In recent years,Polarization SAR(PolSAR)has been widely used in the filed of crop biomass estimation.However,high dimensional features extracted from PolSAR data will lead to information redundancy which will result i...In recent years,Polarization SAR(PolSAR)has been widely used in the filed of crop biomass estimation.However,high dimensional features extracted from PolSAR data will lead to information redundancy which will result in low accuracy and poor transfer ability of the estimation model.Aiming at this problem,we proposed a estimation method of crop biomass based on automatic feature selection method using genetic algorithm(GA).Firstly,the backscattering coefficient,the polarization parameters and texture features were extracted from PolSAR data.Then,these features were automatically pre-selected by GA to obtain the optimal feature subset.Finally,based on this subset,a support vector regression machine(SVR)model was applied to estimate crop biomass.The proposed method was validated using the GaoFen-3(GF-3)QPSΙ(C-band,quad-polarization)SAR data.Based on wheat and rape biomass samples acquired from a synchronous field measurement campaign,the proposed method achieve relative high validation accuracy(over 80%)in both crop types.For further analyzing the improvement of proposed method,validation accuracies of biomass estimation models based on several different feature selection methods were compared.Compared with feature selection based on linear correlation,GA method has increased by 5.77%in wheat biomass estimation and 11.84%in rape biomass estimation.Compared with the method of recursive feature elimination(RFE)selection,the proposed method has improved crops biomass estimation accuracy by 3.90%and 5.21%,respectively.展开更多
Optical remote sensing allows to efficiently monitor forest ecosystems at regional and global scales.However,most of the widely used optical forward models and backward estimation methods are only suitable for forest ...Optical remote sensing allows to efficiently monitor forest ecosystems at regional and global scales.However,most of the widely used optical forward models and backward estimation methods are only suitable for forest canopies in flat areas.To evaluate the recent progress in forest remote sensing over complex terrain,a satellite-airborne-ground synchronous Fine scale Optical Remote sensing Experiment of mixed Stand over complex Terrain(FOREST)was conducted over a 1 km×1 km key experiment area(KEA)located in the Genhe Reserve Areain 2016.Twenty 30 m×30 m elementary sampling units(ESUs)were established to represent the spatiotemporal variations of the KEA.Structural and spectral parameters were simultaneously measured for each ESU.As a case study,we first built two 3D scenes of the KEA with individual-tree and voxel-based approaches,and then simulated the canopy reflectance using the LargE-Scale remote sensing data and image Simulation framework over heterogeneous 3D scenes(LESS).The correlation coefficient between the LESS-simulated reflectance and the airborne-measured reflectance reaches 0.68-0.73 in the red band and 0.56-0.59 in the near-infrared band,indicating a good quality of the experiment dataset.More validation studies of the related forward models and retrieval methods will be done.展开更多
基金National Key R&D Program of China(No.2017YFB0502700)Project of The Technique of Accurate Surface Parameters Inversion Using GF-3 Images(No.03-Y20A11-9001-15/16)National Natural Science Foundation of China(No.41801289)。
文摘In recent years,Polarization SAR(PolSAR)has been widely used in the filed of crop biomass estimation.However,high dimensional features extracted from PolSAR data will lead to information redundancy which will result in low accuracy and poor transfer ability of the estimation model.Aiming at this problem,we proposed a estimation method of crop biomass based on automatic feature selection method using genetic algorithm(GA).Firstly,the backscattering coefficient,the polarization parameters and texture features were extracted from PolSAR data.Then,these features were automatically pre-selected by GA to obtain the optimal feature subset.Finally,based on this subset,a support vector regression machine(SVR)model was applied to estimate crop biomass.The proposed method was validated using the GaoFen-3(GF-3)QPSΙ(C-band,quad-polarization)SAR data.Based on wheat and rape biomass samples acquired from a synchronous field measurement campaign,the proposed method achieve relative high validation accuracy(over 80%)in both crop types.For further analyzing the improvement of proposed method,validation accuracies of biomass estimation models based on several different feature selection methods were compared.Compared with feature selection based on linear correlation,GA method has increased by 5.77%in wheat biomass estimation and 11.84%in rape biomass estimation.Compared with the method of recursive feature elimination(RFE)selection,the proposed method has improved crops biomass estimation accuracy by 3.90%and 5.21%,respectively.
基金supported in part by the National Basic Research Program of China(2013CB733400)in part by the Natural Science Foundation of China(41930111 and 41871258)+1 种基金in part by the Youth Innovation Promotion Association CAS under Grant 2020127in part by the‘Future Star’Talent Plan of the Aerospace Information Research Institute of Chinese Academy of Sciences under Grant Y920570Z1F.
文摘Optical remote sensing allows to efficiently monitor forest ecosystems at regional and global scales.However,most of the widely used optical forward models and backward estimation methods are only suitable for forest canopies in flat areas.To evaluate the recent progress in forest remote sensing over complex terrain,a satellite-airborne-ground synchronous Fine scale Optical Remote sensing Experiment of mixed Stand over complex Terrain(FOREST)was conducted over a 1 km×1 km key experiment area(KEA)located in the Genhe Reserve Areain 2016.Twenty 30 m×30 m elementary sampling units(ESUs)were established to represent the spatiotemporal variations of the KEA.Structural and spectral parameters were simultaneously measured for each ESU.As a case study,we first built two 3D scenes of the KEA with individual-tree and voxel-based approaches,and then simulated the canopy reflectance using the LargE-Scale remote sensing data and image Simulation framework over heterogeneous 3D scenes(LESS).The correlation coefficient between the LESS-simulated reflectance and the airborne-measured reflectance reaches 0.68-0.73 in the red band and 0.56-0.59 in the near-infrared band,indicating a good quality of the experiment dataset.More validation studies of the related forward models and retrieval methods will be done.