AUTHOR=Liu Jun , Wang Xuewei , Miao Wenqing , Liu Guoxu 

TITLE=Tomato Pest Recognition Algorithm Based on Improved YOLOv4

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.814681

DOI=10.3389/fpls.2022.814681

ISSN=1664-462X

ABSTRACT=Abstract.  Tomato plants are infected by diseases and insect pests in the growth process, which will lead to the reduction of to-mato production and economic benefits of growers. At present, the detection of tomato pests is mainly through manual collection and classification of field samples by professionals. This manual classification method is expensive and time-consuming. The existing automatic pests detection based on computer requires the background environment of the pests is simple, and cannot locate pests. To solve these problems, based on the idea of deep learning, a tomato pests identification algorithm based on im-proved YOLOv4 fusing triplet attention mechanism was proposed. At the same time, a labeled dataset of tomato pests was estab-lished. The proposed algorithm was tested on the established dataset, and the average recognition accuracy could reach 95.2%. The experimental results show that the proposed method can effectively improve the accuracy of agricultural pests classification judgment and obtain the accurate positioning of agricultural pests, which is superior to the previous automatic agricultural pests detection methods.
Keywords: image processing; Pests identification; YOLO; Object detection