Design and Implementation of a Corn Pest Detection Method Based on RMS-YOLOv11

Corn serves as one of China’s core staple crops, which guarantees national grain security, yet frequent pest infestations severely impact crop yields. Conventional crop protection measures suffer from low operating efficiency and inevitable environmental contamination. Meanwhile, existing mainstream detection algorithms are restricted by insufficient identification precision and high computational overhead. Targeting the above drawbacks, this research develops an improved RMS-YOLOv11 detection framework to achieve high-precision identification of corn pest individuals. The RFB component is embedded into the backbone’s feature extraction terminal of native YOLOv11 to strengthen feature extraction capacity for tiny pest targets. MobileNetV2 is adopted as the lightweight backbone to reduce the computational cost of backbone feature extraction, and the original standard Conv layer is replaced with a self-designed Conv-SWS module to balance multi-scale feature enhancement for targets with different sizes. Quantitative experimental outcomes reveal that the proposed RMS-YOLOv11 achieves 90% detection precision, 83% recall, 89.1% mAP@50, 65.5% mAP@75 and 59.3% mAP@50–95. These five indicators show absolute percentage-point improvements of 3.1, 4.0, 4.5, 2.4 and 3.6 compared with vanilla YOLOv11, and the proposed framework achieves better overall detection metrics against Faster-RCNN, SSD and other mainstream YOLO variants. Grad-CAM thermal visualization results verify that the optimized network can precisely lock the actual pest area and remedy the original network’s deficiency of inadequate feature attention toward small-size targets. The experimental results demonstrate that the proposed model achieves a competitive detection performance on the corn pest dataset. RMS-YOLOv11 can be regarded as a promising image-level corn pest detection model. Nevertheless, further validations based on independent field datasets, long-term trap monitoring data, multi-camera equipment, multi-regional data and real pest management decision scenarios are still required before practical field application.