Influence Analysis And Research On Correction Method Of Atmospheric Polarization Remote Sensing Image

In this paper, the influence of particle scattering, absorption, refraction and disturbance on the quality of polarization imaging is analyzed, and the correction algorithm of the polarization remote sensing image is established.Firstly, the influence of atmospheric scattering on polarization remote sensing is discussed. In order to study of the atmospheric light scattering and polarization state change, the scattering theories of the spherical particle are introduced. Analyzed the typical spectral distribution of aerosol particle group and particle scattering intensity, the influence of the degree of polarization, and reveals the differences in characteristics of single particles and particle scattering groups.Next, according to the discrepancy between the measured polarization information obtained from polarization remote sensing detection and the real polarization information, the measuring accuracy of the polarimatric bidirectional reflectance distribution function(BRDF), the polarization degree and the polarization phase angle can be affected by three system error sources in polarization remote imaging system, which is discussed. The error analysis model of polarization remote sensing detection is established. The influence of polarizer angle positioning error on measuring Stokes parameters is analyzed, and the influence of polarizer angle orientation on polarization detection accuracy is numerically simulated; based on the feature space interval of polarization remote sensing detection and the resolution ratio of the imaging CCD, the influence of the visual field space coincidence error on BRDF is discussed; then the influence of photon noise of imaging system and other inherent systematic errors on polarization detection results is analyzed. According to error transfer principle, the measuring errors caused by three error sources are compounded, and the general error model with respect to linear polarization degree and polarization phase angle is derived. Experiments show the error model of polarization remote sensing detection can truly reflect the influence of polarization detection system error sources on measuring accuracy of BRDF.Then, when the sky is not in the field of view, the atmospheric correction algorithm based on independent component analysis is proposed. According to the independence of the high frequency information of the atmosphere and scene light, the polarization degree of atmosphere is estimated by independent component analysis and mutual information optimization principle, and the airlight intensity at infinity is estimated with prior knowledge of the source image and atmospheric scattering physical model. Through experimental contrast analysis, the obtained atmospheric information results match the theoretical value, and the reasonability of atmospheric information estimation is verified. Meanwhile, for the influence of sensing image degradation caused by atmospheric medium, the linear polarization degree is fixed based on fuzzy rule, the airlight information is corrected, and the image quality is improved. The polarization remote sensing image reflects the terrain target characteristics more accurately after correction, and the ability of target detection and recognition is improved.Finally, since the atmospheric transmission, especially in the infrared band is not negligible it influences the polarimetric measurement. In order to suppress the influence of the atmosphere, the model of the infrared spectrum polarization parameters calibration algorithm based on infrared polarization transfer theory is proposed. Firstly, by analyzing the atmospheric radiation and transmission characteristics, the effective transmission model of the atmosphere is established, several factors of atmospheric transmittance were analyzed, such as wave bands, the atmospheric environment, the transmission path and the temperature.Experiments show that the algorithm presented in this thesis achieves higher robustness and image quality is improved, the polarization remote sensing image reflects the terrain target characteristics more accurately after correction, the ability of target detection and recognition is improved.