欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
基于图像暗通道先验规律的图像去雾算法是当前一种比较先进的基于模型的图像去雾算法，但也存在一些缺陷和不足，例如：算法计算量较大，处理时间较长；可能出现去雾失败现象。针对此类问题，提出一种自适应的图像去雾算法，使用基于暗点优先膨胀算法提高图像暗通道计算速度；采用引导式滤波算法快速细化透射图，改善了“白边（halo）”现象；根据图像本身的特征自适应地计算去雾参数，有效减少去雾失败的现象。实验结果表明，该算法可以动态地适应图像的特征，自适应地调整相关参数，计算效率和效果更好，可以满足视频去雾的需求。
英文摘要：
Image haze removal algorithm based on priori rules of image dark channel is a rather advanced model-based image haze removal algorithm, but there are also some shortcomings and deficiencies, e.g. heavy computation load, long processing time and possible failure in haze removal. In view of these problems, we propose a self-adaptive image haze removal algorithm, which uses dark dots priority-based expansion algorithm to improve computation speed on dark channel of images ; uses guided filtering algorithm to rapidly refine the transmission graph and ameliorates “halo” phenomenon. Haze removal parameters are calculated adaptively according to the features of the image itself and thus the haze removal failures are effectively cut down. As the result of experiment it is demonstrated that the new algorithm can dynamically adapt to image features and adaptively regulates correlated parameters, its computation efficiency and effect are better, and can meet the requirement of video haze removal.
同期刊论文项目
同项目期刊论文
期刊信息
位置：