planet, to which is added in some regions an irregular regime of precipitation and a scarcity of available water resources, requires precision irrigation systems. The rational use of water and inputs (mainly fertilizers and pesticides) is crucial in some areas of the planet suffering from a deficiency of water. Hence, in these regions where the environmental conditions are harsh to ensure an efficient crop growth. Moreover, plant diseases and pests impact the yields of crops. For these reasons is it why an early detection gives us the opportunity to treat the disease or pest as quickly and effectively as possible, in order, to reduce the impact of these latter. Nowadays, the identification of plant diseases and pest with Artificial Intelligence algorithms on video flow in real conditions with variable exposition are still being a very challenging problem. Researchers classically develop algorithms that are trained on calibrated exposition images, which does not perform well in real conditions. Furthermore, the processing of a video in real time needs specialized computing resources close to the pivot-center irrigation trained with AI algorithms on real images and then analyzes rapidly, detects problem, and then react accordingly. In this paper, we complete our previous proposed IoT system to optimize the water use and we displaced the computing of data at the edge level in order to be able to process videos locally, event the Internet connection is limited. This local computing power also allows us to manage the supply of fertilizers and the treatment of plant diseases, and pests.

Identifiants :
  • DOI : 10.1016/j.procs.2020.10.009

Mots-clés :
  • (Anglais) Water Requirement
  • (Anglais) Center-pivot Irrigation
  • (Anglais) Smart Irrigation
  • (Anglais) Intelligent Irrigation
  • (Anglais) Connected Irrigation