Research on the dehazing effect of orthogonal polarization method based on atmospheric scattering model and considering extinction ratio parameter

ABSTRACT

The hazy weather is currently one of the most common atmospheric environments. It is an urgent problem to realize clear imaging in hazy environment. Orthogonal polarization dehazing is one of the classical dehazing techniques attracting a widespread attention from research workers. However, while using this method, the extinction ratio (ER) parameter of the polarizer was neglected leading to poor dehazing. Therefore, we have introduced the ER of polarizers into the orthogonal polarization dehazing technique in this paper. In the process of sky region division, this study proposes to use HSV, RGB and Lab colour channels to create a comprehensive mask to select the sky region. And the morphological corrosion operation is utilized to increase the region connectivity for better sky region division. Based on subjective visualization, it can be found that the dehazing effect is significantly enhanced with the introduction of ER. The dehazing effect was evaluated based on four parameters of contrast (C₀), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM) and information entropy (E_n). It is further demonstrated the necessity of introducing ER. The experimental results show that the ER can make the dehazing effect of the orthogonal polarization dehazing technique improved, which is valuable for the engineering application of image processing in hazy environment. The experimental results proved to be encouraging.

HIGHLIGHTS

• The main highlights of this work are the following three points:

  The NDC parameter of the polarizer is a crucial factor that many researchers tend to overlook when utilizing polarizers for detection in the polarization dehazing process. This study aims to address this oversight by incorporating polarizers for detection and introducing the NDC parameter to rectify images at varying polarization angles. Consequently, the polarization dehazing operation yields superior results compared to those achieved without NDC polarization dehazing technology.

• The separation of the sky area from the non-sky area is achieved through the utilization of a multi-channel mask. This model employs multi-channel processing, specifically the HSV colour channel, RGB colour channel, and Lab colour channel. By generating an appropriate mask to delineate the sky area and the non-sky area, merging the three masks into a comprehensive mask, and subsequently applying morphological etching on the comprehensive mask to enhance the connectivity of the sky area, it is observed that the effectiveness of separating the sky area from the non-sky area is significantly improved.

• The amalgamation of image restoration and image enhancement is a critical process. Once the polarization dehazing operation is accomplished, we proceed with an image enhancement operation on the restored image. To regulate the filtering level, we employ the standard deviation of the Gaussian function. Furthermore, we utilize the Laplacian operator to intensify the edges of the current channel. By adjusting the parameters, we can control the intensity and scope of the enhancement effect, thereby achieving edge enhancement and augmenting the details of the image.

KEYWORDS:

• Polarization-based dehazing
• extinction ratio
• algorithm research
• image processing

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author agreement

The all authors certify that have seen and approved the final version of the manuscript being submitted. We warrant that the article is our original work, hasn’t received prior publication and isn’t under consideration for publication elsewhere.

Credit authorship contribution statement

Dongdong Shi: Funding acquisition, Project administration, Supervision, Conceptualization, Software, Writing-original draft, Writing-review & editing. Fuyu Huang: Methodology, Investigation. Leilei Jia, Yuandong Niu, Shuangyou Chen & Liting Jiao: Data curation. Yanhua Huang: Conceptualization, Resources, Visualization, Validation. Limin Liu: Supervision, Writing-review, Formal analysis.

Data statement

The data that support the findings of this study are available from the corresponding author, [Limin Liu], upon reasonable request.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [62171467]; Natural Science Foundation of Hebei Province [F2021506004].