【摘要】 基于反转的低照度图像与雾天图像的相似性，结合物理模型和数学优化理论，提出了一种改进的基于去雾模型的低照度图像增强方法.该方法通过设计一种具有结构感知性的权重函数，构建了一个适用于雾天退化模型的透射率目标函数，从而对反转的低照度图像进行去“雾”.与经典低照度图像增强方法相比，该方法增强的图像颜色自然，细节保持良好，同时在信息熵、平均梯度、自然图像质量评价指标表现较好.更多还原

【Abstract】 According to the similarity between the inverted low illumination image and hazy image, an improved low-illumination image enhancement method based on dehazing model is proposed in this paper, which combines physical model and optimization theory.This work designs a structure-aware weighting function and constructs an optimal transmission function suitable for dehazing model, thus, the inverted low-illumination image is ″dehazed″.Compared with some classical low-illumination enhancement methods, images enhanced by this method have natural colors and vivid details.Besides, this method obtains better image quality assessment results in metrics such as IE,AG,and NIQE.更多还原