论文部分内容阅读
目标分类就是在M个假设目标中,确定哪个假设目标的形状与来自散射场的观测数据反映的目标特征最符合,可以说这是逆散射尚待完善的一种应用。本文介绍了一种利用雷达接收机观测到的后向散射波形或目标特征信号来辨别各种形状的目标的方法。虽然近年来一些文献对这个课题很重视,但是,本文却是首先对特定条件下用正在运转的雷达进行目标自动分类作理论分析的文献之一。这种体形分类方法是利用工作在不同载频和极化方式的两个或多个辐射源的散射回波。这种分类法实际上就是将接收数据与目标特征信号储存库中的各种目标特征资料进行比较。目标特征信号储存库中的资料是事先计算或测量好后存入的。在目标特征信号观测值伴有强噪声的情况下,雷达用这种方法对空间目标作了识别试验。这种目标分类算法是判决理论与严密的目标响应特性计算法相结合的产物。对形状简单的空间目标来说,在雷达捕捉目标后的初始段内,这种方法就能作出可靠的分类。这种分类法不需要事先知道目标的运动参数就能自动估计目标的姿态角。它依次对输入数据进行处理,每输入一个信息,它对目标的分类就准确一步,并同时给出即时判决的置信度。这种分类法还能以拒判度来表达目标特征信号库中是否有这个目标。
The target classification is to determine which hypothetical object’s shape best matches the target feature reflected by the observed data from the scattering field among the M hypothetical objects, which can be said to be an application of the inverse scattering to be perfected. This article describes a method of using the backscattered waveform or target signature observed by a radar receiver to discern targets of various shapes. Although some papers attach great importance to this subject in recent years, this paper is one of the first articles to conduct theoretical analysis of the target automatic classification under certain conditions. This body-type classification method utilizes scattered echoes of two or more radiation sources operating in different carrier frequencies and polarizations. In fact, this classification method compares the received data with various target characteristic data in the target characteristic signal repository. The data of the target characteristic signal repository is calculated or measured in advance and stored. In the case that the target characteristic signal observations are accompanied by strong noise, the radar uses this method to test the space target. This target classification algorithm is a combination of decision theory and strict target response characteristic calculation. For simple-shaped spatial targets, this method makes a reliable classification within the initial segment after the radar captures the target. This classification method does not require prior knowledge of the target’s motion parameters can automatically estimate the target attitude angle. It processes the input data in turn, and each time a message is entered, it accurately classifies the target and at the same time gives the confidence of the instantaneous decision. This classification can also reject the target to express whether the target signature library has this goal.