 This paper proposes a novel approach for hyperspectral anomaly detection using a two-branched 3D convolutional autoencoder and spatial filtering. The proposed method was evaluated on both airborne and satellite-borne hyperspectral imagery datasets. The results showed that the proposed method outperformed existing methods in terms of AUC scores. This article was authored by Shui LV, Si Wei Zhao, Dandan Li, and others.