Adaptive image defogging method based on textures

Abstract

The invention discloses an adaptive image defogging method based on textures and belongs to the technical field of imaging processing and computer vision. Corresponding scene transmission coefficients are calculated through carrying out texture classification on an input image and combining texture classification results with dark channel values with different scales to realize quick image defogging based on texture adaption. The adaptive image defogging method has the beneficial effects that the complicated soft cutout optimization step can be avoided, the algorithm complexity of defogging by using an He method is reduced, the Halo phenomenon easily caused after defogging is avoided, and the saturation of a defogged image can be better controlled; meanwhile, the calculation on an atmospheric light vector is corrected to ensure the brightness of the defogged image; and the adaptive image defogging method is high in processing speed and high in defogged image quality and can satisfy the demands of real-time processing application.

Description

technical field [0001] The invention belongs to the technical field of image processing and computer vision, in particular to a texture-based adaptive image defogging method. Background technique [0002] Foggy image restoration is a key technology in the technical fields of image processing and computer vision. Because the classic computer vision algorithms often assume that the input image is the reflected light on the surface of the object, for some foggy images, the mathematical models of these methods will inevitably fail to a certain extent, thus affecting the final image defogging effect. Therefore, it is of great significance to study the imaging degradation model and dehazing method of foggy images. [0003] For a single foggy image, the existing image defogging methods are usually based on the prior information rules of the foggy image. Typical methods are Tan method (local contrast maximization method), Fattal method (scene albedo and transmittance estimation me...

AI Technical Summary

Problems solved by technology

Although the dehazing quality of the image recovered by the He method is relatively ideal, only one scale parameter is considered, and the calculation is extremely complex and the soft matting optimization method is used to estimate the transmission coefficient of the scene, and the calculation amount will appear as the image size increases. The geometric level increases, and it is difficult to achieve real-time processing of fog image removal on the basis of existing hardware, and it is difficult to apply it to practical applications

Images