基于红外目标提取的夜视图像融合实时系统研究  

Research of Real-time Night-Vision Image Fusion System Based on Infrared Object Extraction

作　　者：胡德超[1] 朱尤攀[1] 罗琳[1] 李泽民[1] 范宏波[1] 孙爱平[1] 苏凡[1] 韩娟[1] 

机构地区：[1]昆明物理研究所,云南昆明650223

出　　处：《红外技术》2014年第2期125-130,共6页Infrared Technology

摘　　要：研究的图像融合技术是以微光图像作为背景,红外图像作为目标,两光路图像经过合光棱镜进行光学投影融合的实时融合技术。为了具备良好的融合效果,红外图像经过预处理后,用改进的Otsu算法进行图像分割来实现目标提取,提取后的红外目标经过改进的伪彩色变换方法进行处理后在OLED上显示。融合后的系统实时性好,既可以得到微光细节丰富、较接近可见光视觉感受的背景,也可得到对比度明显的红外伪彩色目标,对于战场上的士兵观察环境和快速寻找目标有很大的帮助。In this paper image fusion technology researched is optical overlay real-time fusion technology, in which low-level-light image is as background, infrared image is as object, and image from double optical path is through prism to fusion. For good fusion performance, infrared image is preprocessed, and then an improved Otsu algorithm is used in the process of image segmentation for object extraction. An improved pseudo-color process is used in the process of infrared object and at last the infrared object pseudo-color image is shown on the OLED. The fusion system is real-time, whose image’s background has more details and is similar to visible light, and high contrast pseudo-color object is distinct, which can help the soldier in the battlefield to observe the surroundings and to fast look for object.

关 键 词：红外图像 微光图像 伪彩色处理 图像融合 目标提取 OTSU算法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]