基于大气散射模型的暗通道先验图像去雾改进算法

打开文本图片集

中图分类号：TP391 文献标志码：A

Abstract：Toaddresstheproblemsthe incompleteedgedehazinghaloartifacts in traditionaldark channel prior image dehazing algorithms for image defogging，an improveddark channel prior image dehazing algorithm basedon the atmospheric scattring model was proposed.Thisalgorithm improved thecalculation method transmittance.First，dark chanel maps with filtering windows diferent scales werecalculatedtoobtain transmittance images， then weighted fusion wasperformonthe transmitance images derived from the filter windows different scales.Subsequently，guided filtering wasemployed to refine the fused transmitance map，a clear haze-free image was derived using the atmospheric scatering model.Theresultsshow that the proposed algorithm exhibits significant advantages in transmitance estimation.Compared with traditional algorithms，itimproves the dehazing performanceatscene edges ffectively avoids halo effects around object boundaries.

words： image dehazing；dark channel； guided filtering； transmitance； atmospheric scattering model

随着计算机视觉和图像处理技术的不断发展，图像去雾一直是一个备受关注的研究方向。（剩余6358字）