YOLO-Gum: a lightweight target detection model for gummosis on tree branches in smart agriculture

Gummosis, a common disease among stone fruits such as peach, plum, and apricot trees, primarily affects the trunks and major branches. Peach trees stand as one of the frequently targeted species for this disease. To address the issues of the inability to observe high branch and trunk lesions directly, as well as the complex morphological features and low differentiation, a lightweight detection model, YOLO-Gum, has been proposed. The objective of this model is to provide an accurate basis for the prevention and scientific management of peach gummosis. Firstly, the SENetV2 module was integrated into the original YOLOv8 backbone network, replacing some of the original convolutional layers to enhance the model’s representative capability. Secondly, the CCFM structure was introduced into the neck structure to integrate detailed features and contextual information, reducing the number of parameters and improving computational efficiency. The fusion of CCFM and SENetV2 structures maintains the lightweight nature of the model while optimizing feature extraction to enhance detection accuracy. The experimental results show that the improved YOLOv8n model attains a precision of 92.5% and an F1 score of 74.3%. Compared to the original YOLOv8n model, there are improvements of 5.3% and 6.2% respectively. Furthermore, the parameters of the improved model are 2.79 M, the model size is 5.57 MB, and the FLOPs are 7.6 G, which are reduced by 12.54%, 35.4%, and 12.64%, respectively, in comparison with the original YOLOv8n model. Therefore, this model, being lightweight, precise, and robust, offers technical support for peach tree growth management and robotic vision systems for disease detection.