An airlight estimation method for image dehazing based on gray projection
- PDF / 1,802,381 Bytes
- 19 Pages / 439.37 x 666.142 pts Page_size
- 85 Downloads / 193 Views
An airlight estimation method for image dehazing based on gray projection Wencheng Wang 1 & Xiaohui Yuan 2 & Xiaojin Wu 1 & Yihua Dong 1 Received: 27 September 2019 / Revised: 16 June 2020 / Accepted: 16 July 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
As the key parameter of dehazing algorithms, airlight value directly affect the calculation accuracy of sky region, and any deviation will lead to the chromatic aberration in the image restoration. Many methods are proposed to address this problem, but the large amount of calculation or large deviations make them difficult to apply to realtime systems. In this paper, a fast algorithm is proposed based on the statistics of sky area’s distribution in hazy images. Firstly, fast mean filter is used to process gray image; and then the anti-interference ability of regional projection is analysied. Through the horizontal projection and vertical projection, the main sky area is quickly located, and finally the sky region are calculated by selecting some special pixels as atmospheric light. A large number of experiments show that the proposed algorithm can obtain the airlight value quickly for the images with sky region, and can be used in real-time conditions. Keywords Airlight estimation . Atmospheric light . Gray projection . Image restoration
1 Introduction Hazy weather is a natural phenomenon caused by the scattering of atmospheric particles in the air. This type of weather often causes a significant decrease in the contrast and visibility of collected images, which affects the performance of outdoor visual systems. The use of information processing techniques for image dehazing can effectively resolve image quality degradation, raise the visibility of observation scenes, and improve the robustness of outdoor systems in target recognition and tracking. At present, the processing methods for hazy images
* Wencheng Wang [email protected]
1
College of Information and Control Engineering, Weifang University, Weifang, China
2
Department of Computer Science and Engineering, University of North Texas, Denton, TX, USA
Multimedia Tools and Applications
are usually divided into two classes: enhancement-based methods and restoration-based methods. Image enhancement-based methods improve the visual effect of images by improving the contrast of degraded images, which mainly include the histogram equalization method [18], Retinex method [5], homomorphic filtering method [33], wavelet transform method [6], etc.. Although these algorithms can visually improve the quality of an image, they do not consider image degradation factors, which often cause color distortion in recovered images. Image restoration-based methods usually compensate for missing information of hazy image via an inversion operation by establishing a degradation model for the physical process. This kind methods have high directivity with the natural effect of a dehazed image, attracted attention from many researchers in recent years, and has achieved productive results [1,
Data Loading...
