Method and device for denoising video image

Abstract

The embodiment of the invention discloses a method and a device for denoising a video image. The method and the device relate to image processing, remove noise of the image to a greater degree, and effectively avoid an image tail phenomenon of a moving region. The concrete scheme comprises the following steps: taking a current processing pixel point as a center, and carrying out intra-frame filtering on all pixel points within an N*N range in a current processing frame to obtain a second pixel block; taking a first processing pixel point as a center, and carrying out intra-frame filtering on all pixel points within an N*N range in a reference frame to obtain a fourth pixel block; acquiring a weighted value of time-domain filtering of the current processing pixel point according to the second pixel block and the fourth pixel block; acquiring a time-domain filtering component of the current processing pixel point according to the weighted value; carrying out intra-frame filtering on the current processing pixel point to obtain a spatial filtering component; and updating a gray value of the current processing pixel point according to the time-domain filtering component, the spatial filtering component and the weighted value. The method and the device are applied in a video image denoising process.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method and device for noise reduction of video images. Background technique [0002] With the continuous development of science and technology, digital cameras have been widely used in people's daily life, so how to restore the original image without distortion as much as possible through image noise reduction technology has become a key topic of research in this field. [0003] In the process of video acquisition, there will be various noises mixed in the acquired image sequence, most of which are random, it is understandable that the correlation between images in the image sequence acquired by digital camera is very High, so the noise of the image sequence can be reduced by time-weighted averaging of the image. This method can remove the random noise between frames very well, but it is simply superimposed and averaged on the image in the time domain. It will blur the area whe...

AI Technical Summary

Problems solved by technology

[0004] In the process of realizing the above-mentioned filtering, the inventors found at least the following problem in the prior art: the noise of the current processing frame obtained according to the equipment gain or the variance of the smooth area in the current processing frame is not accurate, so the calculated current processing frame There will also be a certain error in the distance between the block in the reference frame and the block in the corresponding position in the reference frame, which will lead to errors in the weights of time domain filtering and spatial domain filtering obtained according to the distance, so that when filtering the image During the process, the noise removal is not clean or the smear of the moving area occurs

Images