基于多元统计特性异常的盲元检测算法

Algorithm of blind-pixel detection based on multi-statistical characteristic abnormity

针对红外焦平面阵列(IRFPA)探测器盲元和非均匀性导致系统性能降低的问题,首先建立红外焦平面阵列的多元正态分布时序噪声模型,将盲元看作是不符合模型统计分布特性的异常像素点,游离于多元正态分布超椭球之外。其次对序列图像进行主成分分解,将统计距离与等分线空间角作为异常像素检测的统计判据。最后,利用红外热像仪采集了黑体的序列图像数据,用于盲元检测算法的性能验证,实验结果证明该算法的有效性。

Aiming at the severe problems of system performance degradation that was caused by the blind pixels and the nonuniformity which exists in Infrared Focal Plane Array （IRFPA）, multivariate normal distribution sequence noise model of IRFPA was set up firstly, and blind-pixel was regarded as abnormity pixel for the unconformiting mold statistical distribution characteristic and it was dissociating the multivariate normal distribution ellipsoid. Then the principal component was applied to sequence pattern, and the statistical distance and bisectrix spatial angle are regard as statistical criterion of abnormity pixels detection. Finally, to test and verify the performance of the method on taken full advantage of thermal infrared imager to do sequential multiple frames of noise data collection. The algorithm was applied to actual blind pixel detection of uncooled IRFPA and the validity of the algorithm was proved by the experimental result.