Image Clarification Algorithm for Adverse Weather
Yinzhou Zhang, Jiya Menghe · 2024
In the current era, adverse weather conditions such as dust storms and haze occur frequently. Addressing issues like color bias, uneven color distribution, low contrast, and brightness in images affected by such conditions, we propose an image clarification algorithm tailored for adverse weather. Specifically, for the issue of color bias, we introduce a color correction algorithm based on the Lab color space and a Weighted Fusion Gaussian Model. This algorithm surpasses traditional methods in terms of color correction accuracy and adaptability. To tackle the problems of dull colors and low clarity, we propose a color restoration algorithm based on an improved Multi-Scale Retinex with Color Restoration (MSRCR). Lastly, image dehazing and brightness enhancement are achieved through an enhanced algorithm based on the Dark Channel Prior. Experimental results and analysis demonstrate that our approach not only effectively corrects color bias but also enhances image contrast and clarity, significantly improving the clarification performance of the base model.