改善劣质实时图像预处理效果的优选法研究  

Pre-processing of poor real-time images with optimum selecting algorithm

作　　者：陈慧津[1] 陈哲[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100083

出　　处：《光电工程》2006年第1期89-93,107,共6页Opto-Electronic Engineering

基　　金：航空科学基金资助项目(03D51007);航天科技创新基金项目:智能军用制导多传感器信息融合技术目

摘　　要：针对某次实际飞行实验条件下所采集的有较严重问题的CCD实时景像图预处理算法研究问题,提出一种优选法。该方法先利用已有的Photoshop图形处理软件与人工手动调节相结合对原始图像进行处理,得到期望图像,再根据原图像的具体问题和已有的先验知识选择几种可能的预处理方法对原图像进行仿真,将所得图像与期望图像进行对比选出最适合处理此种劣质图像数据的优化预处理算法。将该算法应用于本次飞行实验数据上,飞行航线的匹配概率由5.41%提高到了86.48%,结果表明本文提出的优选法具有可行性和实用性。An optimum selecting algorithm is presented for the pre-processing of the poor CCD real-time image gotten from a flight experiment in practice. Firstly, the expected images were gotten by manual processing original images with Photoshop software. Secondly, these original images were pre-processed with several probable algorithms based on the practical situation of original images and prior knowledge. Finally, an optimization algorithm, which was most suitable for improving this type of poor images, was chosen by comparing the pre-processing results with the expected images. The algorithm were utilized to preprocess the flight experiment data. Experiment shows that matching probability of one flight line increases from 5.41% to 86.48%.

关 键 词：号像匹配 图像预处理 劣质图像 优选法 

分 类 号：TN911.73[电子电信—通信与信息系统]