基于Q位移二分树复数小波变换算法的图像特征提取  

Extraction of Image Characteristic Based on Arithmetic of Q-shift Dual Tree Complex Wavelet Transformation

作　　者：宋凯[1] 张俸侨[1] 祁艳[1] 张焕君[1] 

机构地区：[1]沈阳理工大学信息科学与工程学院,沈阳110168

出　　处：《计算机科学》2008年第5期253-256,共4页Computer Science

基　　金：基金项目:863计划项目(2003AA41250);辽宁省教委A类基金(20243303)

摘　　要：为了提高图像特征的识别率,研究并设计了基于Q位移二分树复数小波变换算法的图像特征识别系统。论述了Q位移二分树复数小波变换算法的基本原理及滤波器的工作原理。在设计中保留了复数小波的优点,满足完全重构奈件,同时具有良好的对称性。实验样本采用农作物黄瓜霜霉病叶部病斑图像,在Visual C++软件平台上时样本进行了训练和学习的仿真实验,仿真结果验证了系统的有效性。In order to improve the recognition rate of image characteristic, an image characteristic recognition system based on Q-shift dual tree complex wavelet transformation algorithm is studied and designed. The advantages of complex wavelet are reserved in this design. It satisfies the entire reconstruction conditions and has favourable symmetry. The image of downy mildew of cucumber is used as experiment sample. The training and learning simulation experiments are done on Visual C＋＋ software platform. The result of simulation experiment illustrates the efficiency of the proposed system.

关 键 词：Q位移二分树复数小波变换 图像特征 黄瓜霜霉病 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN219[自动化与计算机技术—计算机科学与技术]