Self-Supervised Hyperspectral Anomaly Detection Based on Finite Spatialwise Attention

Hyperspectral anomaly detection (HAD) is of great value in both practical and theoretical terms. However, due to the lack of available semantic labels, previous works mainly relied on unsupervised or semisupervised methods to construct learning models, which inevitably lacked semantic guidance and led to limited AD effectiveness. Besides, few previous methods jointly mine spectral and spatial global dependencies, which limits their effectiveness in practical scenarios. To address the above-mentioned problems, we design a novel self-supervised HAD method, named the Self-Supervised HAD method based on the finite spatialwise attention (FSA). The core of the proposed method is the designed self-supervised HAD transFormer (SSHADFormer). It explores the specific spectral attributes of hyperspectral images (HSIs) to reconstruct background HSI from a given RGB image, which solves the difficulty of acquiring semantic information and enhances the agility of AD models. In addition, we propose an FSA mechanism. The mechanism mines the cluster structure of the background spectrum in a data-driven manner, enhancing the discriminative ability between background and anomalous targets while avoiding anomalous target interference during training. Extensive experiments on six public datasets demonstrate the effectiveness and agility of the proposed method.