基于暗通道原理的图像去雾算法改进研究

Research on Image Dehazing Algorithm Improvement Based on Dark Channel Prior

针对采用暗通道原理的单幅图像去雾算法的处理效率,提出了采用低分辨率暗通道模板进行有雾图像去雾处理的改进方法;该方法利用低分辨率图像估算大气透射率并利用插值还原模板分辨率以估算无雾图像,可以在保持处理效果的同时,有效节省图像去雾的处理时间及与对处理单元资源的需求;通过分析及仿真计算验证了该方法的有效性。该方法还适用于对高分辨率图像去雾处理,具有良好的工程实用性。

In order to improve the processing efficiency of dehazing algorit.hm using dark channel prior principle, a new method is presented in this paper.The method uses a low-resolution dark channel template to estimate the atmospheric transmissivity and restores the template resolution using interpolation.Through analysis and simulation, it is validated that the method can effectively save computing time and reducethe requirements for the processing unit without affecting the processing performance.It can alsobe used for high-resolu- tion image dehazing processing.