基于相关分析的飞机目标识别方法

摘要

提出了一种基于相关分析的飞机目标识别方法.该方法利用飞机图像低频和高频部分合成滤波器模板,能达到很高识别率与很低的等错率.该研究旨在提高飞机识别的准确率和降低出错率,采用一种基于相关分析的飞机目标识别方法.该方法通过对采集的飞机图像做去除背景、降噪、图像增强、二值化和归一化处理,将飞机图像低频和高频部分合成滤波器模板,通过特征比对达到识别飞机的目的.利用Matlab 7.0做10种飞机的识别实验,得出了95.47％识别率和0.04％等错率的结论,识别率和等错率均优于不变矩法、三维识别方法、基于小波分析和矩不变量的方法,印证了笔者提出的基于相关分析的飞机目标识别方法的优越性.在飞机图像数据库上的实验结果表明,该方法是可行的.%Aircraft objection recognition method is proposed in this paper,which uses the low and high frequency features to synthesize the filter templates.The method has very high recognition rate and very low error for aircraft recognition.The goal of this research is to improve the recognition accuracy and lower error rate,adopt an aircraft target recognition method based on correlation analysis.This method through the collection of the plane images for background removal,noise reduction,image enhancement,binarization,and the normalized processing,the plane image synthesis of low frequency part and high frequency filter template,through feature to identify the plane than to achieve.Ten kinds of airplane recognition experiments using Matlab 7.0,the 95.47％ recognition rate is obtained and 0.04％ error rate,the recognition rate and fault rate,etc is superior to the invariant moment method,3D recognition method,the method based on wavelet analysis and moment invariants,confirms the author puts forward the superiority of aircraft target recognition method based on correlation analysis.Experiment results on the aircraft image dataset validate the effectiveness and feasibility of the proposed algorithm.