车载红外视频图像快速彩色化技术研究  

Fast Colorization Technology of Vehicle-Mounnted Infrared Video Images

作　　者：范德营[1] 蒋先刚[1] 丘赟立[1] 熊娟[1] 

机构地区：[1]华东交通大学基础科学学院,江西南昌330013

出　　处：《华东交通大学学报》2013年第2期71-77,共7页Journal of East China Jiaotong University

基　　金：国家自然科学基金项目(61262031)

摘　　要：为了实时高效地对车载红外视频图像进行彩色化处理,提出了一种在Lab颜色空间基于调色板的分类彩色化方法。对增强和融合处理后的红外图像采用K均值聚类方法得到聚类中心,再利用Fisher评价函数获得分类界限以对红外视频图像进行图块景物分割,在Lab颜色空间中采用分类彩色化方法对各类别景物赋予适当的色彩。试验结果表明,采用调色板的分类彩色化方法具有更高的效率,得到的图像色彩符合人眼的视觉特性且便于夜间红外图像的彩色化实现。This paper puts forward a color classification method based on color palette technology in the Lab color space so as to carry out efficiently the color processing of vehicle-mounted infrared video images in time.The system firstly applied k-means clustering method to the preprocessed infrared images and obtained the clustering center.Then it segmented the block scenery of infrared video images with classification boundaries obtained by Fisher evaluation function.The appropriate color was thus given to different scenery frames by the color classification method.Experiments show that the proposed color classification method has much higher efficiency which can obtain realistic image color fit for human vision and is convenient for achieving colorization of infrared video images at night.

关 键 词：K均值聚类 FISHER评价函数 Lab颜色空间 调色板 分类彩色化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]