摘要
根据海上舰船目标的特点,提出了一种基于独立分量分析的目标检测算法,应用于高光谱图像海上舰船检测。首先采用基于峰度的快速独立分量分析方法(FastICA)对高光谱图像进行处理,获得统计独立的独立分量影像,然后以偏度为特征度量指标从上述独立分量中选择特征影像,得到舰船目标的检测结果。应用于海上高光谱舰船图像,能够抑制背景中的海浪杂波及舰船尾迹对目标的影响,取得较好的检测效果。实验结果也进一步验证了基于峰度的FastICA算法在高光谱数据分析中的有效性。
A sea fleet target detection approach based on independent component analysis（ICA） is proposed and applied to ship detection in hyperspectral data.Firstly,FastICA approach is used to obtain components that are statistically independent from each other.Then,fleet target feature image is selected from the independent components which has the maximum skewness.Results obtained by applying the new algorithm on data from the operational modular imaging spectrometer（OMIS） show that,the method has an increased efficiency.The experiments confirm the efficiency of FastICA based on kurtosis as an attractive tool for hyperspectral data processing.
出处
《计算机仿真》
CSCD
2008年第9期196-197,299,共3页
Computer Simulation
