首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于改进直方图不变矩的声呐图像识别方法
引用本文:赵春晖,李誉斐.基于改进直方图不变矩的声呐图像识别方法[J].应用科技,2007,34(11):19-24.
作者姓名:赵春晖  李誉斐
作者单位:哈尔滨工程大学,信息与通信工程学院,黑龙江,哈尔滨,150001
基金项目:国家自然科学基金;高等学校博士学科点专项科研项目;黑龙江省自然科学基金
摘    要:针对声呐图像受噪声干扰大,目标边界较模糊的特点,并结合传统直方图不变矩识别图像快速的优点,以及其不考虑灰度空间信息而易造成误识的缺点,将目标的形状阕子加入不变矩特征向量中,提出了一种对亮度和对比度变化不敏感,具有旋转和尺度不变性的改进的直方图不变矩.实验仿真结果表明,该方法能够快速、有效地识别水下目标.

关 键 词:声呐图像  识别  特征提取  直方图  不变矩  形状因子
文章编号:1009-671X(2007)11-0019-06
修稿时间:2007-07-13

Recognition for sonar image based on improved histogram invariant moments
ZHAO Chun-hui,LI Yu-fei.Recognition for sonar image based on improved histogram invariant moments[J].Applied Science and Technology,2007,34(11):19-24.
Authors:ZHAO Chun-hui  LI Yu-fei
Institution:College of Information and Communications Engineering, Harbin Engineering University, Harbin 150001,China
Abstract:Considering the blurred boundary on the sonar image of target due to the heavy noise, this paper put forward an improved histogram invariant moments for the recognition of sonar image in conjunction with the shape factor from the traditional histogram invariant moments. This method is not sensitive to luminance and contrast. Mo- reover, the histogram invariant moments in this method feature invariant rotation and scale, not only preserving the recognition speed of the original method but avoiding the probability of wrong recognition due to the space informa- tion loss. Experimental results show that the improved method can recognize the underwater target quickly and effectively.
Keywords:sonar image  recognition  feature extraction  histogram  invariant moment  shape factor
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号