A Fusion Dehazing Method Based on Variable Step-Size Guided Filtering and Contrast Stretching

Abstract

The invention discloses a fusion defogging method based on variable step-length guided filtering and contrast stretching. The steps include: firstly downsampling and compressing the image, then performing the defogging operation, and finally using bilinear difference Up-sampling and zooming in; the defogging operations include: RGB value conversion operations, linear transformation operations, high-intensity filtering operations for large windows, and white balance adjustment; linear transformation operations include: grayscale transformation, grayscale image traversal to obtain the maximum value and minimum value value, traversal division; large-window supercharged filtering operations include: guided filtering and traversal multiply-accumulate. The invention can optimize the defogging steps and increase the defogging speed under the premise of basically not affecting the defogging quality of the image.

Description

technical field [0001] The invention belongs to the field of image defogging in image processing, in particular to a guided filter with variable step size and an image fusion defogging method using the guided filter. Background technique [0002] With the rapid development of society, people's various requirements for material life continue to increase, and the number of various motor vehicles continues to increase, resulting in a large amount of pollution. At the same time, the waste gas produced by heavy industry cannot be well treated, which also leads to increasing air pollution. It has become an extremely serious environmental problem. [0003] In the field of aviation, smog will seriously affect the pilot's vision, which makes the aircraft have great risks whether it is taking off or landing, and during the flight, the smog will cause the pilot's perceived distance to deviate from the actual distance, increasing the The possibility of accidents; in the field of video ...

AI Technical Summary

Problems solved by technology

[0005] 1. When processing full HD images, 2K images, and 4K images, the time taken for traversal in the linear stretching process is too long, and too many filtering windows are required in the filtering process;

[0006] 2. It is necessary to traverse and filter the maximum and minimum values ​​of the R, G, and B channels, and then traverse and divide each channel, which is too cumbersome;

[0007] 3. The guided filtering method with a step size of 1 cannot meet the real-time requirements of high frame rate processing scenes

Images