基于多源遥感数据的森林生物量定量评价研究

Research on Quantitative Evaluation of Forest Biomass Based on Multi-Source Remote Sensing Data

以ALOS PALSAR L波段双极化FBD微波遥感数据及多光谱光学遥感数据AVNIR-2为基础,对数据进行预处理,利用地面云南松林样地坐标,提取HH、HV双极化后向散射系数及极化比值3个因子,结合光学遥感数据提取4个波段值及NDVI、RVI 2个植被指数,作为云南松林生物量估测因子。分别以微波数据、光学数据、微波及光学数据结合的多源遥感数据,建立3个云南松林生物量估测模型。结果表明:所建模型经方差分析均达到显著相关或极显著相关水平;PALSAR L波段双极化后向散射系数,可以反映森林生物量的变化,但反演精度有待进一步提高;AVNIR-2数据模型优于PALSAR L波段双极化数据模型;多源数据模型与光学数据模型的估测精度相近。

Based on ALOS PALSAR L-band dual-polarization FBD microwave remote sensing data and multi- spectral optical remote sensing data AVNIR-2, preprocessing was conducted, then three microwave data factors including HH and HV dual-polarization backscattering coefficients and polarization ratio, and four original bands and two vegetation indices of NDVI, RVI which combined with optical remote sensing data were extracted with the coordinates of Pinus yunnanensis forest sample plots. After that, the three estimation models of Pinus yunnanensis forest biomass were built by using microwave data, optical data, and the multi-source remote sensing data combined with microwave and optical data. Research suggests that and the models reached the significant or extremely significant correlation level by variance analysis. PALSAR L-band dual-polarization backscattering coefficient could be reflected the changes of forest biomass, but inversion accuracy needs to be improved. AVNIR-2 data model was better than PALSAR L-band dual-polarizatlon data model. The estimation accuracy of multl-source data model was similar to optical data model.