Adaptive processing method for foggy sky and white areas in UAV aerial images

Abstract

The invention discloses a self-adaptive processing method for foggy sky and white areas of UAV aerial images. Firstly, the foggy images to be processed are obtained; then dark channel images and grayscale images of the foggy images are obtained; according to the dark channel The image calculates the atmospheric light value A of the image; the dark channel image is used as a rough estimate of the atmospheric dissipation function, and the adaptive threshold ThrB that can segment the close-up and the sky or white area is obtained according to the gray image; then the image is processed in different regions, and the calculation Correct the coefficient and substitute it into the improved atmospheric dissipation function formula to obtain the improved atmospheric dissipation function value; then refine the obtained improved atmospheric dissipation function through bilateral filtering; then calculate the image according to the transmittance estimation formula Transmittance t(x); Finally, the haze-free image is restored according to the image degradation model.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a self-adaptive processing method for foggy sky and white areas of aerial photographed images of drones. Background technique [0002] At present, the hotspot in the field of UAV research is the application of UAV in surveying and mapping, target recognition, geological disaster prevention and other fields. The execution of UAV tasks largely depends on UAV images with high imaging quality. With the deterioration of weather conditions, the images obtained by aerial photography in foggy weather are blurred and the definition is not high. The basic information features of the images obtained by aerial photography are seriously distorted and damaged, and useful information cannot be captured. The demand for defogging is getting higher and higher. [0003] In recent years, great progress has been made in single-image dehazing methods, and image dehazing methods can be divide...

AI Technical Summary

Problems solved by technology

This method has an ideal defogging effect, but bright areas such as the sky or white do not satisfy the prior law of the dark primary color, resulting in image color distortion in these areas after defogging. Therefore, the color distortion of bright areas such as sky or white in the image is solved problem becomes an important task in the restoration of existing images

[0004] In view of the shortcomings of the current algorithm, the key parameters in the atmospheric scattering model are studied, and the bilateral filter or guided filter with better edge-preserving effect is used to accurately estimate the atmospheric dissipation function, retain the edge details of the image, and can It is too natural between the flat area and the halo effect or fog residue after the image is restored, but the problem of distortion in the bright area such as the sky or white is still not solved; some people use a fixed threshold to repair the transmittance and protect the sky fog image to a certain extent However, it is easy to cause insufficient defogging of the image without sky; some people also use segmented sky area to avoid the color distortion of the image after defogging, but because the maximum connected area is used as the identified sky area, it is easy to cause some sky parts. Missed detection of blocks, resulting in distorted color of the missed detection part of the sky area

These algorithms do not fundamentally solve the problem of defogging color distortion, so the algorithm for improving the effect of foggy sky or white areas needs to be improved

Images