Remote sensing image automatic haze removal method and system based on marked dark channels

Abstract

The invention provides a remote sensing image automatic haze removal method based on a marked dark channel, and the method comprises the steps: obtaining the dark channel of any pixel in the remote sensing image, and setting a mark for a waveband corresponding to the dark channel; layering all the dark channels according to wave bands, and taking a layer corresponding to the wave band with the maximum number of the marks as a starting wave band layer; obtaining a proportional relation between the initial waveband layer and other layers according to the number of marks of each layer; convertingthe dark channels of the other layers to the initial waveband layer according to the proportional relation so as to obtain an initial haze background image corresponding to the initial waveband layer; converting the initial haze background image into other haze background images corresponding to the other layers according to the proportional relation; and removing the initial haze background image and the other haze background images from the remote sensing image to obtain a result image. The invention further provides a remote sensing image automatic haze removal system based on the marked dark channel, and a data processing device for carrying out automatic haze removal processing based on the marked dark channel on the remote sensing image.

Description

technical field [0001] The invention relates to the field of remote sensing image processing and application, in particular to a remote sensing image haze removal method and image analysis application. Background technique [0002] Remote sensing technology is often affected by weather conditions and atmospheric conditions, and it is difficult for satellite sensors to obtain high-quality images under completely clear sky conditions. Atmospheric pollutants, aerosols, thin clouds, and haze near the ground absorb, refract, reflect, and scatter electromagnetic waves, resulting in deviations in the reflected radiation detected by satellite sensors, especially in the visible light band. The contrast, color and other characteristics of the image acquired under haze or cloudy conditions will change, which will make the visual effect of the image worse and reduce the viewing value, which will affect the post-processing of the image and the interpretation and interpretation of image i...

AI Technical Summary

Problems solved by technology

However, there are still some aspects of these algorithms that need further research and improvement: 1) The complexity of the haze removal algorithm needs to be reduced, some algorithms involve complex physical transmission processes or mathematical operations, and it is difficult for non-professionals to use them; 2) Algorithm automation It needs to be improved. There are too many key parameters in the current algorithm that need to be manually adjusted by the user to obtain the optimal output, which is not conducive to automation and batch processing; 3) The haze removal process will lead to the loss of original image information and changes in spectral characteristics to a certain extent. Most algorithms and application systems lack effective result analysis and objective evaluation

Images