基于四阶PDE的去噪方法在煤矿井下视频监控图像中的应用

Denoising Method for Application in Coal Mine Monitoring Image Based on Fourth-order Partial Differential Equations

煤矿井下特殊的应用环境直接影响着图像的质量,导致图像视觉效果较差。文章提出一种改进的基于四阶PDE的去噪方法,该方法将图像区域分为闪点区域和非闪点区域,并针对不同的区域采用不同的传导系数。与传统的各向异性扩散方法相比较,仿真实验结果表明,该方法在抑制噪声的同时能够有效克服二阶偏微分方程引起的块效应、消除闪点并保持图像的边缘、线条纹理等细节,从而达到更好的图像质量和视觉效果。

The final monitoring images in video monitoring system are usually in low contrast and brightness because of special application environment in underground mine. A improved denoising method based fourth-order Partial Differential Equations （PDE） is proposed in this paper. The proposed method segments the image domain into speckle domain and non-speckle domain, then uses different conductance coefficients depending on the domain type. Comparing to traditional anisotropic diffusion method, the simulation results show that this method will suppress the noise so as to achieve a better image quality and visual effect. At the same time, it can effectively remove isolated speckles, eliminates flash point ,keeps edge details and line texture in the image.