讨论了使用最佳鉴别准则作为特征提取的方法识别飞机的问题。在介绍最佳鉴别集求法之后，通过计算距离可分性测度得出该方法比一般的Ｋ－Ｌ变换提取特征具有更好的可分性；最后用模糊ＡＲＴＭＡＰ神经网络作为分分类器以处理后的数据进行识别。对某相控阵雷达所得到的三类过航飞机的数据识别结果表明，该方法可以得到更高的识别率。

This paper discusses the utilization of Foley-Sammon optimal set as feature extraction to identify plane targets. Alter the method for computing the Foley-Sammon optimal set are presented, the measure based on range shows this way is more effective than the K-L transform. At last fuzzy ARTMAP neural network which is a classifier is used to complete target classification. Experimental data from a phased array radar shows that this way is more effective.

TN957.52