A Fast Image Dehazing Algorithm for Highway Tunnel Based on Artificial Multi-exposure Image Fusion

Abstract

Single image dehazing has been a challenging problem due to its ill-posed nature. This paper proposed a fast image dehazing algorithm for highway tunnel based on artificial multi-exposure image fusion. The original blurred images are transformed into under-exposed images using a series of gamma-correction operations. Meanwhile, a multi-exposure image fusion algorithm based on two-scale decomposition is proposed. It exploits the guided filter to separate each multi-exposure image into base and detail layers. The fused base and detail layers are integrated into the fused images. Then, a linear saturation adjustment method is utilized to enhance the saturation in spatial domain. Through the above image processing steps, the final haze-free images can be obtained. In this paper, the superiority of the proposed image dehazing algorithm based on artificial multi-exposure image fusion is demonstrated through theoretical analysis and experimental verification. It can quickly improve the visibility of foggy images, and make the processed images colorful and clear.