论文部分内容阅读
本文针对毫米波雷达距离像识别这一应用背景,提出了一种应用极化信息使距离像幅值趋于平稳,用神经学习来提取距离像角不变特征的模式识别新方法;其特点是通过极化处理减弱斑纹可增强对目标的确认,经神经学习获取的子类型特征与常规方法相比载有更多信息因而识别率高;实验结果证明了该方法的有效性.
Aiming at the application of distance radar image recognition in millimeter wave radar, this paper proposes a new pattern recognition method using polarized information to stabilize distance image amplitude and neural learning to extract invariant features of distance image angle Weakening the speckle by polarization can enhance the recognition of the target. The subtype characteristics obtained by neural learning contain more information than the conventional method, so the recognition rate is high. The experimental results show the effectiveness of the method.