An Image Dehazing Algorithm Based on Dark Channel Prior

Abstract

The invention discloses an image defogging method based on dark primary color prior, which includes the following steps: Step 1: Find the dark primary color map of the foggy degraded image; Step 2: Find the rough transmission of the image from the dark primary color map rate; Step 3: Use an improved tolerance mechanism transmittance adjustment algorithm based on color attenuation prior to amplify and correct the transmittance of bright areas such as the sky, and obtain the corrected transmittance. Step 4: Use guided filtering to refine the transmittance; Step 5: Combine the halo operator and the dark primary color image to obtain the atmospheric light value; Step 6: Use the atmospheric scattering model to obtain the restored image J. The defogging algorithm designed by the present invention can effectively solve the problem of color cast in bright areas such as the sky in the restored image that exists in the traditional defogging algorithm. At the same time, it can also solve the problem that the traditional atmospheric light value calculation method is susceptible to interference by white objects, and improve the image quality. .

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image defogging algorithm based on dark channel prior. Background technique [0002] In the case of haze weather, the images acquired by the vision system will be affected by the scattering of visible light by haze particles, resulting in degradation phenomena such as low contrast and blurred details, which will affect the normal operation of the vision system. Therefore, image restoration in foggy weather is an urgent research work. At present, the image defogging algorithm mainly has two directions: model-based dehazing and non-model-based dehazing. The essence of non-model-based dehazing is to enhance image contrast, but this type of method may lose image information and cause image distortion, which is represented by the Retinex algorithm. The essence of the model-based dehazing method is to construct a physical model of the foggy image, and reverse the degradation proces...

AI Technical Summary

Problems solved by technology

Based on the prior knowledge that the brightness and saturation of foggy image pixels change sharply with the change of haze concentration, Zhu et al. constructed a linear model to obtain the depth information of the image, so as to realize image defogging, but the algorithm involved many parameters. , poor adaptability

Wang et al. proposed to adjust the transmittance of the sky region by identifying and extracting the sky region by setting the gradient and brightness threshold according to the characteristics of the sky region. However, this algorithm uses the largest connected region as the sky for recognition, which is time-consuming and easy. Missing part of the sky, resulting in color distortion in the missing area

Jiang Jianguo introduced a tolerance mechanism to judge bright areas such as the sky, and improved the transmittance of the area, which better solved the problem of color distortion in the sky area and became a mainstream solution. The difference parameter selection is affected by whether there is a sky area in the image, and the transmittance after correction will be insufficiently enhanced or overcompensated

Images