采用背景抑制和自适应阈值分割的高光谱异常目标检测

摘要

高光谱图像小目标异常检测存在着大面积背景异常的干扰，直接采用传统的阈值分割方法会产生较高的虚警。针对核RX异常检测算法存在较大面积背景干扰的现象，结合形态学滤波方法提取大面积背景杂波干扰并对其进行抑制，滤除背景干扰。然后采用自适应阈值方法对处理后的灰度值图像进行异常目标的分离。仿真实验结果表明，该方法较好地实现了对大面积背景干扰的抑制和对异常目标的保持，改善了现有的核RX算法用于高光谱异常检测的性能。%Hyperspectral small target anomaly detection algorithms have the problem of a large area of background interference, and using the traditional threshold segmentation yields a high rate of false alarms. In order to better handle a large area of background interference in a Kernel RX detector, this paper developed a new morphology fil⁃ter method to effectively extract and suppress large cluttered background areas. Further an adaptive threshold meth⁃od was used to segment anomaly targets on grayscale images. Simulation experiments show that the proposed method provides very good anomaly detection with the advantage of large area cluttered background suppression. This dra⁃matically improves the performance of the Kernel RX detector.