暗原色先验图像去雾算法研究

Research on the image haze-removal algorithm based on the prior dark-channel

为了减小有雾天气对户外成像系统尤其是航拍视觉系统成像的影响,提出了一种新颖快速的基于暗原色先验和雾天图像退化模型的图像去雾方法。结合航拍图像特点,对原暗原色先验去雾算法做了优化和改进。选取一系列户外带雾图像进行大量实验,结果表明,该方法能从物理特性上快速明显的去除雾的干扰,提高图像的清晰度,增强图像色彩和细节,复原得到高质量的图像。实际工程测试表明,文中方法能有效提高雾天航拍成像系统的能见度,降低气象条件对航空成像测量方面的限制,对提高其雾天工作的可靠性和稳定性具有重要意义。

In this paper we propose a fast and effective method to process the haze degraded image based on prior dark-channel and the physical model of hazy image in order to reduce the bad effect in the outdoor visibility system especially the aerial visibility imaging system in foggy days.The method optimizes and improves the original dark-channel prior algorithm which combines the characteristics of the aerial images.Using this method,we can recover a high quality hazefree image.The experiments of processing many outdoor foggy images have been conducted,and the results demonstrate that it can efficiently remove the haze from the physical character of the haze,increase the definition of haze-degraded images,enhance the details and color,and recover high quality haze-free images.The test of project indicates that it can improve the visibility of the aerial imaging system and reduce the bad weather effect on the aerial imaging and measurement.