基于高亮区域自适应处理的监控图像去雾算法

IMAGE HAZE REMOVAL METHOD BASED ON HIGHLIGHTED AREA ADAPTIVE PROCESSING

针对现有图像去雾算法处理监控图像失真的问题,提出一种高亮区域自适应处理的监控图像去雾算法。通过自适应识别分割高亮区域,根据高亮区域所占图像的比例优化大气光强参数的取值;在暗原色先验的条件下,使用修正的大气光强参数与导向滤波进行透射率的计算和边缘细化;引入自适应容差机制恢复雾化图像,再对恢复后的雾化图像进行色彩增强处理。实验结果表明,算法能够消除恢复图像的色彩失真与块状边缘,较大程度上避免了高亮区域的光晕,保护了天空区域的恢复质量,使恢复图像保持清晰和自然,视觉效果得到较大提高。

Aiming at the monitor image distortion problem of the existing image de-fogging algorithm,a monitoring image de-fogging algorithm with highlight area adaptive processing was proposed.Firstly,the highlighted area was divided by adaptive recognition,and the value of the atmospheric light intensity parameter was optimized according to the proportion of the image occupied by the highlighted area.Secondly,under the condition of dark channel prior,the modified atmospheric light intensity parameters and the directional filter were used to calculate the transmittance and refined the edge.At last,the adaptive tolerance mechanism was introduced to recover the atomized image,and then the restored atomized image was color-enhanced.The experimental results showed that the algorithm could eliminate the color distortion and block edges of the restored image,and the halo of highlighted area was avoided largely.The sky area recovery quality was protected to keep the image clear and natural,so that the visibility of de-fogging image was elevated obviously.