针对非线性混合下的亚像元目标检测问题, 提出一种基于核函数的信号空间正交投影方法(KSSP)。该方法作为信号空间正交投影方法(SSP)的非线性推广, 首先将原空间中像元矢量经非线性映射转换到高维特征空间,然后在特征空间中用线性信号空间正交投影进行目标检测。通过核技巧, 核信号空间正交投影不必知道具体的非线性映射形式。经模拟数据与真实高光谱图像数据实验证明, KSSP 方法在目标检测性能上优于SSP, 且对噪声的抑制也有很好的效果。

A kernel-based signature space orthogonal projection (KSSP) technique is proposed for nonlinear subpixel target detection in hyperspectral imagery. As a nonlinear version of the signature space orthogonal projection (SSP), the SSP is adopted in a high-dimension feature space after the pixels of input space are mapped into the feature space via nonlinear mapping. The kernel trick allows the KSSP ignor the actual nonlinear mapping. Experimental results of simulated and real data prove that the proposed KSSP approach outperforms the SSP method in target detection, and improves the robustness to noise.