盐沼植被是蓝碳生态系统的重要组成部分,具有很强的固碳、储碳能力,叶面积指数决定了其生长、光合有效辐射的吸收比例和生物量。为准确、快速地提高叶面积指数估算精度,选取黄河三角洲碱蓬滩湿地,以我国土著植物碱蓬为研究对象,获取无人机高光谱遥感影像和测定地面光谱,结合区域土壤因子、植被光谱特征、高光谱影像纹理特征和植被覆盖度构建多模态数据,发展双优化策略的RF-PSO-DELM算法构建滨海湿地碱蓬叶面积指数反演模型,决定系数( )和均方根误差(RMSE)分别为0.9546,0.1341。与基于SVM、BP、ELM、DELM、PSO-DELM五种算法构建的碱蓬叶面积指数反演模型精度相比,决定系数 最高提高了0.2654,均方根误差RMSE最大降低了0.0828。与传统的反演模型SVM相比,RF-PSO-DELM模型具有更好的泛化性,融合多模态数据则可以有效提高反演模型精度。该研究进一步丰富了基于无人机高光谱遥感技术实现盐沼植被精准监测的理论和技术。

Salt marsh vegetation is an important part of the blue carbon ecosystem, with strong carbon sequestration and carbon storage capacity. The leaf area index determines its growth, photosynthetic active radiation absorption ratio and biomass. In order to improve the accuracy of leaf area index estimation accurately and rapidly, the Yellow River Delta Suaeda salsa shoal wetland was selected, the indigenous plant Suaeda salsa as the research object, the UAV hyperspectral remote sensing image was obtained and the ground spectrum was measured, combined with regional soil factors, vegetation spectral characteristics, hyperspectral image texture characteristics and vegetation coverage. Multi-modal data, through Random Forest (RF) feature selection for multi-modal data, and the RF-PSO-DELM algorithm with the dual optimization strategy was developed to construct the inversion model of the leaf area index of Suaeda salsa in coastal wetlands. The coefficient of determination ( ) and the root mean square error (RMSE) were 0.9546 and 0.1341, respectively. Compared with the inversion model accuracy of Suaeda salsa leaf area index constructed based on the five algorithms of SVM, BP, ELM, DELM and PSO-DELM, the coefficient of determination is increased by 0.2654 at most, and the root mean square error RMSE is reduced by 0.0828 at most. Compared with the traditional inversion model SVM, the RF-PSO-DELM model has better generalization, and the fusion of multimodal data can effectively improve the accuracy of the inversion model. This study further enriched the theory and technology for accurate monitoring of salt marsh vegetation based on UAV hyperspectral remote sensing technology.