Noise reduction method after histogram equalization of low-illumination image

Abstract

The invention discloses a noise reduction method after histogram equalization of a low-illumination image. The method takes a plurality of minimum brightness values as a reserved set and traverses all pixels, and comprises the following three situations of 1: for the pixel brightness values belonging to the reserved set, keeping the pixel brightness values unchanged; 2, for the pixel brightness values which do not belong to the reserved set, if the brightness values of eight pixels in a 5*5 area around the pixel belong to the reserved set, valuing the pixel brightness in the reserved set; and 3, for other situations, according to the pixel characteristics in the 5*5 area around the pixel, adaptively constructing a 5*5 filtering template, and calculating the brightness value of the pixel by using the filtering template. Compared with the noise reduction via a Gaussian low-pass filter, the method has the advantage that the information in a dark environment can be visually identified better.

Description

technical field [0001] The invention relates to a noise reduction method after histogram equalization of a low-illuminance image. Background technique [0002] Digital image acquisition is to use sensors such as cameras to convert optical signals into electrical signals and store, transmit and display them in digital form. Digital image processing is to optimize the collected digital images according to the purpose and scene of use. The common methods include: image Enhancement and restoration, image coding compression, image description, etc. [0003] Low-illumination images refer to images collected under low ambient light conditions. In order to be able to display clearly, it is necessary to enhance the brightness and reduce noise of the low-illumination images, so that the naked eye can obtain useful information in the images. The brightness enhancement of low-illumination images is playing an increasingly important role in practical applications. In the field of outdoo...

AI Technical Summary

Problems solved by technology

[0007] Although the histogram equalization algorithm is suitable for brightness enhancement of low-illumination pictures, and the Gaussian low-pass filter is suitable for noise reduction of ordinary pictures, the effect of Gaussian low-pass filter noise reduction on the image after the histogram equalization algorithm is not good. Because in the image after histogram equalization, low-brightness information is missing, high-brightness noise is amplified or crossed, and the image and noise are sparsely distributed and mixed together. At this time, using a Gaussian low-pass filter will cause image blur

Images