基于大光斑雷达数据和光学遥感数据估算江西省森林树高和森林生物量-estimation of forest tree height and forest biomass in jiangxi province based on large spot radar data and optical remote sensing data.docxVIP

AbstractUsing forest in Jiangxi province as the research area, estimate the forest canopy average height and biomass by multispectral Landsat TM data and GLAS waveform data. Extracted waveform extent, leading edge extent and trailing edge extent from the GLAS waveform data as the waveform characteristic parameters to represent waveform features, also extracted terrain index and terrain standard deviation from ASTER GDEM data to reduce the influence of terrain factors on the waveform data. Using waveform characteristic parameters (waveform extent, leading edge extent and trailing edge extent) and terrain feature parameters (terrain index and terrain standard deviation) build model to estimate tree height, completed the GLAS full waveform data processing algorithms, set up forest canopy height estimation model that can reduce the influence of terrain factors. Considering the large-footprint LiDAR data were spatially discrete and lack of imaging capability, the study brought forward the estimation model of regional forest canopy height by combine with optical remote sensing images. Finally estimated regional scale forest biomass of the study area by canopy height and field measured biomass data.The research results show: using GLAS data, ASTER GDEM data and measured data to estimate forest canopy height have high accuracy. The model which contain waveform extent, leading edge extent, trailing edge extent and terrain standard deviation has the highest correlation coefficient (0.6505). Build up the linear, power function, index and logarithmic 6 kinds of continuous forest canopy height models by large laser radar GLAS data and optical Landsat TM data , model results show that the linear correction the highest correlation coefficient (0.7219); build up the linear, power function, index and logarithmic 6 kinds of forest biomass estimate models by forest biomass measured data and tree height, the results show that the power function model has the highest correlation coeffic