Method and device for eliminating image noise

Abstract

The invention discloses a method for eliminating image noise, comprising the steps: a sub-block image to be processed generates a plurality of image blocks, adjacent image blocks contain image information of same position; transformation of a two-dimensional transform domain is conducted on the image blocks to obtain a coefficient matrix of the two-dimensional transform domain and form a three-dimensional array; transformation of a one-dimensional transform domain is conducted on the three-dimensional array to obtain a three-dimensional transform domain coefficient; threshold operation is conducted on the obtained three-dimensional transform domain coefficient, and then inverse transformation of a transform domain is conducted to obtain the processed image block information; and image information of the same position in the processed image blocks is averaged to obtain a first processed image. The invention also discloses a device for eliminating image noise. In the invention, pixel information of a plurality of adjacent image blocks contained in one image block is utilized and the pixel information of the same position is averaged, thus avoiding feature matching error caused by conducting feature matching on the image blocks, and being capable of effectively improving the effect of image noise elimination and the image quality.

Description

technical field [0001] The invention relates to image processing technology, in particular to a method and device for eliminating image noise. Background technique [0002] During the imaging process, due to the sensitivity and dynamic range of each photosensitive device in the imaging system, as well as the illumination changes in the actual environment and other factors, for example, the dark current and mechanical movement of the Charge Coupled Device (CCD) in the imaging system , the transmission circuit signal, and the transmission environment where the CCD is located, etc., make the imaged image always have different types and different intensities of noise. The noise in the image not only annihilates the information of the image, but also produces false information. The greater the noise intensity, the more image information is annihilated, which reduces the clarity and quality of the image and affects the application of the image. [0003] As a processing technology...

AI Technical Summary

Problems solved by technology

[0020] It can be seen from the above that in the prior art methods for eliminating image noise, the effect of eliminating image noise depends on the accuracy of feature matching of local image reference blocks, and when the noise intensity is high, the accuracy of feature matching is poor, which easily leads to matching Ineffective, thus affecting the effect of image noise removal, further, the threshold value operation is performed on the acquired first transformation coefficient to eliminate noise, the threshold value is preset by the user, and cannot reflect the actual environment of the noise in the image to be processed; while using noise estimation In this method, because the wavelet decomposition is directly performed on the noise-containing image to be processed, the high-frequency sub-bands obtained by the decomposition contain noise signals, which makes the accuracy of the estimated noise parameters poor, resulting in a decrease in image noise removal effect and poor image quality. poor

Images