Automated Detection and Identification of Olive Fruit Fly Using YOLOv7 Algorithm

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2023

Autores

Victoriano, M (Autor) Outra; Oliveira, L (Autor) Outra; Oliveira, HP (Autor) FCUP

Ata de Conferência Internacional

Título: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Páginas: 211-222

11th Iberian Conference on Pattern Recognition and Image Analysis, IbPRIA 2023

Alicante, 27 June 2023 through 30 June 2023

Indexação

Scopus - 6 Citações

Outras Informações

ID Authenticus: P-00Y-JZY

DOI: 10.1007/978-3-031-36616-1_17

Abstract (EN): The impact of climate change on global temperature and precipitation patterns can lead to an increase in extreme environmental events. These events can create favourable conditions for the spread of plant pests and diseases, leading to significant production losses in agriculture. To mitigate these losses, early detection of pests is crucial in order to implement effective and safe control management strategies, to protect the crops, public health and the environment. Our work focuses on the development of a computer vision framework to detect and classify the olive fruit fly, also known as Bactrocera oleae, from images, which is a serious concern to the EU¿s olive tree industry. The images of the olive fruit fly were obtained from traps placed throughout olive orchards located in Greece. The approach entails augmenting the dataset and fine-tuning the YOLOv7 model to improve the model performance, in identifying and classifying olive fruit flies. A Portuguese dataset was also used to further perform detection. To assess the model, a set of metrics were calculated, and the experimental results indicated that the model can precisely identify the positive class, which is the olive fruit fly.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 11