Zheng Huiwen, Liu Changjiang, Zhong Lei, Wang Jie, Huang Junming, Lin Fang, Ma Xu, Tan Suiyan
College of Electronic Engineering, South China Agricultural University, Guangzhou, Guangdong, China.
College of Mechanical and Electrical Engineering, Xinjiang Agricultural University, Urumqi, China.
Front Plant Sci. 2025 Apr 16;16:1561632. doi: 10.3389/fpls.2025.1561632. eCollection 2025.
Detection of rice panicles and recognition of rice growth stages can significantly improve precision field management, which is crucial for maximizing grain yield. This study explores the use of deep learning on mobile phones as a platform for rice phenotype applications.
An improved YOLOv8 model, named YOLO_Efficient Computation Optimization (YOLO_ECO), was proposed to detect rice panicles at the booting, heading, and filling stages, and to recognize growth stages. YOLO_ECO introduced key improvements, including the C2f-FasterBlock-Effective Multi-scale Attention (C2f-Faster-EMA) replacing the original C2f module in the backbone, adoption of Slim Neck to reduce neck complexity, and the use of a Lightweight Shared Convolutional Detection (LSCD) head to enhance efficiency. An Android application, YOLO-RPD, was developed to facilitate rice phenotype detection in complex field environments.
The performance impact of YOLO-RPD using models with different backbone networks, quantitative models, and input image sizes was analyzed. Experimental results demonstrated that YOLO_ECO outperformed traditional deep learning models, achieving average precision values of 96.4%, 93.2%, and 81.5% at the booting, heading, and filling stages, respectively. Furthermore, YOLO_ECO exhibited advantages in detecting occlusion and small panicles, while significantly optimizing parameter count, computational demand, and model size. The YOLO_ECO FP32-1280 achieved a mean average precision (mAP) of 90.4%, with 1.8 million parameters and 4.1 billion floating-point operations (FLOPs). The YOLO-RPD application demonstrates the feasibility of deploying deep learning models on mobile devices for precision agriculture, providing rice growers with a practical, lightweight tool for real-time monitoring.
检测水稻穗粒并识别水稻生长阶段可显著提高精准田间管理水平，这对于实现粮食产量最大化至关重要。本研究探索将深度学习应用于手机平台，以实现水稻表型分析。
提出一种改进的YOLOv8模型，即YOLO高效计算优化模型（YOLO_ECO），用于检测孕穗期、抽穗期和灌浆期的水稻穗粒，并识别生长阶段。YOLO_ECO进行了关键改进，包括在主干网络中用C2f-更快块-有效多尺度注意力模块（C2f-更快-EMA）替换原始的C2f模块，采用精简颈部以降低颈部复杂度，并使用轻量级共享卷积检测（LSCD）头来提高效率。开发了一个安卓应用程序YOLO-RPD，以方便在复杂田间环境中进行水稻表型检测。
分析了使用不同主干网络、量化模型和输入图像大小的模型对YOLO-RPD性能的影响。实验结果表明，YOLO_ECO优于传统深度学习模型，在孕穗期、抽穗期和灌浆期的平均精度分别达到96.4%、93.2%和81.5%。此外，YOLO_ECO在检测遮挡穗粒和小穗粒方面具有优势，同时显著优化了参数数量、计算需求和模型大小。YOLO_ECO FP32-1280的平均精度（mAP）为90.4%，有180万个参数和41亿次浮点运算（FLOP）。YOLO-RPD应用程序证明了在移动设备上部署深度学习模型用于精准农业的可行性，为水稻种植者提供了一个实用的轻量级实时监测工具。
