Improved Lightweight YOLOv8 Model for Rice Disease Detection in Multi-Scale Scenarios

Agronomy, Journal Year: 2025, Volume and Issue: 15(2), P. 445 - 445

Published: Feb. 11, 2025

In response to the challenges of detecting rice pests and diseases at different scales difficulties associated with deploying running models on embedded devices limited computational resources, this study proposes a multi-scale pest disease recognition model (RGC-YOLO). Based YOLOv8n network, which includes an SPPF layer, introduces structural reparameterization module (RepGhost) achieve implicit feature reuse through reparameterization. GhostConv layers replace some standard convolutions, reducing model’s cost improving inference speed. A Hybrid Attention Module (CBAM) is incorporated into backbone network enhance ability extract important features. The RGC-YOLO evaluated for accuracy time dataset, including bacterial blight, blast, brown spot, planthopper. Experimental results show that achieves precision (P) 86.2%, recall (R) 90.8%, mean average Intersection over Union 0.5(mAP50) 93.2%. terms size, parameters are reduced by 33.2%, GFLOPs decrease 29.27% compared base model. Finally, deployed Jetson Nano device, where per image 21.3% model, reaching 170 milliseconds. This develops successfully field devices, achieving high-accuracy real-time monitoring providing valuable reference intelligent equipment in unmanned farms.

Language: Английский

Integrating machine learning models for optimizing ecosystem health assessments through prediction of nitrate–N concentrations in the lower stretch of Ganga River, India

Environmental Science and Pollution Research, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 30, 2025

Language: Английский