| CPC G01S 13/9011 (2013.01) [G01S 13/9088 (2019.05)] | 21 Claims |
|
1. A method, comprising:
receiving raw synthetic aperture radar (SAR) data associated with observed views of a scene, wherein the raw SAR data comprises information captured via an automatic target recognition (ATR) system having a SAR system;
radio frequency (RF) preprocessing the received raw SAR data to produce a processed raw SAR data which includes two-dimensional data that includes frequency band data and pulse domain data;
converting the processed raw SAR data to a complex SAR range profile data;
processing the complex SAR range profile data with a recurrent neural network (RNN) having RNN states; and
mapping the RNN states to a target class with a multi-layer linear network;
wherein the processed raw SAR data is represented by a two-dimensional array represented by PH(m,n), wherein m represents frequency band values of a row vector Pf of the frequency band data and n represents pulse domain values of a column vector Pr of the pulse domain data; and
wherein converting the processed raw SAR data to the complex SAR range profile data comprises:
determining a range profile vector rpr(m) defined by a relationship defined as
 where:
Pw(k) is derived from Pr(i)=PH(i,r),
M represents a maximal range bin that is determined by a number of sampling frequencies in the frequency band data; and
m represents range bins;
wherein the complex SAR range profile data is obtained based on the range profile vector rpr(m).
|