高级检索

考虑分数阶梯度的雾天图像增强偏微分方程模型

A PDE Model for Foggy Image Enhancement Based on Fractional Gradient Field

  • 摘要: 为了在去雾的同时增强图像中的纹理细节信息,提高图像亮度,改善图像质量,提出一个雾天图像增强的分数阶偏微分方程模型.将分数阶微分与大气散射物理模型结合,建立了去雾图像的分数阶梯度场;为了突出图像的纹理细节信息,避免出现边缘过度增强或细节纹理增强不够的现象,构造了分数阶梯度场的增强函数,使分数阶梯度场随着梯度模的变化达到非线性增强的效果;在梯度域建立能量泛函,使雾天图像梯度场逼近增强梯度场,通过变分法得到分数阶偏微分方程图像增强模型;最后用有限差分法对模型进行数值求解.实验结果表明,文中模型在去雾的同时,能够有效地提高图像的对比度和清晰度,是一种有效的雾天图像增强模型.

     

    Abstract: In order to enhance the texture detail information and brightness, improve the image quality while re- move the fog, a fractional partial differential equation (PDE) model for foggy image enhancement is proposed in this paper. The fractional gradient field of fog-free image is established by combining fractional differential with atmospheric scattering physical model. For highlighting the texture detail information of the image, avoiding the phenomenon that the edge is excessively enhanced or the detail texture is not enhanced com- pletely, an enhancement function of the fractional gradient field is constructed, which can make the frac- tional gradient field enhance non-linearly with the change of the gradient modulus. In order to approximate the gradient field of the fog image to the enhancement gradient field, an energy functional is established in the gradient domain. And then the fractional partial differential equation model for foggy image enhancement is obtained by the variational method. Finally, the finite difference method is used to solve the model nume- rically. The experimental results show that the proposed model can effectively improve the contrast and sharpness of the image while removing the fog. Hence the model is an effective fog image enhancement model.

     

/

返回文章
返回