| In order to retrieve surface parameters, such as surface reflectance and temperature etc., it is very important to improve the quantitative processing accuracy of airborne remote sensing data with high spatial resolution and spectral resolution. After the major bottlenecks are summarized at first, this paper is devoted to solve these problems based on two aspects: the atmospheric attenuation and the surface complexity.The atmospheric influences of the airborne remote sensing data are mainly due to the variability of low-layer atmosphere. According to the analysis of features of the data acquired from airplane, a new algorithm for retrieving atmospheric column water vapor and temperature profile is suggested based on measurements of downward sky thermal radiation above the surface. The results show that the new algorithm can estimate atmospheric column water vapor accurately, and air temperature at lower than 3 kilometers can also be retrieved.The complexity of surface parameters can be characterized by cross radiance and BRDF in VIS/NIR band. At first, the proportion of cross radiance to the radiance received by sensor is analyzed using MODTRAN4 code, and then the influences of a set of factors on cross radiance are simulated. The main factors considered in this paper include: variation of aerosol phase function, aerosol optical thickness, sensor IFOV and viewing angle, flying height, and wavelength. The results show that the variance of cross radiance could be resulted from every factor mentioned above, so it must be computed based measurement condition. A new methodology is developed to eliminate cross radiance from total radiance based on a modification of analytical solution of PSF and combination of it with radiation transfer code of MODTRAN. The applications to AMTIS image show a good precision.For the retrieval of BRDF information from airborne AMTIS data, a software system is modularized based on features of AMTIS data, and can also be used for processing of other types remote sensing data.As-for the thermal infrared band, the thermal emissivity is the major factor of surface complexity. In this thesis a series of tests have been done in order to find out the IR feature of mixed pixel composed by objects with different temperatures. The results show that the augment of the effective emissivity exists when we define the temperature as an area1Uweighted average temperature. The test validated Li-Wang-Strahler's scale-corrected Planck law for non-isothermal surfaces. And also in the dissertation, an experiment method is suggested to measure surface emissivity in field based on ISSTES algorithm (Iterative spectrally smooth temperature-emissvity separation). A deliberately designed set of measurements was carried out to study feature of the spectral emissivity of soil with different grain size or with different SiO2 and water content. The result shows that emissvity becomes lower with increasing sand content and higher with increasing water content in 8-9.5um region; In 1 l-14um region, the variance of emissivity is not as obvious as in 8-11 jam region. According to the result, a method is proposed for retrieving parameters of soil from TTR emissivity ratio.By the end of this dissertation, the processing flow of airborne thermal multi-band data is represented in details. The noises in the radiative calibration of airborne thermal infrared data are analyzed systematically, and a new algorithm used for eliminating the striped noise based on the spline interpolation is developed. The results show a good effect in the processing of airborne OMIS thermal infrared data.Our research could improve the quantitative research of airborne multi-angle and multi-band remote sensing data.
|
PDF Full Text Request