暗通道去雾技术和聚类算法的优化研究

Research on Optimization of Defogging and Clustering Algorithm Based on Dark Channel

在交通运输行业中,每当天气中含有大量雾尘的时候,大气会发生散射,会对驾驶人员的视觉造成一定的干扰,对交通安全有着极大隐患;为避免交通事故的发生,我们在原有算法的基础之上,结合机器学习中的K—means聚类算法进行了优化研究,对含有雾的图像进行了去雾处理,使得图像的能见度增加;根据现实研究当中问题的实际情况,我们建立了暗原色模型,并利用暗原色算法拟合透射率,用K—means算法进行聚类分析进行处理强化图像的特征,结合估计出来的大气光强,利用去雾算法得到最终的无雾图像;最后我们通过Matlab进行分析演示,并且与其他算法的去雾处理图像进行对比与分析,图片更加清晰,可以很好的应用在航海交通,公路运输,气象遥感等方面的去雾领域,具有一定的应用价值。

In the transportation industry, whenever the weather contains a lot of fog, the atmosphere will scatter, which will cause certain interference to the driver s vision, which has great hidden dangers to traffic safety. In order to avoid the occurrence of traffic accidents, we carried out optimization research based on the original algorithm and K-means clustering algorithm in machine learning, and defogged the image containing fog to increase the visibility of the image. According to the actual situation of the problem in the real research, we established the dark primary color model, and used the dark primary color algorithm to fit the transmittance. The K-means algorithm was used for cluster analysis to process the characteristics of the enhanced image, combined with the estimated atmospheric light intensity. The final fog-free image is obtained using a defogging algorithm. Finally, we analyze and demonstrate through Matlab, and compare and analyze with the defogged images of other algorithms. The pictures are clearer and can be applied to the defogging field in navigation, road transportation, meteorological remote sensing, etc. Value.