International Journal on Advanced Science, Engineering and Information Technology, Vol. 6 (2016) No. 6, pages: 1081-1088, DOI:10.18517/ijaseit.6.6.1469

A Study of Atmospheric Particles Removal in a Low Visibility Outdoor Single Image

Muhamad Lazim Bin Talib, Mohammad Faidzul Nasirudin, Siti Norul Huda Sheikh Abdullah

Abstract

Maximum limit of human visibility without the assistance of equipment is 1000 m based on International Commission on Illumination. The use of camera in the outdoor for the purpose of navigation, monitoring, remote sensing and robotic movement sometimes may yield images that are interrupted by haze, fog, smoke, steam and water drops. Fog is the random movement of water drops in the air that normally exists in the early morning. This disorder causes a differential image observed experiences low contrast, obscure, and difficult to identify targets. Analysis of the interference image can restore damaged image as a result of obstacles from atmospheric particles or drops of water during image observation. Generally, images with atmospheric particles contain a homogeneous texture like brightness and a heterogeneous texture which is the object that exists in the atmosphere. Pre-processing method based on the dark channel prior statistical measure of contrast vision and prior knowledge, still produces good image quality but less effective to overcome Halo problem or ring light, and strong lighting. This study aims to propel the development of machine vision industry aimed at navigation or monitoring for ground transportation, air or sea.

Keywords:

atmospheric particles removal; single image; haze removal; fog removal; image enhancement.

cite this paper     download