基于贝叶斯估计的目标特征识别扩散参数挖掘模型

Mining Model of Time-frequency Target Characteristics Identification Diffusion Parameters Based on Bayesian Estimation

为了有效从海量的特征和噪声数据提取能表征目标特征的有用数据,提高对目标的检测识别能力,需要对目标回波特征的时频TID(time interference domain)域扩散参数进行快速挖掘,达到提取目标特征的目的。传统方法中采用频繁项挖掘方法实现扩散参数挖掘,导致在速度模糊的情况下对基带回波特征参数估计困难,性能不好。提出一种基于贝叶斯估计的目标特征识别扩散参数挖掘模型。有效挖掘出局部离群点,对基带回波特征进行频域变标脉冲压缩处理,对优化后的特征矢量进行累加限幅,并计算互补累积分布函数,基于贝叶斯估计构建检测统计量和统计函数,从而挖掘出时频TID域扩散参数,提高对特征参数的识别能力,仿真结果表明,该算法对时频TID域扩散参数挖掘精度较高,能有效提高对运动状态目标的参数估计精度及目标识别的能力。

Noise from vast amounts of data and target data can represent the target feature extracting useful data, improve recognition ability of target detection, the need to the time-frequency characteristic of target echo dar （Time Interference Domain） diffusion parameters for mining, achieve the goal of target feature extracting. Traditional methods of mining frequent items mining method was adopted to realize the diffusion parameters, resulting in speed under the condition of the fuzzy characteristics of baseband echo parameter estimation is difficult, the performance is not good. In this paper, a target feature recognition based on bayesian estimation diffusion parameters mining model. Effective, dig out the local outliers of baseband echo characteristic frequency variable pulse compression processing, the characteristic vector of the optimized ac- cumulative limiter, and calculate the complementary cumulative distribution function, based on the bayesian estimation to construct test statistics and statistical functions, to excavate the time-frequency dar diffusion parameter, improve the ability to identify the parameters of the target data simulation results show that the algorithm pair high precision frequency dar field diffusion parameters mining, can effectively improve the motion target parameter estimation precision and the ability to target recognition.