This is the code for the paper nemed 'Anomaly Detection for Hyperspectral Imagery Based on the Regularized Subspace Method and Collaborative Representation'
Hello, thank you very much for sharing the code about hyperspectral anomaly detection, and would like to ask you a few questions.
In the local_rx algorithm, why did you not subtract the mean vector when calculating the Mahalanobis distance?
CRD and LRX algorithms fill the image with 0 or mirror image, does this have a big impact on the result? How other people deal with when do compare algorithm?