期刊
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
卷 61, 期 -, 页码 -出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2023.3269097
关键词
Anomaly detection; Dictionaries; Matrix decomposition; Hyperspectral imaging; Kernel; Detectors; Atmospheric modeling; Antinoise; hierarchical structure; hyperspectral anomaly detection (AD); incoherent constraint; low-rank representation (LRR)
This paper proposes an antinoise hierarchical mutual-incoherence-induced discriminative learning method for hyperspectral image anomaly detection. By applying structural incoherence constraint and first-order statistic constraint, the separability between background and anomalies is improved. A mixed noise model is used to enhance antinoise performance. Experimental results show that the proposed method outperforms other methods.
Hyperspectral image (HSI) anomaly detection (AD) generally considers background pixels as low-rank distribution and anomaly pixels as sparse distribution. However, it is usually difficult to construct an accurate background dictionary for the background pixels composed of different land covers, and completely separate sparse anomaly targets from various complicated background pixels with complex mixed noise interference. To address these challenges, we propose an antinoise hierarchical mutual-incoherence-induced discriminative learning (AHMID) method for the AD of HSI. A structural incoherence constraint is designed to constrain the inherent dissimilarity and incoherence between the background and anomalies for improving their separability. Then, a first-order statistic constraint is conducted on targets to enhance the anomaly representation, and a decentralization constraint is used on the background to suppress the background representation. Meanwhile, a mixed noise model is constructed by $\ell _{1,1}$ -norm and Frobenius norm to improve the antinoise performance. Finally, a hierarchical alternative strategy is developed to gradually optimize the background and anomalies. Experiments on six HSI AD datasets show that the proposed method outperforms a few state-of-the-art AD algorithms. Code: https://github.com/HalongL/HAD-AHMID
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据