Image demisting method based on sky segmentation and transmissivity adaptive correction

Abstract

The invention discloses an image demisting method based on sky segmentation and transmissivity adaptive correction, an iterative threshold segmentation algorithm is used to divide the fogged image into sky region and non-sky region, and the brightness of the segmented sky region is adjusted. A dark channel prior algorithm based on fast bilateral filtering and cubic logarithmic polynomial correctedtransmittance is used to de-fogging the non-sky region. At the same time, the average value of high luminance pixel set in dark channel image is used to ensure that the estimation result of atmospheric light value tends to reasonable value as far as possible. The output image is further enhanced by fast bilateral filter and cubic logarithmic polynomial transform of transmittance, so as to achievethe purpose of denoising, preserving edge and eliminating the edge halo of the scene. The experiment results show that the invention has the advantages of good demisting effect, high color fidelity of the sky region and strong edge retention ability.

Description

technical field [0001] The invention relates to the field of outdoor monitoring image processing, in particular to an outdoor monitoring image defogging method based on sky segmentation and adaptive correction of transmittance, which has good defogging effect, high color fidelity in the sky area, and strong edge retention ability. Background technique [0002] In recent years, the application range of video surveillance systems has become more and more extensive, such as road monitoring, vehicle monitoring, public safety monitoring, etc. For video surveillance systems in indoor scenes, it can capture high-definition video images in most cases; for video surveillance systems in outdoor scenes, the shooting quality depends largely on the weather environment. Under fog, haze and other weather conditions, a large number of tiny water droplets and aerosols suspended in the atmosphere have a scattering effect, resulting in color distortion, contrast and visibility reduction of ima...

AI Technical Summary

Problems solved by technology

However, the transmittance map refined by guided filtering is blurry, and the edges of foreground objects are not sharp enough; and, because the dark channel value in some areas is much greater than 0, which is close to the atmospheric light value, so that using the atmospheric light curtain guided map and The transmittance calculated by the guided filter has a large deviation, and the color of the sky area is obviously distorted

Images