Authors

Abstract

Accurate and efficient rice-pest detection is essential for field scouting, early warning and precision control. This study presents a lightweight YOLOv11n-based framework for accurate edge deployment. It integrates a Multi-cognitive Visual Adapter (Mona), a Detail-Preserving Contextual Fusion module (DPCF) and Wise-IoU-based non-maximum suppression (Wise-IoU-NMS). Mona improves multi-scale feature representation with limited overhead; DPCF preserves fine-grained pest details during contextual fusion; and Wise-IoU-NMS stabilizes localization of overlapping and densely distributed targets. On a public dataset of 5,212 images from six near-balanced rice-pest classes, divided 8:1:1 for training, validation and testing, the proposed model achieved 95.4% mAP@50, 94.8% precision and 94.7% recall with 2.65 M parameters and 6.5 GFLOPs. The F1-confidence curve identified 0.476 as the alarm threshold. Deployment on Jetson Nano using ONNX and frame-level logging demonstrated a practical workflow for real-time monitoring and selective spraying.

Abstract in English

水稻害虫的快速、准确检测对于田间巡检、虫情预警和精准防控具有重要意义。本文提出一种面向边缘端高精度部署的轻量化 YOLOv11n 检测框架，融合多认知视觉适配器（Multi-cognitive Visual Adapter, Mona）、细节保留型上下文融合模块（Detail-Preserving Contextual Fusion, DPCF）和基于 Wise-IoU 的非极大值抑制策略（Wise-IoU-NMS）。Mona 以较低开销增强多尺度特征表达，DPCF 在上下文融合中保留细粒度虫体信息，Wise-IoU-NMS 则提高重叠及密集目标的定位稳定性。在包含 5,212 张图像、6 类近均衡水稻害虫并按 8:1:1 划分训练集、验证集和测试集的公开数据集上，所提出模型取得 95.4% 的 mAP@50、94.8% 的精确率和 94.7% 的召回率，参数量为 2.65 M，计算量为 6.5 GFLOPs。F1-confidence 曲线确定 0.476 为报警阈值；基于 ONNX 的 Jetson Nano 部署及逐帧记录验证了该方法用于实时监测和选择性施药的可行流程。