基于ODNHP的毫米波探测器地面目标识别算法

Millimeter Wave Detector Ground Target Recognition Algorithm Based on ODNHP

针对毫米波探测器识别地面目标时的强地杂波问题,在正交判别邻域保持投影(ODNPP)的基础上,提出正交判别邻域Hessian投影(ODNHP)算法。ODNHP在正交化约束和判别信息的基础上,利用样本的切空间坐标构造Hessian二次型矩阵,实现目标函数的优化。将ODNHP分别与NPP、ONPP和ODNPP进行了毫米波地面目标识别仿真实验,结果表明,ODNHP降低了杂波的影响,在较低的维数下实现更高的识别率。

In view of strong ground clutter when millimeter wave detector recognizing the ground object,based on the orthogonal discriminant neighborhood Hessian projection（ODNHP） algorithm,the orthogonal discriminant neighborhood keep projection（ODNPP） was presented.Introducing Hessian operator in ODNHP can ensure the optimization of object function basis on keeping information of ODNPP Boolean manipulation in restraint and discrimination.The experiment of millimeter wave detector ground target recognition simulation is respectively used the algorithms of NPP、ONPP、ODNPP and ODNHP.The results show that ODNHP can reduce the influence of clutters and get higher recognition rate at the lower dimension.