Finer three-dimensional radiative transfer simulation using surface texture mapping
Three-dimensional(3D)Radiative Transfer(RT)simulation is an important method to use in studying the mechanism of quantitative remote sensing and improving inversion accuracy.Computer graphics combined with laser scanning data or photogrammetry images can help produce realistic centimeter-level vegetation scenes,including leaves,branches,trunks,and ground polygons.Current 3D RT models can utilize these kinds of reconstructed scenes to simulate decimeter-level spaceborne or airborne remote sensing images.However,sparse forest areas still exhibit some deviations between the simulated and actual images.Such deviation is somehow due to the reconstruction error of the 3D structure of the trees or understorey,while the limitation of the average component spectrum of soil background plays another role,which hardly reflects the random heterogeneity of surface reflectance.To solve the second problem,this study extends the 3D radiosity applicable to porous individual objects for directional reflectance over complex vegetated scenes(RAPID)model,adding two weight-coefficient texture matrices based on the end-member spectra of bare soil and dense vegetation,realizing the finer simulation of surface reflection spectrum.A multispectral or hyperspectral image or at least an RGB image containing the background texture information should be given as input.First,the Normalized Difference Vegetation Index(NDVI)can be used to extract the pixel spectra of bare soil(NDVI:～0.2)and dense vegetation(～maximum NDVI).Second,the two weight coefficients for each background pixel are fitted.Last,RAPID simulates the reflectance image using the texture matrices filled with the weight coefficients.By using real images from the Tianfeng Mountain of Yunnan Province and Dayekou of Gansu Province,the improvement of image quality by considering heterogeneity on image simulation has been evaluated.The Tianfeng Mountain scene represents a sparse Yunnan pine(Pinus yunnanensi)forest with heterogeneous background(bare soil,shrub,grass,or dead wood)in the stand,with a hyperspectral image used as input.The Dayekou scene shows the Qinghai spruce(Picea carassifolia)forest segmented by a large grass patch,with a Quickbird image used as input.Results show the following:(1)the two weight coefficients can be solved directly from real images when the end member of bare soil and dense vegetation are chosen properly.(2)If two heterogeneous weight coefficients are assigned to each surface pixel,high coincidence exists between the synthetic spectrum and the spectral curve of the real image(R2 is better than 0.98,RMSE=0.016).(3)The simulated image qualities are remarkably improved in the Tianfeng Mountain and Dayekou scenes(R2 increased by 0.096 and 0.041,respectively,and RMSE decreased by 0.015 and 0.01,respectively).(4)Near infrared(NIR)band images can also be predicted from three-band RGB images.When only RGB images are used as input,a similar texture is reproduced,but a certain deviation exists in the NIR reflectance values(0.006-0.027).The method proposed in this study can simulate more precise submeter high resolution satellite images,which can provide high-quality samples for deep learning and improved training data for quantitative inversion.
NETL
NSTL
万方数据
