Rain Detection and Removal of Sequential Images

摘要：

Detection and removal of rain in image is a difficult and crucial problem due to the complexity of rain and its negative effects on image. The spatio-temporal property and the chromatic property of rain are comprehensively analyzed. Using the two properties, a simple but effective algorithm is proposed to detect and remove the rain of sequential images. Firstly time complexity of k-means is reduced by our optimized algorithm. Then each pixel is classified by the optimized k-means according to the spatio-temporal property. Next the chromatic constraint is used to identify accurately the rain. Finally the blending parameter α is computed and the rain in each frame is removed. The method can handle both light and heavy rain in image. Experimental results prove that our algorithm performs simply but effectively.

关键词： Rain detection rain removal image restoration image procession chromatic property K-means

作者: Ming Zhou Zhichao Zhu Rong Deng Shuai Fang

作者单位: School of Computer and Information, Hefei University of Technology, Hefei 230009, China

会议类型: 国际会议

会议名称: 2011 China Control and Decision Conference(2011中国控制与决策会议 CCDC)

会议地点: 四川绵阳

会议语种:英文

页码: 615-618

在线出版日期: 2011-05-23（万方平台首次上网日期，不代表论文的发表时间）