北京航空航天大学学报 ›› 2008, Vol. 34 ›› Issue (09): 1012-1015.

• 论文 • 上一篇    下一篇

改进PCA在热波图像处理中的应用

孙延春1, 马齐爽1, 刘跃明2, 姚红宇3   

  1. 1. 北京航空航天大学 自动化科学与电气工程学院, 北京 100191;
    2. 中国国际航空股份有限公司工程技术分公司 成都维修基地, 成都 610200;
    3. 中国民用航空总局 航空安全技术中心, 北京 100028
  • 收稿日期:2007-08-06 出版日期:2008-09-30 发布日期:2010-09-17
  • 作者简介:孙延春(1978-),男,山东定陶人,博士生,ace9024@126.com.
  • 基金资助:

    国家自然科学基金资助项目(60572181)

Application of improved PCA to thermal wave image processing

Sun Yanchun1, Ma Qishuang1, Liu Yueming2, Yao Hongyu3   

  1. 1. School of Automation Science and Electrical Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100191, China;
    2. Chengdu Aircraft Maintenance Base of Air China Limited, Engineering Technics Branch, Chengdu 610200, China;
    3. Center of Aviation Safety Technology, Civil Aviation Administration of China, Beijing 100028, China
  • Received:2007-08-06 Online:2008-09-30 Published:2010-09-17

摘要: 针对飞机结构隐藏腐蚀热波成像检测中出现的热像仪中有用数据较少、图像序列中腐蚀区域与非腐蚀区域对比不明显、信噪比较低等问题,提出了基于小波变换的改进PCA(Principal Component Analysis)算法.该算法对热图像序列中的每帧图像进行二进制离散小波变换,提取出低频部分的系数;对低频系数进行改进PCA分析,通过固定点算法并引入Gram-Schmidt正交化过程求解特征向量;令高频系数为零,抑制图像的噪声.实验结果表明:该算法充分利用了热像仪的有用数据,提高了图像的质量和信噪比,且计算时间较短.

Abstract: To solve the problems less valid data in thermal infrared imager, uneasy dipartite corrosion area in image and lower the signal to noise ratio which happened in detection of structure parts- hidden corrosion by thermal wave imaging, the improved PCA (principal component analysis) based on the wavelet was presented. The every frame image was transformed by dyadic discrete wavelet and the characteristic of wavelet transform coefficient was analyzed. Low frequency coefficients were analyzed by fast principal component method which quoted Gram-Schmidt orthonormalization procedure in fixed-point algorithm then distilled principal component and reconstructed the low frequency coefficients. To suppress image noise, the high frequency coefficients were equated to zero. The experimental results indicated that the method made best of thermal infrared imager-s dada, improved the contrast of infrared thermal image and reduced noise of infrared thermal image, and that the method-s compute-time was short.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发