Nº 7366 – 1er rejeb 1446 (2-1-2025)
BULLETIN OFFICIEL
161
En outre, notre système encourage l'adoption de technologies innovantes dans le secteur agricole, contribuant ainsi à la modernisation de l'agriculture et à l'amélioration des compétences des agriculteurs. Cela ouvre la voie à une transition vers une agriculture plus durable et respectueuse de l'environnement, en encourageant l'utilisation efficace des ressources naturelles et la réduction des impacts environnementaux négatifs.
Projet ISA (Imagerie Satellite Data Mining et Applications Agricoles) : Le projet ISA bénéficie d’un financement de l’Académie Hassan II des sciences et techniques, dans le cadre d’un appel à projets. Les partenaires impliqués dans ce projet sont la fondation MAScIR (Moroccan Fondation for Advanced Science Innovation and Research-Rabat), l’ENSIAS (École Nationale Supérieure d’Informatique et d’Analyse des Systèmes- Rabat), l’IAV (Institut Agronomique et Vétérinaire- Rabat), l’HBRS (Hochschule- Bonn-Rhein-Sieg Allemagne) et les Domaines Agricoles. Références : (1) https://www.agriculture.gov.ma/fr/programme/exportation
Etude de cas 2: “Smart mobile application to recognize tomato leaf diseases using Convolutional Neural Networks”94 Introduction Despite worldwide efforts to reduce plant loss and improve food security, plant diseases still account for over 20% of crop losses. Pollution and climate change have exacerbated the severity of this problem during the last decade. Farmers now use plant disease databases and phone consultations with local pathologists instead of sending plants to diagnostic laboratories for treatment recommendations. Several papers have used standard CNN design architectures for plant disease detection, including SqueezeNet, ResNeXt, ResNet, NiN, GoogLeNet, VGGNet, ZFNet, AlexNet, and others. Several strategies and applications have been developed to prevent crops loss due to disease. Several approaches, including Convolutional Neural Networks (CNNs), have been developed for identifying plant illnesses, particularly in tomatoes. All related approaches need significant computing resources. For example, [1] presented an intelligent warning system for disease-pest detection in fruit melon, whereas [2] employed CNNs to construct a comprehensive diagnostic system for all viral infections in cucumbers. Furthermore, [3] employs AlexNet and GoogleNet assisted-CNN models to detect 14 crop species and 26 diseases. A collection of 7176 photos of healthy and healthy plant leaves was utilized. To overcome the intricacy of the analysis and make it more accessible internationally, and inspired from the MobileNet architecture, we created a smartphone application, that identifies ten distinct tomato leaf diseases using the Quantized CNN-aided model. Previous smart systems rely on deep CNNs, making them only suitable for computer usage. It is worthwhile to mention that there is no similar smart smartphone application that has been developed so far.
94 Azeddine Elhassouny, ENSIAS, Mohammed V University in Rabat, Florentin Smarandache, University of New Mexico, USA
45