Research On The Methods Of Polarization Imaging Simulation And Validation For Optical Remote Sensing

Polarization is one of the essential properties of light, and the combination of the information contained in light polarization, intensity and spectrum is important for improving the inversion precision of atmospheric aerosol optical properties as well as the ground target recognition for an optical remote sensing system. Currently, the orbiting polarization detectors are mostly serviced in cloud and aerosol parameters’detection for their low spatial resolution of about one kilometer or less, so it is not quite suitable for the target feature identification, in which the spatial resolution are expected to be several or tens of meters. For the feature of vectorization, the influence of observational geometric angle and the aerosol models on the polarized radiative transfer should be considered to make the judgement of whether the designation of a high resolution optical remote sensing system or the procedure of signal procession is reasonable or not. Simulation is a scientific and effective technique and a method of validation, so the key work in the paper is the simulation of polarization imaging.According to the physical processes of polarization imaging, two methods for simulating optical polarization remote sensing images are proposed in combination with the existing technical conditions. They are respectively the method based on airboard polarized images and the method based on intensity images. The principles and the processes of realization are also provided. Then, with the consideration of atmospheric vector radiative transfer model and the polarized Bidirectional Reflectance Distribution Function (p-BRDF) models of ground target, the influence of atmospheric model, aerosol model, ground target p-BRDF model and geometric conditions on polarization imaging are analyzed in detail. Results show that the polarization imaging data are influenced noticeably by the geometric conditions and aerosol models.As for the simulation method for space-based polarization imaging based on airborne polarization images, to get the aerial polarization image is the precondition, meanwhile the inversion of ground targets’p-BRDF through atmospheric correction is the key problem. This paper introduces the pretreatment processes of Directional Polarimetric Camera (DPC) aerial image data, and realizes the quantification of aerial images by the research on the image registration of aerial image and polarization calibration, radiation calibration, the three polarization direction images of the polarization band into reflectance images and polarized reflectance image. Based on the works above, the method of how to determine aerosol model and how to retrieve aerosol optical thickness are introduced in detail. Ground polarized reflectance images are obtained by excluding atmospheric effects, and the space-based polarized reflectance images are simulated and obtained in different atmospheric conditions. The images are used for analyzing the influence of atmospheric condition on polarized detection.As for the polarization imaging simulation based on the intensity images, the methods of retrieving p-BRDF of ground targets and remote sensing image classification are introduced. The simulated polarization images are obtained under different atmospheric conditions and different geometrical conditions by using the vector radiative transfer model, and the simulated intensity and polarization images under the same conditions are compared quantitatively. Results show that although the simulated intensity and polarization images blurs gradually with the decrease of atmospheric visibility, the degradation of the resolution of polarization images are weaker. The advantage of polarization imaging appears more prominent when the visibility is low or in the condition of backscattering or some particular directions. The fact that the clarity of polarization imaging is more sensitive to observational direction suggests that the simulation is helpful for the selection of a specific direction for polarization imaging. Eventually, it is helpful for enhancing the ability of identifying the ground target in haze conditions through polarization imaging.To test the validity of the polarization simulation results, a variety of evaluation methods are given and the polarization simulation images are compared with the experimental airboard images, which verifies the effectiveness of the polarization imaging simulation process.