Development of an artificial vision algorithm based on neural networks for the identification of oidium in the blueberry plant

Abstract

In this study, we introduce a revolutionary approach that combines computer vision and neural networks in order to design a computer vision algorithm that can distinguish between blueberry leaves that have powdery mildew and those that are healthy without this disease. which could significantly improve the quality and efficiency of blueberry production in the agroindustrial industry. Methodology. The Python programming language and the YOLOv7 library were used, it was executed on an Intel Corei5 processor, 8 GB of RAM, Nvidia Geforce 1650 graphics card, webcam with 1080p resolution. Using an experimental and quantitative approach, we evaluate the effectiveness of the algorithm. Results. An efficiency of 91% was demonstrated in the detection of the Oidio or Oidiosis fungus in blueberries, validating the effectiveness of the method and highlighting its potential to transform the agro-industrial industry. Conclusion. It will allow for more efficient classification of products, which could have a positive impact on the quality and competitiveness of the blueberry market.