A Novel Haze Line Prior Haze Image Restoration Algorithm with Refine Transmission Map

With a focus on the problems of contrast degradation and color shift in image dehazing algorithms, we propose a line-ahead haze image restoration algorithm with a new refined transmission map and theoretically analyze its refined transmission map model. First, a compensation term for the illumination-sensitive transmission map is introduced to compensate for the effect of image luminance degradation caused by dense haze in the scene. Next, an optimization algorithm is introduced to regularize the transmission map and remove the effect of noise on the estimation results. Finally, the image is restored in combination with an atmospheric scattering model. Subjective and objective experiments show that the proposed algorithm not only has a high subjective dehazing effect, adequate brightness and high color fidelity, but also performs well on objective performance measures under different haze concentrations.