集成Landsat OLI和机载LiDAR条带数据的亚热带森林生物量制图  

Mapping biomass by integrating Landsat OLI and airborne LiDAR transect data in subtropical forests

作　　者：曹林[1] 徐婷[1] 申鑫[1] 佘光辉[1] 

机构地区：[1]南京林业大学南方现代林业协同创新中心,江苏南京210037

出　　处：《遥感学报》2016年第4期665-678,共14页NATIONAL REMOTE SENSING BULLETIN

基　　金：江苏省自然科学基金(编号:BK20151515);江苏省高校自然科学研究项目(编号:14KJB220002);江苏高校优势学科建设工程资助项目(PAPD)~~

摘　　要：以亚热带天然次生林为研究对象,借助一个条带的少量LiDAR点云数据和覆盖整个研究区的免费Landsat OLI多光谱数据,并结合地面实测数据,探索森林生物量低成本高精度制图方法。首先,提取了OLI和LiDAR特征变量,并与地上和地下生物量进行相关分析以筛选变量;然后,借助LiDAR数据覆盖区的样地和条带LiDAR数据构建"LiDAR生物量模型";再从LiDAR反演生物量的结果中进行采样,结合OLI特征变量构建"LiDAR-OLI模型";最后,与单独使用OLI多光谱数据建立的"OLI估算模型"结果进行比较,分析精度并验证新方法的效果。结果表明,"LiDAR-OLI模型"对地上和地下生物量的模型拟合效果较好且均优于"OLI模型",且其交叉验证的精度也较高并优于"OLI模型",从而证明了新方法的可靠性及有效性。本研究为主、被动遥感技术在中小尺度上协同反演森林参数提供了实验基础,也为基于全覆盖免费OLI多光谱数据及条带LiDAR数据的低成本森林生物量制图探索了技术路线。Accurate estimation of forest biomass is critical for modeling the carbon cycle and mitigating climate changes. Integration of multi-spectral satellite data and airborne LiDAR data can accurately estimate the biomass. However, the application of this strategy is lim-ited in subtropical forests, particularly in China. In this study, a novelapproach was assessed using one strip of LiDAR point cloud and ＂wall- to-wall＂ Landsat OLI free multi-spectral data combined with field-measured plot data to generate a low-cost and high-accuracy forest bio- mass map in a subtropical secondary forest in southeast China. Sixty square plots （30 m×30 m） were established across the study site. First, the OLI data were processed by atmospheric and geometric correction, and LiDAR point clouds were extracted from the raw full-waveform LiDAR data. Second, fivesets of OLI and three sets of LiDAR metrics were extracted, and correlation analysis was performed with the field estimates of above- and below-ground biomass foroptimal metrics selection. Third, the ＂LiDAR biomass model＂ was fitted to LiDAR met- rics extracted from the strip of LiDAR point cloud and the field plots within the strip. The ＂LiDAR-OLI biomass model＂ was fitted tothe OLI metrics and forest biomass estimated by the LiDAR data. Finally, the performance of the predictive models and the accuracy of the cross-validation results were evaluated through comparison with the accuracy assessment results of the ＂OLI biomass model.＂ [Result] The ＂LiDAR-OLI biomass model＂ （R2 of above- and below-ground biomass estimation=0.69 and 0.56, respectively） exhibited improved per- forrnance than the ＂OLI biomass model＂ （Rz of above- and below-ground biomass estimation=0.69 and 0.56, respectively）. The relative bi- ases of above- and below-ground biomass estimation increased by 14% and 15%, respectively. The mean differences in the cross-validation results for the ＂LiDAR-OLI biomass model＂ （mean differences in above- and below-ground biomass estimat

关 键 词：升尺度 生物量反演 亚热带森林 机载LiDAR数据 LANDSAT 8 OLI影像 多元回归模型 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]