摘要
以江西省某森林资源管理工程为研究对象,探索激光雷达与多光谱遥感数据融合技术在森林参数提取中的应用效果。通过对1m分辨率激光雷达数据和10m分辨率多光谱遥感影像数据的融合,提出数据预处理、特征提取、融合策略及模型优化等方法,并使用支持向量机、随机森林、极限梯度提升和长短期记忆网络等模型进行分析。结果表明,融合方法在树高提取精度上达到了±1m,树冠面积估算精度为±5m^(2),森林密度误差为±0.1树木/m^(2),归一化植被指数误差降低至±0.05。此外,该方法在提高森林参数提取精度的同时,展现了比传统方法更高的精度,且模型稳定性更高,具有较高的推广价值。
A forest resource management project in Jiangxi Province is selected as the research object to investigate the application effectiveness of LiDAR and multispectral remote sensing data fusion technology in forest parameter extraction.Through the fusion of 1 m resolution LiDAR data and 10 m resolution multispectral remote sensing imagery,methods including data preprocessing,feature extraction,fusion strategies,and model optimization are proposed,and models such as Support Vector Machine,Random Forest,Extreme Gradient Boosting,and Long Short-Term Memory network are utilized for analysis.The results indicate that the fusion method achieves an accuracy of±1 m in tree height extraction,±5 m^(2)in canopy area estimation,and±0.1 trees/m^(2)in forest density error,while reducing the normalized difference vegetation index error to±0.05.Moreover,this method not only improves the accuracy of forest parameter extraction but also demonstrates higher precision and greater model stability compared to traditional methods,showing strong potential for widespread application.
作者
游晋卿
YOU Jinqing(PowerChina Jiangxi Electric Power Engineering Co.,Ltd.,Nanchang,Jiangxi 330000,China)
出处
《自动化应用》
2025年第16期197-200,共4页
Automation Application
关键词
激光雷达
多光谱遥感
数据融合
森林参数提取
LiDAR
multispectral remote sensing
data fusion
forest parameter extraction
