DETECTION OF VINEYARD DISEASES USING THE INTERNET OF THINGS TECHNOLOGY AND MACHINE LEARNING ALGORITHMS

• Published in: Aerul si Apa. Componente ale Mediului pp. 128 - 139
• Main Authors: Roşcăneanu, Roxana, Streche, Robert, Osiac, Filip, Bălăceanu, Cristina, Suciu, George, Drăgulinescu, Ana Maria, Marcu, Ioana
• Format: Journal Article
• Language: English
• Published: Cluj-Napoca Babes Bolyai University Faculty of Geography 01.01.2022
• Subjects: Airborne microorganisms; Algorithms; Artificial intelligence; Crop production; Diseases; Downy mildew; Environmental conditions; Graphs; Identification methods; Image processing; Internet; Internet of Things; Learning algorithms; Leaves; Machine learning; Medical imaging; Methods; New technology; Operating costs; Parameters; Plants (botany); Powdery mildew; Production costs; Rainfall; Sensors; Streamlines; Technology; Unmanned aerial vehicles; Viticulture; Websites
• ISSN: 2067-743X; 2344-4401
• DOI: 10.24193/AWC202213

In recent years, the Internet of Things concept has rapidly spread in most fields because of the benefits it offers, motivating viticulturists to implement new technologies that increase crop production and quality, as well as streamline production costs. The study's purpose is to monitor, using Internet of Things technology, two methods of identifying vine-specific diseases, which can be determined by environmental conditions (temperature, humidity, rainfall) or by analyzing diseased leaves from the vine. The first method is associated with a field study that involves placing Internet of Things sensors inside crops to measure environmental and plant parameters, which are then sent and stored in the Cloud. Based on these parameters, a correlation is made with the values that determine the occurrence of a specific vine disease (powdery mildew, downy mildew, and grey rot). The second method involves the use of Unmanned Aerial Vehicle imaging to take images containing healthy and diseased leaves from different parts of the vine. To analyze these images, a web page has been developed integrating a machine learning algorithm that detects the leaf state from the drone image footage. After the analysis all the values are stored in a database and the results are displayed as graphs and charts that are visualized by the viticulturist so that he can take the necessary actions. This study is an important step in the implementation of Internet of Things technology in viticulture, helping to monitor the main environmental and plant parameters, as well as detecting the occurrence of diseases among the vine cultures.