海面溢油对海洋环境会生成许多有害物质, 导致位于食物链底端的浮游生物大量死亡, 进而危害整个海洋生态系统。 高光谱遥感可以获取溢油和海水的空间分布信息和光谱分布信息, 利用高光谱图像进行溢油检测具有更好的检测性能。 针对高光谱图像的海面溢油检测过程中需要人工辅助的缺点, 提出一种基于自适应匹配滤波的高光谱图像海面溢油自动检测方法。 首先, 利用场景中溢油的C—H键光谱特性, 设计一种溢油参考光谱特征提取模型, 获取溢油检测过程中的参考光谱。 其次, 依据海水、 云以及溢油光谱反射特性的差异, 选取海水、 云和溢油差异最大的波段, 分割海水像素, 计算溢油检测过程中的背景参数。 最后, 将目标检测中经典的自适应匹配滤波方法进行适当的改进应用于溢油的检测。 将该方法应用于实际墨西哥湾airborne visible infrared imaging spectrometer（AVIRIS）高光谱遥感影像的海面溢油的检测, 结果表明该方法具有计算速度快, 不需要人工辅助等方面的特点, 可用于海面溢油的自动检测。

To reduce artificial auxiliary works in oil spills detection process, an automatic oil spill detection method based on adaptive matched filter is presented. Firstly, the characteristics of reflectance spectral signature of C—H bond in oil spill are analyzed. And an oil spill spectral signature extraction model is designed by using the spectral feature of C—H bond. It is then used to obtain the reference spectral signature for the following oil spill detection step. Secondly, the characteristics of reflectance spectral signature of sea water, clouds, and oil spill are compared. The bands which have large difference in reflectance spectral signatures of the sea water, clouds, and oil spill are selected. By using these bands, the sea water pixels are segmented. And the background parameters are then calculated. Finally, the classical adaptive matched filter from target detection algorithms is improved and introduced for oil spill detection. The proposed method is applied to the real airborne visible infrared imaging spectrometer (AVIRIS) hyperspectral image captured during the deepwater horizon oil spill in the Gulf of Mexico for oil spill detection. The results show that the proposed method has, high efficiency, does not need artificial auxiliary work, and can be used for automatic detection of marine oil spill.