高光谱图像海上舰船目标检测方法研究

Method for Ship Detection in Hyperspectral Image

提出了一种基于光谱异常检测的高光谱海上舰船目标自动检测算法。首先提取图像中的海域,确定舰船检测的目标区;然后在通过主成分变换降低高光谱图像数据维数后,由RX算法检测出目标区中的异常点;最后提取具有一定几何特征的面状异常目标,即为疑似舰船目标。实验结果表明:算法能够有效实现复杂背景下的舰船目标检测。

A method based on spectral anomaly for the ship target detection in hyperspectral image was proposed. Firstly,the area covered by seawater was picked out and confirm the target searching area for ships in the sea. Then,principal components analysis（ PCA） was used to reduce the dimension of hyperspectral image. RX detection was carried out on the PCA translated image to find the spectral anomaly pixels. At last,anomaly pixels clustered in a certain shape were detected as the ship targets. Experiment results indicate that the proposed method is efficient for sea ship target detection in complex background.