基于光谱角背景纯化的高光谱异常检测算法

为了解决利用高光谱图像进行异常检测时结果不准确、虚警率较高的问题，提出了一种基于光谱角背景纯化的异常检测算法。该算法以局部RX算法为基础，根据光谱角距离分离出内外窗口间背景像元中的异常成分，得到纯化后的背景像元，然后进行异常检测。为验证算法的有效性，选取了两组机载可见光/红外光成像光谱仪真实高光谱数据进行仿真实验，并与经典的全局RX、局部RX算法进行对比。结果表明，与局部RX算法相比，该算法在两组数据下的曲线下面积分别提高了0.0317和0.0053。这些结果为下一步的研究方向提供了参考。

Hyperspectral anomaly detection algorithm based on spectral angle background purification

In order to solve the problem of inaccurate results and high false alarm rate when using hyperspectral image for anomaly detection, an anomaly detection algorithm based on spectral angle background purification was proposed. With this algorithm, which is based on the local RX algorithm, the the anomalous components in the background pixels between the inner and outer windows could be separated according to the spectral angular distance. The purified background pixels were then obtained, following which the anomaly detection could be performed. In order to verify the effectiveness of the algorithm, two sets of airborne visible infrared imaging spectrometer real hyperspectral data were selected for simulation experiments. The corresponding data was then compared with that of the classical global RX and local RX algorithms. The results show that, the area under the curve of the two sets of data is respectively increased by 0.0317 and 0.0053 compared with that of the local RX algorithm. These results provide a reference for the next research direction.