Hyperspectral anomaly detection using Modified Principal component analysis reconstruction error

摘要

Nowadays, with the emerging of the new applications in hyperspectral imagery, Anomaly detection has been an interesting topic. In this paper, we propose an anomaly detection algorithm based on Modified PCA reconstruction error. This method, named, Modified Principal component analysis (PCA) Reconstruction-Error-Based Anomaly Detector (MPREBAD), detects anomalies by computing errors associated with reconstructing the original image using PCA projections. An iterative approach is used to filter detected anomalies from the covariance matrix computation in next iterations that lead us to better accuracy for detecting anomalies in subsequent stages. We tested the proposed algorithms with San Diego airport hyperspectral data. MPREBAD results are provided using receiver-operating-characteristic (ROC) curves and intuitive images. Comparing the results of the MPREBAD method with three popular and previous methods, proves the superiority of the proposed method.