Cascade filter and dynamic setting method for calibrated denoising intensity thereof

Abstract

The present invention relates to the field of video processing. Disclosed are a cascade filter and a method for dynamically setting the calibrated denoising strength thereof. In the present invention, the denoising capability of a preceding cascade filter is estimated, an estimated value is subtracted from the calibrated denoising strength of the preceding cascade filter to acquire a difference, and the calibrated denoising strength of a subsequent cascade filter is set on the basis of the difference. This allows for self-adapting adjustment of the denoising strength of the subsequent cascade filter for different input images, thereby balancing the denoising strengths of the cascaded cascade filters to acquire improved overall denoising effect. Preferably, the cascade filter is provided with a time-domain filter and with a space-domain filter.

Description

technical field [0001] The invention relates to the field of video processing, in particular to a dynamic setting technique for cascading filter calibration denoising strength. Background technique [0002] In specific application scenarios such as video surveillance, the actual collected images often contain a lot of noise. In order to facilitate the application or post-processing of the image, filters are usually used to filter the image to reduce noise. [0003] There are many types of filters, such as time-domain filters, spatial-domain filters, time-space domain filters, and so on. However, no matter which type it is, the filtering effect of a single filter is often limited, so some people use cascaded filters to filter images. [0004] A cascaded filter refers to a system in which multiple filters are cascaded together. In the cascade filter, the image will be filtered in multiple filters successively, which can usually achieve a better filtering effect than a single...

AI Technical Summary

Problems solved by technology

[0006] However, in practical application scenarios, the actual noise level of the image is often unknown, or is changing, and the type of noise is also changing, or different areas of the image have different noise levels, so the existing technology usually Set a fixed calibration denoising strength for each filter in the cascade filter. In this way, the multi-stage filter adopts the same noise level, so that some areas of the denoising result will be too smooth, and some areas will be insufficiently denoised. Poor denoising effect

Images