Rain removing method for single image

Abstract

The invention discloses a rain removing method for a single image, and the method comprises the steps: firstly decomposing a rain image into a high-frequency image and a low-frequency image based on bilateral filtering; secondly introducing the incoherence of a dictionary at a dictionary learning state so as to reduce the similarity between atoms with rain and atoms without rain, and constructinga new target function, thereby guaranteeing the separability of a rain dictionary and a no-rain dictionary during clustering. Moreover, the learning incoherent dictionary has the attributes similar tothe attributes of a tight frame, and can approach to an equal-angle tight frame. The method achieves the sparse expression of the high-frequency image through the rain dictionary and the no-rain dictionary, can separate a rain component and a no-rain component in the high-frequency image in a better way, achieves the superposing of the no-rain component of the high-frequency image with the low-frequency image, and achieves the removing of rain of the image. An experiment result indicates that the incoherent dictionary learned through the method is better in sparse expression performances, there are fewer the residual rain lines on the image after rain removal, the method maintains the edge details in a better way, and the visual effect is clearer and more natural.

Description

Technical field [0001] The invention relates to the field of image processing, in particular to a method for removing rain from a single image. Background technique [0002] In rainy weather, raindrop particles are generally larger than 100μm and are easily captured by the lens. Therefore, affected by rainy weather conditions, part of the texture and detail information of the image collected by the outdoor lens is easily obscured by the rain, causing problems such as excessively bright local areas and blurred background images. The degradation of image quality in rainy days greatly restricts the functions of outdoor intelligent vision systems such as visual monitoring, visual navigation and target tracking. In addition, the state of raindrop particles is changeable, and the direction and thickness of the rainline under different conditions are different. Therefore, studying how to recover high-quality images from various rain-degraded images has extremely high research and appli...

AI Technical Summary

Problems solved by technology

Based on literature [15], Huang et al. [16] The affine propagation method is used to realize the clustering of dictionary atoms. Each type of dictionary atom contains similar edge and texture information, and the high-frequency no-rain component is obtained by using the large variance of the no-rain atom class, and then superimposed with the low-frequency image to realize image rain removal. , this method improves the image edge preservation, but there are some rain lines left, and the dictionary clustering method leads to poor real-time performance of the algorithm

Luo et al. [17] Using the color filtering principle in PS (Phtoshop), a nonlinear rain image model is proposed, and the image is derained by using discriminative sparse coding. This method is realized by constraining the correlation between the rain-free image and the sparse representation coefficient of the rain component. The image removes the rain, the phenomenon of residual rain lines is serious, and the rain is not completely removed

[0005] Image deraining based on dictionary learning and sparse representation has achieved certain results, but there are still the following problems: (1) There is overlap between the rain line and the background image texture in the rainy image, and the dictionary obtained by the existing dictionary learning method The similarity between atoms is high

The orthogonal base coherence is equal to 0, but the orthogonal base cannot guarantee the sparsity of the sample representation

Images