摘要： 针对传统的雾霾污染等级检测方法中实时性较差、以点带面及成本较高等问题,提出一种基于图像灰度差分统计的检测方法。利用人眼能够对雾霾污染图像进行辨识的特点,通过分层处理将图像置于RGB空间下,运用图像灰度差分统计的方法计算图像的特征参量熵值和 Canberra距离,对雾霾污染等级进行分类识别。实验结果表明,该算法能有效识别雾霾天气,具有识别率高、实时性强的特点。

关键词: 灰度差分统计, 大气气溶胶, 雾霾污染, 熵值, Canberra距离

Abstract: In view of the problems of pollution level detection of the traditional method,such as poor real-time performance,fan out from point to area,and higher cost,this paper presents haze pollution detection algorithm based on the image gray differential statistics.Using the characteristics that human eye can be used for haze pollution identification,the image is assigned by slicing under the RGB space,followed by the use of the image gray differential statistical method for image processing,calculation of entropy characteristic parameters for the processed image,the final image by calculating the distance to Canberra haze of pollution classification.Experimental results show that the algorithm can effectively identify fog and haze,with a high recognition rate,and high real-time characteristics.

Key words: gray differential statistics;atmospheric aerosol;haze pollution;entropy;Canberra distance gray differential statistics, atmospheric aerosol, haze pollution, entropy, Canberra distance

中图分类号: 

• TP391.41