Research On Microvave Scattering Mechanism And Inversion Of Biomass Of Wheat

Microwave remote sensing, which has advantages of observing the physical and geometric characteristics of surface objects, as well as that penetrating vegetation, working all the time under different weather conditions, has been applied to many fields, such as vegetation monitoring, environmental assessment, hazards warning. However, microwave scattering mechanism is so extremely complex that different surface objects present various scattering characteristics. For making most of the unique properties of the interaction between microwave and surface objects and the advantages of large-scale and full-temporal synthetic aperture radar(SAR) images, exploring the scattering mechanism, modeling the parameters and developing the quantitative inversion methods of parameters have therefore become a hotspot for surface objects in scientific research.In this dissertation, based on the experimental data from the scatterometer and the surface parameters collection, the research studied wheat, which has the second production in the world. Based on the analysis of microwave scattering characteristics of wheat from different growth periods; the scattering models for wheat were improved and developed. Further, the inversion methods were compared, and two proper methods were applied to retrieve the wheat biomass from advanced synthetic aperture radar(ASAR) images. In addition, the procedure for cereal parameters inversion based on SAR images was improved. The main work of the dissertation is summarized as follows:(1)The research selected two groups of the measured data from 2008-2009 and 2010-2011 as the experimental data, which were measured in the experimental location in west Sichuan province. The measured data included two categories: backscattering coefficients and surface parameters. The backscattering coefficients were derived from four bands(L, S, C, and X), full-polarization(HH, HV, VH, and VV), different incidence angles(0-90°) and different azimuth angles(0-360°) at each wheat growth stage; The collected surface parameters included parameters of wheat components(length, thickness, width, biomass, and moisture of wheat stems, leaves and ears), leaf area index(LAI), stem density, leaf inclination, and soil parameters(moisture, roughness and composition) at each wheat growth stage. Under different conditions with various input data including the parameters of incidence wave(frequency, incidence angle, azimuth angle, and polarization), macro-structures(row and column) and microscopic features(parameters of wheat and soil), the scattering characteristics research and theoretical analysis were performed.(2)Based on the previous study of microwave scattering mechanism of wheat, the microwave scattering models including the empirical model, the semi-empirical models and the theoretical model were developed. The empirical one was established with respect to backscattering coefficients and wheat parameters; An entire wheat growth cycle is divided into two periods, ie., with and without wheat ears, which is an important component of wheat. Thus, the semi-empirical models were built by modifying the Water Cloud(WC) model and the Michigan Microwave Canopy Scattering(MIMICS) model corresponding to the two periods, and then the two semi-empirical models were established. Further, based on the vector radiative transfer(VRT) theory, the theoretical model of wheat was constructed by dividing wheat canopy into three layers, which are wheat ears, wheat stems and leaves and underlying soil surface. All the models were analyzed and verified with measured data.(3)The inversion methods were built from three kinds of wheat scattering models and the measured surface parameters. First, an empirical inversion algorithm was established based on the relationship between wheat parameters and backscattering coefficients, and then the algorithm was verified by the measured data. Second, two semiempirical inversion methods were performed based on the modified WC model and the simplified MIMICS model, and the effectiveness of such two methods were compared by using the measured data. Third, the inversion method of Neural Network was established, where the training data were generated by the established theoretical model. The accuracy of the inversion method was verified by comparing the measured data with the inversion results obtained by the trained neural network method.(4)The backscattering coefficients of wheat were retrieved from the preprocessed ASAR images acquired synchronously with ground-based scatterometers measurements from the experimental area. Further, the wheat biomasswere retrieved from the ASAR images based on the simplified MIMICS model and the Neural Network, and such retrieved wheat biomass were verified by measured biomass from the field investigation and the retrieved biomass from the optical TM images of the corresponding experimental area.Wheat is one of the regular planting row crops. Based on the study on the microwave scattering mechanism and the parametric inversion of wheat, the research found that row wheat at different growth stages has different influence on radar backscatter for different microwave band and polarizations, which can lay a foundation for modeling row cereal crops through comparing the backscatter differences and analyzing the differences based on VRT theory. Further, considering the influence of wheat ears on radar backscatter, the WC model was modified and the MIMICS model was simplified. In Addition, theoretical scattering model of wheat based on VRT was established. The research improved the process of inversion of cereal parameters by integrating measurements, analysis of scattering characteristics, scattering modeling, research on inversion methods, cereal parameters inversion from SAR images and verification by measured and optical data.This dissertation enriches the research of wheat microwave scattering mechanism and makes certain contributions to quantitative inversion of wheat parameters with largescale SAR images. Further, it provides the valuable scientific research documentsfor monitoring crop growth status and vegetation eco-system based on SAR images.