An effective approach for low-light image enhancement

Abstract

     Low-light image enhancement plays an essential role in many applications of image processing. According to our observations, although many methods have been proposed to solve this problem, the obtained enhanced image may be lost in detail. This paper presents a new approach to solve the above problem. First, we introduce a method of decomposing an image into two components based on a Gaussian filter. Then, this image decomposition method is applied to decompose the input image into two components, the base layer ( ) and detail layer ( ). Local Brightening ( ) method is applied to the component, obtaining the component. The Laplacian edge detection (LED) operator is applied to the detail component ( ), yielding the feature component ( ). The Marine predators algorithm (MPA) is applied to find the suitable parameters ( , and ) for the , , and components. Finally, the enhanced image is calculated as the sum of the components ( , , and ) multiplied by the corresponding adaptation coefficients. Four evaluation indexes, three image enhancement algorithms, three other optimization algorithms, and 95 low-light images were used and evaluated. The experimental results show that the proposed method effectively enhances image quality and ensures sharp images.