Infrared and Laser Engineering, Volume. 48, Issue 11, 1126001(2019)

Optimization for a polarimetic dehazing method based on the circularly polarized light

Wang Hui, Wang Jin, Li Xiaobo, Hu Haofeng, and Liu Tiegen
Author Affiliations
  • [in Chinese]
  • show less

    In a typical polarimetric dehazing method, the estimation of the degree of polarization (DoP) of the backward scattered light is a key factor to affect the dehazing effect. In the present paper, based on the polarization maintaining characteristic of the circularly polarized light in the Mie scatting media, the traditional polarimetric dehazing method was optimized which was suitable for use in the strong scattering environment. The variation patterns of the DoP of the light field received by the CCD, in the light source illuminations with different polarization states and in different densities of the scattering particles, were discussed, based on which a simple and easy method for estimating the DoP of the scattered light was proposed. This method can improve the effect of polarimetric dehazing method, without increasing the complexity of the system. It can be seen from the experiment results that under a condition of strong scattering, the present method can provide the dehazed images with an EME value being 20.4% higher than that by the traditional method. In addition, in this method, it is not necessary to determine the background region(s) as in the traditional method for polarimetric dehazing, thus lowering the calculation complexity.

    Tools

    Get Citation

    Copy Citation Text

    Wang Hui, Wang Jin, Li Xiaobo, Hu Haofeng, Liu Tiegen. Optimization for a polarimetic dehazing method based on the circularly polarized light[J]. Infrared and Laser Engineering, 2019, 48(11): 1126001

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: 图像处理

    Received: Feb. 5, 2019

    Accepted: May. 15, 2019

    Published Online: Dec. 9, 2019

    The Author Email:

    DOI:10.3788/irla201948.1126001

    Topics