研究了一种基于分数阶微分理论的图像增强算法来提高恶劣天气及复杂场景情况下模糊交通视频图像的清晰度。研究了分数阶微积分Tiansi算子模板的特性, 将现有的分数阶微分模板中被忽略的像素点应用到新的分数阶微分处理之中改进和更新了分数阶微分算子。对改进后的算子与近期常用的和公认的5种图像增强算法进行了比较分析, 结果显示: 改进的分数阶微分算子能够显著增强局部细节信息(如纹理等), 而且图像中的噪声不明显。与Tiansi算子相比, 新算子在处理彩色图像时, 分数阶阶数的增高并不明显地增加图像噪声, 而且锐化结果好, 颜色更真; 处理灰度图像时, 在高频信息较多的情况下, 增加模板中心像素权值可以抑制图像中的噪声。对试验及算法的分析比较证明: 研究的新算法对增强模糊交通视频图像清晰度具有较好的效果, 且处理速度也能满足实际应用的要求。

A new image enhancement algorithm based on fractional differential was proposed to improve the lower visibility of traffic video images in bad weather or complicated situations. The characteristics of fractional differential operator (Tiansi) were researched and some points to be ignore in the existing fractional order differential template were applied to the new fractional order differential treatment to improve and update the fractional differential operator. The improved operator was compared with the Tiansi operator and 5 kinds of common image enhancement algorithms. The comparing results show that the proposed algorithm can enhance the details of an image with less noise. When a color image is processed, the improved Tiansi operator can maintain the image colors well , sharpen the grey level images with high frequency signals and do not increase the noise more. When a gray image is processed, the noise can be suppressed by increasing the pixel weight in the template center with much high frequency information. The different fuzzy traffic video images under various situations are tested. The testing results show that the studied algorithm is suitable for the vague traffic video images under bad weather, and the processing speed of the algorithm can meet the requirements of the real applications.