SCPest-YOLO: A Lightweight Network for Cotton Pest Detection Based on YOLOv11

Cotton pest detection is very important to ensure cotton yield. In the actual field pest detection, the pests are small and many. How to realize the lightweight model under the premise of ensuring the detection accuracy has become a key bottleneck in practical applications. To solve this problem, a lightweight detection network SCPest-YOLO was proposed. Firstly, the lightweight StarNet was used to replace the Backbone network in YOLOv11, which reduced the number of model parameters while ensuring the effect of pest detection. Secondly, the improved CFBE module is used in the Neck part to achieve more efficient feature extraction. Finally, it was verified on Insect9, the self-built data set and the public data set. Compared with the current mainstream detection models, SCPest-YOLO achieves an effective balance between detection accuracy and model complexity, showing better performance.