Deep Learning Techniques for Automation and Industrial Applications

6Removal of Haze from Synthetic and Real Scenes Using Deep Learning and Other AI Techniques

Pushpa Koranga^1*, Ravindra Singh Koranga², Sumitra Singar¹ and Sandeep Gupta³

¹Faculty of Computing Skills Education, Bhartiya Skill Development University, Jaipur, Rajasthan, India

²Department of Computer Science and Engineering Graphic Era Hill University, Bhimtal, Uttarakhand, India

³Department of Electronics and Communication Engineering, Sunstone (JECRC), Jaipur, Rajasthan, India

Abstract

Images are the main source for all image-processing fields like surveillance, detection, recognition, and satellite. Good visibility of images captured by sensors becomes crucial for all computer vision tasks. Sometimes, the scene quality is degraded by bad weather conditions like haze, fog or smoke; therefore, making it difficult for the computer vision area to obtain actual information. Haze can be removed from a single input scene by using single image dehazing methods. Synthetic hazy images are created by a haze generator. Currently, most image dehazing techniques are applied for synthetic haze. Various single-image dehazing techniques are being developed and tested on real-world scenes captured in hazy environments using cameras. These techniques aim to be practical solutions for removing haze from images. This study focuses on dehazing methods for both synthetic and real datasets totaling 45 hazy scenes. The output qualities of different techniques are measured using different parameters, ...