有监督的高光谱图像伪装目标检测方法

针对伪装目标检测问题，提出了一种有监督的高光谱伪装目标检测方法。以植被型伪装目标为研究对象，在分析伪装材料与绿色植被光谱之间特性的基础上，先通过光谱重排、光谱微分以及光谱差异性增强处理，对植被型伪装材料与真实植被（背景）之间的光谱差异进行放大，然后利用主成分分析（PCA）变换进行降维，从而实现了一种适用于大面积植被型伪装目标的高光谱检测方法。实验结果表明，该检测方法在检测时间和检测效果上要优于基于加权的约束能量最小化法（WCM-CEM）和基于非监督目标生成处理的正交子空间投影法（UTGP-OSP）。

Supervised method for hyperspectral image camouflage target detection

Aiming at camouflage target detection problem, a supervised method for hyperspectral image camouflage target detection was proposed. The plant camouflage targets were taken as study objects, and then based on the spectral characteristics analysis of camouflage materials and plants, camouflage materials and plant's spectral differences were magnified through spectrum rearrangement, spectral derivative and spectrum difference enhancement. Then, principal components analysis(PCA) was used for dimensionality reduction, thus a detection method for big camouflage target in hyperspectral image was realized. The experimental result shows that the method outperforms weighted correlation matric-constrained energy minimization(WCM-CEM) and unsupervised target generation process-orthogonal subspace projection(UTGP-OSP) both in the detection time and detection result.