| Wetland as the most important ecological system,not only has a strong productivity,but also for human beings to provide a valuable living environment.Because of the important role of wetland vegetation in the energy exchange of wetland ecosystems,the remote sensing extraction and biomass estimation of wetland vegetation information has become the hotspot of wetland research.Based on the GF-1 data,Landsat-OLI data and field data,this paper uses the 3S technology and statistical analysis method to realize the accurate extraction of the phragmites australis information and the remote sensing estimation of aboveground biomass in the Wuliangsuhai,the main conclusions as follows:1?The Landsat-OLI data and GF-1 data are collected in June,July and August.Then,remote sensing image processing are geometric correction,radiometric calibration and atmospheric correction,resampling.Above all of those processes are the basis for this study.Field sampling carried out during the same period of the satellite overpass.A total of 60 sampling points were selected,40 of them were used for model construction,and 20 were used as precision verification.The points could meet the requirements of model construction and accuracy verification.2?In this study,we first proposed normalized differential inverse vegetation index NDVI-1 =(aB3-B4)/(B3 + B4)(a is undetermined coefficients).The construction of the index is based on the principle of remote sensing optics,and the NDVI formula is used as the prototype.Through the inverse construction of the NDVI index,the vegetation information is weakened,the non-vegetation information is enhanced,and the contrast between vegetation and non-vegetation is enhanced,To achieve the extraction of vegetation information.The application of inverse vegetation index in Wuliangsuhai has effectively improved the accuracy of phragmites classification and the accuracy of biomass estimation.3?Based on the multi-seasonal GF-1 data and object-oriented multi-seasonal classification method in 2016,the NDVI-1?NDVI?RVI?DVI?ARVI and various shape features are selected to analyze the spatial distribution of phragmites in Wuliangsuhai.The results showed that the area of phragmites in Wuliangsuhai is 134.243 km,the area of water is 161.092 km,the area of phragmites is about 1/2 of the total area of Wuliangsuhai;the spatial extraction information accuracy of phragmites was 0.933,the Kappa coefficient was 0.906.The main distribution area of phragmites is the northern part and the middle part where the Irrigation wastewater pour into Wuliangsuhai.4?The data of NDVI,RVI,DVI,ARVI,NDVI-1 vegetation indices and GF-1 fourth band reflectance B4 were used to calculate the linear and non-linearity of the dry and wet biomass.The results show:(1)The stem biomass of phragmites was highly correlated with the total amount of dry biomass,and the formula is:total dry biomass = 1.542×stem biomass + 102.320(R2 = 0.982).The relevance was significantly higher than that of leaf dry biomass and total dry biomass(R2 = 0.939).(2)The accuracy of the multi-linear model of dry aboveground biomass is higher than that of the univariate nonlinear regression model.The prediction accuracy is 0.809,R2 = 0.758 and RMSE is 64.150g/m2.The aboveground dry biomass estimation model is:y = 1.542 ×(-889.621-2742.368×NDVI-1+ 1515.993 x NDVI + 26.763×RVI-397.458×ARVI-0.023 x DVI + 0.106×B4)+ 102.320;According to the model,we can calculatethe total amount of aboveground dry biomass of phragmites in 2016 is 7.922 x 104t,the average weight of phragmites is 1053.851 g/m2.The spatial information extraction method and the biomass prediction model used in this study can be used for other wetlands,and have certain promotion prospects. |
PDF Full Text Request