一种基于Gabor小波的制导精度预测方法  

An Algorithm for Estimating the Precision of Scene Guidance Based on Gabor Wavelet

作　　者：杨昕[1] 杨新[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200030

出　　处：《计算机工程》2003年第18期72-73,103,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60072026)

摘　　要：在无人驾驶飞行器的自主制导系统中，一般采用下视景像匹配制导技术对飞行器的具体位置进行精确自主定位。为了提高制导精度，减少误匹配，飞行前需要在基准图上进行航迹规划，就是在飞行区域中选择出一些匹配概率高的匹配区。在选择过程中，匹配概率的计算尤为重要，用相关匹配定义计算虽然准确但耗时巨大。文章提出了一种基于Gabor小波特征的匹配概率预测方法。实验表明，该方法能较快地预测出匹配概率，精度也较高。In the self-service guidance system of unmanned aerocraft, it usually adopts down-ward looking scene matching technique to locate the spot precisely . Aiming at improving the precision and reducing the probability of mismatching , guidance system needs path planning before launch in the reference image , namely , selecting some high matching probability areas of flying region as the base of image registration . During the course of selection , it is very important to calculate the matching probability , however , the result is accurate using the definition of correlation matching but quite calculatively time-consuming . This paper proposes an algorithm for estimating the matching probability based on Gabor wavelet .Experiments show it can estimate the probability fastly with high precision .

关 键 词：景象匹配制导 制导精度 GABOR小波特征 Fisher线性分类器 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]