Remote Sensing Quantitative Retrieval And Software Implementation Of Moso Bamboo Parameters

The study is part of the National Natural Science Foundation (grant, 30700638) and'948'item of national forestry bureau (grants, 2008-4-49). Bamboo is a special forest type mainly distributed in semi-tropical area of China, which plays an important role in regional ecosystem. Study of remote sensing methods in the moso bamboo forest is remote sensing information extraction and directly estimates aboveground biomass in moso bamboo forest using of statistical methods presently. Except for biomass, Moso bamboo forest biophysical parameter retrieval based on the remote sensing data is rarely reported. The studies in mapping spatial distribution of biophysical parameter can response the moso bamboo forest information and calculate other biophysical parameter, such as aboveground biomass, carbon content. Therefore, research in this area is important.Based on remote sensing data and supported by geographic information system (GIS) and global position system (GPS), the main contents are as follows:1. Using BP artificial neural network and maximum likelihood methods extract the moso bamboo forest information in the study area;2. Simplified the single moso bamboo model and establish the three-dimensional scene model using 9 moso bamboo forest samples;3. Based on the remote sensing images, obtained four-component endmember of moso bamboo forest using scene simulation results and unmixing the image in the study area;4. Inversion of moso bamboo forest clown closure, number of stems, leaf area index and the weighted average age using virtual reality scene as the result of the decomposition;5. Prepared remote sensing quantitative retrieval software for moso bamboo forest parameters using Matlab language.The main results are as follows:1. Three-dimensional scene model were established using simplified bamboo model combined with plot spatial information. The crown closure was used to assessment the accuracy of scene simulation, which indicated a high precision (R~2=0.78). The advantage of spectral mixture analysis method based on the simulation of real scenario is to use field data as priori knowledge and applied them to extract endmember. In addition, three-dimensional simulation model are introduced to two-dimensional linear spectral decomposition too;2. A compare analysis of the different methods of extraction of endmember such as simulated endmember, image endmember, and reference endmember, indicate that the simulated endmember based on the real scenario simulation method has the highest estimation accuracy(image endmember R~2=0.0001, reference endmember R~2=0.0034, modeled endmember R~2=0.2053), minimum relative error, and better robust;3. Inverted the moso bamboo forest clown closure (R~2=0.21), number of stems (R~2=0.49), leaf area index (R~2=0.59) and the weighted average age (R~2=0.40) using virtual reality scene as the result of the decomposition. This result not only provides a new technology for obtain accurate biophysical parameters from remote sensing data, but also provides a new reference for the management of moso bamboo forest;4. The spatial distribution maps of bamboo forest biophysical parameters in Anji country are obtained by inversion model. Such as in the southwest of Anji country, average age of bamboo forest is smaller, crown closure is higher, leaf area index is higher and stem number is relatively more.