福建电脑2025，Vol.41Issue(3)：12-15,4.DOI:10.16707/j.cnki.fjpc.2025.03.003
无人机影像作物与杂草分割技术研究
Research on Crop and Weed Segmentation Techniques for UAV Imagery
潘涛 1王俊玲1
作者信息
- 1. 南京理工大学紫金学院计算机与人工智能学院 南京 210000
- 折叠
摘要
Abstract
Weeds have a serious impact on crop yield and quality,and traditional weed control methods have problems such as high labor demand and environmental pollution.To address this issue,this paper proposes an efficient semi supervised semantic segmentation model to achieve automated differentiation between crops and weeds.Adopting the semi supervised segmentation model SWeedNet,the segmentation accuracy is improved through multi-scale enhancement modules and online difficult sample mining techniques,while utilizing consistency regularization to enhance the robustness of the model to environmental changes.The experimental results show that SWeedNet has excellent segmentation performance on public datasets and has good potential for application in complex agricultural environments.关键词
农作物识别/杂草分割/半监督作物与杂草分割模型/智能农业Key words
Crop Identification/Weed Segmentation/SWeedNet/Intelligent Agriculture分类
林学引用本文复制引用
潘涛,王俊玲..无人机影像作物与杂草分割技术研究[J].福建电脑,2025,41(3):12-15,4.
