基于cICA的红外点目标检测方法

Method of Infrared Point Target Detection Based on cICA

摘要: 针对红外点目标检测的要求,提出了一种基于约束独立分量分析(cICA)的红外点目标检测方法. 该方法只采用单帧扫描图像作为数据源,利用峭度约束的独立分量分析方法,消除了传统独立分量分析(ICA)方法的提取顺序的不确定性,从而直接提取点状目标的位置. 实验结果表明,利用该方法能够有效地进行红外点目标检测.

Abstract: To solve infrared point target detection problems, a novel approach called cICA(constraint independent component analysis) is proposed. It adopts only one image as data source and introduces the constraint of kurtosis to the ICA (independent component analysis) to eliminate the permutation ambiguity of ICA. Experiments utilizing this method showed that it could detect point targets effectively.