Video stream noise reduction method, device and storage medium

Abstract

The present invention provides a video stream noise reduction method, device and storage medium. By acquiring the first image texture and the second image texture after noise reduction in the video stream, according to the current real-time filtering parameters of the filtering system, the first image texture and The second image texture performs linear interpolation calculation to output the noise-reduced first image texture, wherein the first image is an image currently to be processed in the video stream, and the second image is a previous frame image of the first image. Real-time denoising processing of the video stream image is realized by the above method, the denoising effect is obvious, and the problem of afterimage existing in the video stream is solved.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image processing, and in particular, to a video stream noise reduction method, device, and storage medium. Background technique [0002] Image and video denoising is to remove the noise introduced in the process of image and video acquisition and transmission, so as to improve image quality and facilitate subsequent processing. Therefore, image and video denoising is a very important link in the image processing process. [0003] In terms of video denoising, the existing technical solutions use an adaptive Kalman filter video denoising system. Video denoising itself is a video filtering process. The system uses a Kalman filter with a fixed variance parameter, but if the measurement variance is relative to If the prediction variance is small, the video noise reduction effect of the system is not obvious, and the image after filtering will have afterimages. In addition, existing video...

AI Technical Summary

Problems solved by technology

[0003] In terms of video denoising, the existing technical solutions use an adaptive Kalman filter video denoising system. Video denoising itself is a video filtering process. The system uses a Kalman filter with a fixed variance parameter, but if the measurement variance is relative to If the prediction variance is small, the video noise reduction effect of the system is not obvious, and the image after filtering will have afterimages

In addition, the existing video stream noise reduction solutions require high computational consumption, which cannot meet the real-time noise reduction processing of the mobile terminal.

Images