摘要
提出了一种基于分级特征模型的多尺度目标序贯识别新方法；研究了不变矩特征的尺度变化规律；建立了多尺度目标不变矩的分级特征模型，进而完成多尺度目标的序贯识别．多尺度舰船目标的模拟实验表明，这种基于不变矩分级特征模型的序贯识别算法，能有效地完成多尺度目标的识别，从而证实了分级特征模型符合多尺度目标特征的变化规律，有利于表达多尺度目标的特征，提高目标的检测、识别概率，减少计算复杂性．
new method for sequential recognition of a multiscale target based on the hierarchicalfeautre model is proposed. The scale variation of the characteristic of invariant moment is studied andthe bier archical feature model and sequential recognition method are developed. Simulation experiments on multiscale ships as targets show that the new method works well and the probability of target detection and recognition is improved and the computing complexity reduced.
出处
《华中理工大学学报》
CSCD
北大核心
1996年第2期19-22,共4页
Journal of Huazhong University of Science and Technology
基金
国防科工委基金
关键词
目标识别
多尺度
分级模型
遥感图像
序贯识别
multiscale
target recognitions hierarchical moded
feature modelling
sequential recognition
