Zobrazit v EDS

AN EFFECTIVE SINGLE IMAGE DEHAZING METHOD TO ENHANCE IMAGE VISIBILITY

Uloženo v:
Podrobná bibliografie
Název: AN EFFECTIVE SINGLE IMAGE DEHAZING METHOD TO ENHANCE IMAGE VISIBILITY
Autoři: Ahmed, Md. Toukir, Bappy, Sabbir Hossain, Tuly, Jinnurain Khanom, Billah, Tazrian, Bari, S.M.Sadakatul
Zdroj: Journal of Information Technology; Vol. 12 No. 1 (2023); 1-12
Informace o vydavateli: Jahangirnagar University
Rok vydání: 2023
Témata: Haze Removal, Histogram Equalization, Airlight Co-efficient Map
Popis: Foggy or hazy weather conditions, including distance, environmental particle density, wavelength, scattering, and attenuation of air particles, reduce contrast, brightness, and visibility, alter color, and produces unrealistic color grading, which causes issues with image quality. Low-quality image is a major issue in image processing and computer vision, compromising the detection of an object from images. So, image fog or haze removal is one of the most basic and important tasks in digital image processing. We aim to remove fog or haze from digital images. The study evaluated four methods, including the Dark Channel Prior Based method, which showed good results for gray and color images. However, it only processed dehaze images if some portion showed the sky region and had high time complexity. This study provided a method for dehazing and enhancing dehazed images to address these drawbacks. In The method, we first calculated the histogram equalized image, then the minimum channel image, then calculated the transmission map using fog level and, as post-processing, applied a guided filter and the PIL image improvement method, and finally, dehazed an image. It was created for RGB pictures without user input or prior knowledge and is faster and lightweight for real-time applications. The method produced enhanced images with higher contrast, brightness, and high-quality color grading, making it useful in traffic surveillance systems. The implementation was a procedural structured programming model, testing only one image at a time using a 0.9 airlight coefficient.
Druh dokumentu: article in journal/newspaper
Popis souboru: application/pdf
Jazyk: English
Relation: https://journals.juniv.edu/index.php/jit/article/view/37/17; https://journals.juniv.edu/index.php/jit/article/view/37
Dostupnost: https://journals.juniv.edu/index.php/jit/article/view/37
Rights: Copyright (c) 2024 Journal of Information Technology
Přístupové číslo: edsbas.6EEB9FF2
Databáze: BASE
Popis
Abstrakt:Foggy or hazy weather conditions, including distance, environmental particle density, wavelength, scattering, and attenuation of air particles, reduce contrast, brightness, and visibility, alter color, and produces unrealistic color grading, which causes issues with image quality. Low-quality image is a major issue in image processing and computer vision, compromising the detection of an object from images. So, image fog or haze removal is one of the most basic and important tasks in digital image processing. We aim to remove fog or haze from digital images. The study evaluated four methods, including the Dark Channel Prior Based method, which showed good results for gray and color images. However, it only processed dehaze images if some portion showed the sky region and had high time complexity. This study provided a method for dehazing and enhancing dehazed images to address these drawbacks. In The method, we first calculated the histogram equalized image, then the minimum channel image, then calculated the transmission map using fog level and, as post-processing, applied a guided filter and the PIL image improvement method, and finally, dehazed an image. It was created for RGB pictures without user input or prior knowledge and is faster and lightweight for real-time applications. The method produced enhanced images with higher contrast, brightness, and high-quality color grading, making it useful in traffic surveillance systems. The implementation was a procedural structured programming model, testing only one image at a time using a 0.9 airlight coefficient.