摘要
介绍了电子稳像技术的发展概况以及相关算法.对KLT算法进行改进,在特征点提取过程中采用测量相异量的方法量化特征点在最初图像和当前图像之间的变化,从而解决了因为外界因素带来的的特征点提取困难的问题;引入亮度适应方法,用相对亮度差替代绝对亮度差,较好地解决了光照变化给特征匹配带来的困难.在移动机器人视觉平台上进行了实验,验证了算法具有实时性的稳像效果.
The development of electronic image stabilization （EIS） and some algorithms was introduced. Dissimilarity method was used to describe the difference of feature point between the initial picture and the current picture based on the KLT algorithm, which solved the difficult feature catching problem induced by outer factor. According to luminance adaptation method, opposite luminance comparison was substituted by absolute luminance comparison, which solved the feature matching difficulty from illumination. Experiment was done on the vision platform of the mobile robot, and verified that this algorithm had a time - stability effect.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2007年第9期1379-1382,共4页
Journal of Harbin Institute of Technology
基金
国家高技术研究发展计划资助项目(2005AA420290)
关键词
移动机器人
电子稳像
KLT
mobile robot
electronic image stabilization
KLT
