Development of artificial vision for the detection of the Prodiplosis Longifila or Caracha pest in tomato cultivation

León León R.A.; Bravo Aranda M.B.; Castañeda Guanilo X.; Juárez Horna S.E.; Silva Gamboa H.E. · CISCI - Conferencia Iberoamericana en Sistemas, Cibernetica e Informatica (2024) · DOI: 10.54808/cisci2024.01.424

Resumen

The purpose of this article is the detection of the pest prodiplosis in tomato fruit, a problem that needs to be detected in time to avoid crop losses, improving the training of a convolutional neural network with a set of data such as photos taken in the growing area to the various plants containing the pest and other healthy plants being necessary for better detection, using a Python programming language, as it is a language with greater ease of learning, complementing the correct development and elaboration of the expected model. The neural network used is Yolo v8, one of the detection versions with greater precision and speed for use in different sectors. For the development of this article, a total of 352 photos of tomato fruit in full color were taken into account, including fruit with pests and fruit without pests, divided into 234, 59 and 59, for training, validation and testing, respectively. Finally, it is concluded that the neural network obtained the desired result with a precision of 98.32% and an accuracy of 98.01%. © CISCI - Simposium Iberoamericano en Educacion, Cibernetica e Informatica 2024.

Instituciones: Universidad Privada del Norte