论文部分内容阅读
以往目标识别技术多针对独立目标,但事实上目标有时会以成群的方式出现,目标之间呈现出明显的有规则的阵列分布关系,即目标阵列群。针对该类目标群的识别问题,可以通过提取目标群中子目标之间的空间分布关系的规律与特征,来弥补子目标的其他特征的不确定性。提出一种可有效识别低分辨率环境下(遥感、红外、多光谱等图像中)具有规则阵列分布特征的斑点目标集合的方法。该方法使用ARG模型描述目标阵列群,将目标相互之间的空间分布关系(距离关系)与目标自身的简单属性相结合,通过寻找空间分布关系特征与自身属性的统计规律,即估计特征的联合概率密度分布,利用其分布特点识别目标阵列群。
In the past, the target recognition technology mostly aimed at the independent target, but in fact the target sometimes appeared in groups, and the target showed an obvious regular array distribution relationship, that is, the target array group. In order to solve the problem of identifying the target groups, we can make up for the uncertainty of other features of the sub-targets by extracting the rules and features of the spatial distribution of the sub-targets in the target group. A new method to identify speckle target sets with regular array distribution in low resolution environment (remote sensing, infrared, multispectral and other images) is proposed. This method uses ARG model to describe the target array group, combines the spatial distribution relationship (distance relation) of the targets with the simple attributes of the target itself, and through the statistical law of the relationship between the spatial distribution and its own attributes, that is, the combination of the estimated features Probability density distribution, using its distribution characteristics to identify the target array group.