
Research Article
Performance Evaluation of Multiwavelet Transform for Single Image Dehazing
@INPROCEEDINGS{10.1007/978-3-031-28975-0_10, author={M. Ravi Sankar and P. Srinivas and V. Praveena and D. Bhavani and M. Sri Uma Suseela and Y. Srinivas and Ch. Venkateswara Rao}, title={Performance Evaluation of Multiwavelet Transform for Single Image Dehazing}, proceedings={Cognitive Computing and Cyber Physical Systems. Third EAI International Conference, IC4S 2022, Virtual Event, November 26-27, 2022, Proceedings}, proceedings_a={IC4S}, year={2023}, month={3}, keywords={Computational Time Computer Vision Dark Channel Prior Dehazing Image Dimension Image Quality Object Detection Transmission map Visibility Wavelet Transform}, doi={10.1007/978-3-031-28975-0_10} }
- M. Ravi Sankar
P. Srinivas
V. Praveena
D. Bhavani
M. Sri Uma Suseela
Y. Srinivas
Ch. Venkateswara Rao
Year: 2023
Performance Evaluation of Multiwavelet Transform for Single Image Dehazing
IC4S
Springer
DOI: 10.1007/978-3-031-28975-0_10
Abstract
Images captured in poor lighting conditions (such haze, fog, mist, or smog) have a lower level of visibility because air particles deflect light. Single picture dehazing techniques can restore clarity to a single hazy image. Even though Dark channel prior (DCP) has been the most used method for image dehazing algorithms, it has poor picture quality and requires a lot of adjustments for real time applications such as; computer vision, object detection. However, previous studies such as; Dark Channel Prior (DCP) on image dehazing agonises from a gigantic processing time and sleaze of the original image. In addition, the dimensions of the images have a significant effect on the performance of the dehazing algorithms. Hence, the main motive of this work is to minimize the computational overload either by reducing the size of the image or by accurate transmission map estimation without compromising the image quality. Hence, this work is focused on reducing the computational time for image dehazing through multilevel discrete (Haar) wavelet transform, in which the image dimensions has been reduced without degrading quality. The performance measures of proposed algorithm have been analysed in terms of PSNR, MMSE and SSIM and compared with existing DCP algorithm. The simulation results proven that, the computational time for proposed algorithm has been reduced by 90% when compared to DCP without degrading image quality.