Improved image dehazing model with color correction transform-based dark channel prior

The precise identification and positioning of objects require the restoration of hazy images. This paper proposes the implementation of Color Correction Transform based Dark Channel Prior (CCTDCP), a new prior-based dehazing technique. The CCTDCP strategy is intended to consider the brightening impact of haze and blue shading cast of hazy pictures. The proposed technique has introduced white balance color correction transform in order to preserve the originality of the image. This transform is responsible for generating properly color-corrected hazy images employing Euclidean norm and radial basis function. An integration of dark channel prior and gamma correction techniques is incorporated for effectual restoration. The experimental outcomes on benchmark datasets demonstrate superior PSNR and SSIM performance, better visual quality, and lesser execution time of the CCTDCP algorithm over the state-of-the-art techniques. The highest performance achieved using the suggested CCTDCP strategy has a PSNR of 28.3798 dB and SSIM value of 0.8317 on the O-Haze and I-Haze datasets, respectively.