摘要： 在野外雾天环境下，由于大气粒子的散射作用导致图像降质严重，直接影响图像的视觉效果和应用价值，因此有必要对雾天图像进行去雾处理，以提高雾天图像的清晰度和保真度。为此，提出一种野外视频监控图像去雾新方法。基于暗原色先验去雾的原理，采用区域生长算法准确快速估计雾天图像的深度信息，应用雾天图像物理模型对图像去雾处理，并进行亮度补偿。实验结果表明，该算法能有效改善雾天图像的质量，大幅提高运算速度。
关键词:
雾天图像物理模型,
暗原色先验,
区域生长,
大气光,
高斯平滑,
图像去雾
Abstract: In field conditions, the scattering effect of atmospheric particles leads to serious image degradation and seriously influences the visual effect and application value, which makes it absolutely necessary to defog the image, so as to improve the definition and fidelity. Based on the principle of dark channel priority, this paper uses the region growth algorithm to accurately estimate the depth information of the fog image, and uses the fog image physical model to defog the image and compensate the luminance of the image. Experimental result shows that the algorithm efficiently improves image quality and greatly reduces the time needed in the image fog removal process compared.
Key words:
fog image physical model,
dark channel priority,
region growth,
atmospheric light,
Gaussian smooth,
image defogging
中图分类号: