A comparative study of single image fog removal methods

  • PDF / 3,219,839 Bytes
  • 17 Pages / 595.276 x 790.866 pts Page_size
  • 0 Downloads / 255 Views

DOWNLOAD

REPORT


ORIGINAL ARTICLE

A comparative study of single image fog removal methods Bijaylaxmi Das1 · Joshua Peter Ebenezer1,2 · Sudipta Mukhopadhyay1 Accepted: 26 October 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract The presence of fog degrades visibility in natural scene conditions. Computer vision applications like navigation, tracking, and surveillance need clear atmospheric images or videos as prerequisites for optimal performance. However, foggy atmosphere creates problems for computer vision applications due to reduced visibility. Different fog removal techniques are used to improve the visual quality of images and videos. The fog density depends on the depth information. Scene depth information estimation needs multiple images, which limits its real-life application. Hence, a single image fog removal requires some prior knowledge and/or assumptions to get the depth information. In this paper, the recent fog removal techniques are grouped into three broad categories: (1) filter-based methods, (2) color correction based methods, and (3) learning-based methods, for ease of understanding. The primary objective is to provide an introduction to this field and compare performance (both qualitative and quantitative) of representative techniques for each category. It is found that filter-based methods are doing overall better compared to other categories. Keywords Fog removal · Image restoration · Transmission map · Color correction · Contrast enhancement · Deep learning

1 Introduction Fog and haze cause visibility reduction leading to accidents. According to the Federal Highway Administration in US,1 in the year 2007–2016, an average of 8902 persons were injured, and 464 persons died in 451 crashes due to fog. Similarly, according to The Times of India,2 in India, 9317 people died due to fog-related crashes in the year 2016. The death toll increased to 11,090 in 2017, which is a jump of almost 20%. Fog removal algorithms are needed to assist drivers in reducing fog-related risks. Outdoor scenes are usually degraded 1 2

https://ops.fhwa.dot.gov/weather/q1_roadimpact.htm. https://timesofindia.indiatimes.com/india/over-10000-lives-lost-infog-related-road-crashes/articleshow/67391588.cms.

B

Sudipta Mukhopadhyay [email protected] Bijaylaxmi Das [email protected] Joshua Peter Ebenezer [email protected]

1

Electronics and Electrical Communication Engineering Department, Indian Institute of Technology Kharagpur, Kharagpur, India

2

The University of Texas, Austin, USA

by the presence of dust particles and water droplets which cause atmospheric absorption and scattering of light. The bad weather can be due to fog, haze, mist, and/or rain. Fog and haze are created by suspended water droplets of size 1–10 µm and 10−2 –1 µm, respectively. Fog and haze belong to the steady bad weather category. The presence of fog and haze attenuates the radiance of the objects received by the camera from the scene point. It causes a reduction in visibility, increasing travel time, and increasing the