摘要
研究了多信息融合技术在SAR图像目标识别中的应用。将扩展分形特征(ExtendedFrac tal)与双参数恒虚警特征(DoubleParameterCFAR)形成的多信息进行融合处理,运用Dempster Shafer证据理论,在决策层对SAR图像中的像素进行识别分类。实验结果表明通过融合对像素分类的准确性明显好于单特征的检测结果,减少了虚警概率,提高了系统的识别能力。
In this paper ,multi-feature fusion technique is used to recognize and classify the SAR image target. The EF feature and DP-CFAR feature are fused by using the Dempster-Shafer' Evidential Reasoning .The decision strategies are applied to classify the pixel of the SAR image.The experiment results indicate that the fusion method's accuracy is better than the detection results of single feature .The fusion can diminish the false alarm rate and enhance the recognition performance of SAR automatic target recognition (ATR) system
出处
《雷达与对抗》
2004年第1期22-25,共4页
Radar & ECM
关键词
信息融合
SAR
目标识别
扩展分形
双参数恒虚警
合成孔径雷达
synthetic aperture radar(SAR)
target detection
data fusion
Extended fractal(EF)
Double Parameter-CFAR(DP-CFAR)