Low-light image enhancement method and device based on optimized Bezier curve and equipment

Abstract

The invention discloses a low-light image enhancement method based on an optimized Bezier curve. The method comprises the following steps: inputting a low-light image to be processed into a Bezier curve model; setting an enhancement parameter of the low-light image, and optimizing the enhancement parameter based on the whale algorithm model; calculating a weighted probability distribution function of the low-light image according to the optimized enhancement parameter; calculating a weighted cumulative distribution function of the low-light image according to the weighted probability distribution function; calculating a weighted average exposure value of the low-light image according to the weighted cumulative distribution function; and optimizing the Bezier curve model according to the weighted cumulative distribution function and the weighted average exposure value to obtain an enhanced image. According to the method, the enhancement parameters of the low-light image are optimized through the whale algorithm model, the enhancement parameters are controlled at appropriate positions, excessive enhancement and insufficient enhancement can be avoided, strong noise in the low-light scene can be reduced through the over-weighted average exposure value, and the enhanced image with rich details and better quality is obtained.

Description

Background technique [0001] The invention belongs to the technical field of image processing, and in particular relates to a low-light image enhancement method, device and equipment based on optimized Bezier curves. [0002] For images acquired under low-light conditions, digital imaging systems usually have problems with low visibility, poor contrast, and color distortion. decline. In computer vision systems and video surveillance systems, it is particularly important to solve low-light image enhancement for low-light images. [0003] In the prior art, for the enhancement of low-light images, there are mainly the following methods: [0004] 1) Use histogram equalization to enhance images in low-light scenes. For example, the contrast-limited adaptive histogram equalization method, the weighted threshold histogram equalization method, the exposure-based sub-image histogram equalization method and the edge-based texture histogram equalization method have all been used for ...

AI Technical Summary

Problems solved by technology

For example, deep autoencoder-based methods are used to enhance naturally degraded and low-light images, however, such methods often produce unwanted artifacts in the processed images.

For example, image decomposition algorithms based on biimage total variation are used for denoising and contrast enhancement, however, the color enhancement obtained by this method in many low-light images is not natural enough

[0007] In summary, the above-mentioned methods in the prior art all have problems such as being unable to properly enhance the image, unable to disclose hidden details in low-light images, and the image obtained after color enhancement is not natural enough, etc.

Images