Study On The Monitoring Of Urban Forest Biomass Based On Hyperspectral

Urban forest plays an important role in improving and maintaining urban ecological environment,and studying urban forest biomass and distribution has positive significance to urban ecosystem service evaluation.However,due to the particularity of urban forest in distribution pattern and operation mode,how to estimate its biomass accurately and efficiently is a challenge in both technology and method.But the traditional forest biomass estimation method is time and power consuming.Hyperspectral remote sensing technology can be effectively mining biomass and spectral characteristics of vegetation relations,implement effective monitoring of urban forest biomass.It can overcome the limitations of the spatial scale and timeliness of traditional biomass estimation methods.Based on the remote sensing technology,the measured ground data,the UAV hyperspectral image data and the OLI satellite image data source measured by the Avafield-2 portable hyperspectral terrestrial spectrometer were used as the research areas in Shanghai and Shanghai Jiaotong University.The spectral characteristics of the lower vegetation were analyzed and studied.Based on the biomass data,the optimal bands and vegetation indices were screened,and the regression model was established to study the hyperspectral inversion model of forest biomass in Shanghai and to estimate the biomass.The main research contents and results are as follows:(1)Based on the measured spectra of the vegetation in the study area,the measured vegetation has significant spectral characteristics of the vegetation,that is,there is a blue valley at the 510 nm band and a small reflection peak at 550 nm green band.In the 690 nm band Red Valley exists at 690~750 nm at the red edge.Based on the continuum removal method,the vegetation distinguishable wave band is obtained,most of the sub-wave bands fall in the visible wave band.(2)Extract spectral values of different landforms in remote sensing images to compare different differences.The results show that the UAV hyperspectral data can well characterize the differences between different landforms and different vegetation types.And the vegetation spectral curve and the ground measured spectral curve have a good consistency,but also in each band shows the spectral characteristics of the measured vegetation on the ground.(3)Based on the analysis of the vegetation biomass and the spectral bands,12 selected vegetation indices were calculated.Based on the statistical analysis of biomass and vegetation index,4 vegetation indices were determined as the modeling factors of the inversion model.The linear model and the nonlinear model were established respectively,and the model accuracy is verified.The quadratic polynomial biomass inversion model with RVI as univariate was determined as y=0.1816x~2-1.2368x+21.655,which was used as a model for estimating urban forest biomass in Shanghai.(4)Based on the forest biomass inversion model obtained above,combined with landsat 8 OLI satellite imagery,the urban forest biomass in Shanghai was statistically calculated,with a total of 4.75Tg and an average biomass density of 45.74t/ha.