基于多窗二值局部熵的特征提取方法及其在红外目标识别中的应用  

Multi-Window Binary Local Entropy Based Feature Extraction Algorithm and Its Application to Infrared Target Recognition

作　　者：张世俊[1] 敬忠良[1] 李建勋[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院航空航天信息与控制研究所,上海200030

出　　处：《模式识别与人工智能》2005年第5期621-627,共7页Pattern Recognition and Artificial Intelligence

基　　金：国家自然科学基金(No.60375007;60375008);航天支撑技术基金(No.2003-1.3 02);上海市科技发展基金重点(No.015115038);教育部高等学校博士学科点专项科研基金(No.20020248029)

摘　　要：特征提取是模式分类及识别的重要组成部分。结合局部熵描述图像局部性质的特性,本文提出一种基于多窗二值局部熵的特征提取算法。通过几何矩的规范性变化,使得由多窗局部熵提取出的特征量同时具备了平移及比例不变性,并引入圆形局部窗使得特征量具备旋转不变性。利用最近邻法对标准飞机模式库中的12种飞机及真实红外目标的分类识别实验结果表明,这种方法提取的特征量具有计算简单且对噪声影响不敏感的特性,其性能优于矩不变量及Zernike矩。Feature extraction constitutes the basis of pattern classification and recognition. Using the property of local entropy of describing local image properties, a feature extraction algorithm based, on multi-window binary local entropy is proposed. By normalizing the target using its geometrical moments, the feature vectors have invarianee of translation and scale, and the circle local window is introduced to make the feature vectors be rotation invariant. Feature extraction and recognition are performed with 12 kinds of airplanes in standard pattern library and real-world infrared targets. The experimental results verify the simple calculation, insensitivity to noise and superiority of multi-window binary local entropy over moment invariants and Zernike moments.

关 键 词：多窗二值局部熵 特征提取 红外目标识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]