Haze Removal Algorithm to Enhance Low Light Images
Smt. K. Maheswari, Associate professor Dept. of ECE, Sanskrithi School of Engineering
K. Charan, K. Annamayya, K. Naveen, K. Manoj Kumar
Dept of ECE, Sanskrithi School of engineering , Puttaparthi 515134
Abstract:
Image captured in hazy, visually degraded visibility caused by foggy atmospheric conditions; it obscures image quality. Instead of producing clear images, by the pixel-based metrics does not guarantee, this updated image is used as input in computer vision of low-level tasks like segmentation. To upgrade this, for image dehazing we are introducing a novel called end-to-end approach, so maintain the generated images visual quality. So here we take a step further to explore possibility of exploiting a network to perform semantic segmentation technique with U-Net. To further improve the perceptual quality of output U-Net going to designed and used in this model.
Keywords: Dehaze, Image Segmentation, U-Net
Introduction:
Haze is formed due to is very fine, widely dispersed particles in atmosphere, in relatively dry air, it gives air an opalescent appearance. Haze is traditionally an atmospheric phenomenon in which smoke, dry particles and other dust particles obscure sky clarity. The World Meteorological Organization manual codes have a classification of horizontal obscuration into categories of fog, ice fog, steam fog, mist, haze, smoke, volcanic ash, dust, sand, and snow. Haze particles having sources like traffic, wildfire, dry weather, industry. For example, an approaching airplane and haze can be appeared brownish or bluish depending on the direction of view with respect to the Sun, while mist become a bluish grey. Haze is formed by dry air and mist is formed by humid air. However, mist droplets are formed by the subsequent of the condensation nuclei(haze particles), such produces of haze are known as “wet haze”.
In general, haze is just like a fog in nature when we capture images in this situation. Image losses its actual visibility, changing colors and obscure scenes. Ultimately it degrades the image quality. It is because of scattering and absorption of small particle in atmosphere. This we can refer as “Atmospheric Scattering Model”. This haze is an undesired and problematic to photography and also some applications like Object detection, scene understanding, and surveillance etc. To get better understanding and identify objects in image has to be done. For this haze must be removed from the image, it is not a simple task in computer vision.
haze-removal-algorithm-to-enhance-low-light-images
FREE DOWNLOAD
@ engpaper.com published paper
PUBLICATION PROCEDURE WITH US ENGPAPER.COM