高温目标红外测量图像伪彩色显示方法  

Pseudo color display method for infrared measurement image of high temperature target

作　　者：吕游[1,2] 何昕[1] 魏仲慧[1] 常松涛[1,2] 孙志远[1] 何家维[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院大学,北京100039

出　　处：《系统工程与电子技术》2015年第6期1424-1430,共7页Systems Engineering and Electronics

基　　金：吉林省自然科学基金(201115124)资助课题

摘　　要：红外测量图像一般为灰度图像,对其进行伪彩色处理是一种非常重要的增强手段。传统的伪彩色处理方法未考虑到实际图像中高温目标与低温背景间的灰度级断点,处理后的伪彩色图像往往颜色数偏少,细节信息损失较多。为更好地显示高温目标及其边缘细节并兼顾低温背景的显示效果,本文首先根据灰度级将图像分为目标区、背景区和目标-背景过渡区,背景区包含多个背景及背景间过渡区。利用K-均值聚类理论对背景区进行聚类,根据聚类结果,采用灰度区间压缩方法确定多个背景及背景间过渡区的灰度级范围。根据各区间内灰度统计特征对整幅图像进行多段线性灰度映射,有效地均衡图像的灰度分布。引入视觉差异因素,采用CIE94颜色差异公式对传统颜色路径进行改进,实现等颜色差异的红外伪彩色显示。实验结果表明,本文方法较传统方法更适合高温目标红外测量图像,处理后的图像目标突出,图像细节色彩丰富,背景对比度较高,大大提高人眼对高温目标红外测量图像细节信息的辨识能力。The infrared measurement image is usually a gray image, and pseudo color processing is a very important means of enhancement. The traditional pseudo color processing methods do not consider grayscale breakpoints between the high temperature target and low temperature background in the actual image, after pseudo color processing, the number of colors of the image is always small, and lots of details information is lost. In order to get better display effects of the high temperature target, its edge details and the low temperature background, according to the gray level, the image is divided into the target area, background area and target background transition area, and the background area contains multiple transition areas between the back- grounds. The K-means clustering theory is used to cluster the background. According to the clustering results, confirm image gray level ranges of the multiple background area and transition area by the gray level compres- sion method. According to the statistical characteristics of each interval gray, map the multistage linear gray level for the whole image, and balance the image grayscale distribution effectively. Introducing visual difference factors, use the CIE94 color difference formula modify the traditional color path, realize pseudo color display of the image in equal color difference. Experimental results show that, the method is more suitable for the high temperature target infrared measurement image than the traditional method, the target of the processed image is outstanding, the color of image details is abundant, and the contrast of background is high, it can largely en hance the ability of recognition with infrared measurement image details for high temperature targets.

关 键 词：红外测量图像 伪彩色显示 K-均值聚类 多段线性拉伸 CIE94颜色差异公式 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN216[自动化与计算机技术—计算机科学与技术]