用于HRRP多类目标识别的D距离分类器

在雷达自动目标识别（radar automatic target recognition,RATR）领域，为了保证基于高分辨距离像（high-resolution range profile,HRRP）的目标识别算法在进行小样本多类目标识别时仍然具有优异的识别性能，需要提出一种同时具备优异泛化性能与低运算复杂度的识别算法。利用比值计算两个向量之间的比值距离，并将比值距离应用于距离分类器中，称之为D距离分类器。然后利用八类地面目标实测数据将D距离分类器与其他一些RATR统计模型进行比较，分别分析其在小样本与多类目标时的识别精度。最终结果验证出D距离分类器在训练样本有限且多类目标识别时仍然具有优异的泛化性能与很低的运算复杂度。

D Distance Classifier for HRRP Multi-Class Recognition

In the field of Radar Automatic Target Recognition (RATR), in order to ensure that the target recognition algorithm based on High-Resolution Range Profile (HRRP) still has excellent recognition performance when performing small-sample and multi-class target recognition , it is necessary to propose a recognition algorithm with excellent generalization performance and low computational complexity. Use the ratio to calculate a ratio distance between two vectors, and apply the ratio distance to a distance classifier, which is called D distance classifier. Then, the D distance classifier is compared with some other RATR statistical models using the measured data of eight ground targets, and its recognition accuracy in small samples and multi-class targets is analyzed respectively. The final result verifies that the D distance classifier still has excellent generalization performance and low computational complexity when recognition is performed with small-sample and multi-class target.