BYOL框架下的自监督高光谱图像分类  

BYOL-based self-supervised learning for hyperspectral image classification

作　　者：韩希珍 蒋振刚[1] 刘媛媛 赵建 孙强[3] 刘建卓[3] HAN Xizhen;JIANG Zhengang;LIU Yuanyuan;ZHAO Jian;SUN Qiang;LIU Jianzhuo(Changchun University of Science and Technology,Changchun 130000,China;Suzhou East Clotho Opto-Electronic Technology Co.Ltd.Zhangjiagang 215600,China;Changchun Institute of Optics,Fine Mechanics and Physics,Chinese Academy of Sciences,Changchun 130033,China;Suzhou Institute of Biomedical Engineering and Technology,Chinese Academy of Sciences,Suzhou 215000,China)

机构地区：[1]长春理工大学计算机科学技术学院,吉林长春130000 [2]苏州东方克洛托光电技术有限公司,江苏张家港215600 [3]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [4]中国科学院苏州生物医学工程技术研究所,江苏苏州215000

出　　处：《红外与激光工程》2024年第10期244-259,共16页Infrared and Laser Engineering

摘　　要：高光谱图像可以获取波段连续的图谱合一的立体数据,其具有丰富的图谱信息,能区分不同物质的类别,被广泛应用于各种遥感勘测领域。但在实际中高光谱图像的标注需要耗费大量的人力、财力和时间,可用的标注样本数量较少,难以通过训练来获得准确的分类结果,所以针对于只有少量标记样本的高光谱图像分类是一个挑战。近年来,自监督学习(Self-supervised Learning,SSL)已成为一种有效的方法,可以减少高光谱图像分类对昂贵的数据标注的依赖。SSL方法通过学习在同一图像的不同视图之间产生的潜在特征,在自然图像分类中取得了较高的分类精度。为了探索SSL方法在高光谱图像分类中的潜力,一种Bootstrap Your Own Latent(BYOL)框架下的自监督高光谱图像分类方法(BSSL)被提出。该方法通过引用自监督的图像特征学习框架BYOL,可以不需要负样本对,利用空间光谱相似的同类样本对进行网络训练及参数微调,提取到更具判别性特征。具体来说,该方法主要包括四个部分:BYOL的预训练、超像素聚类、基于“相似对”的BYOL的再训练和最终分类。为了验证该方法的有效性,在三个公开数据集上进行测试,并与五种先进的无监督、自监督分类方法SuperPCA、S3PCA、ContrastNet、SSCL和N2SSL进行对比,在Indian Pines和Salinas数据集上,BSSL方法的总体分类精度(OA)、平均分类精度(AA)、Kappa系数、召回率(recall)和f1分数(f1-score)都取得了更优值。其中在Indian Pines数据集上,OA分别比SuperPCA,S3PCA,ContrastNet,SSCL和N2SSL提高了1.32%,1.05%,5.68%,3.12%和1.27%。而在University of Pavia数据集上,BSSL方法表现没有那么出色,但在综合分类性能上也表现最优。这表明BSSL方法更适用于地物区域面积较大且分布较集中的场景,因为这对于超像素聚类来说更友好。Objective Hyperspectral images can acquire continuous spectral bands integrated into a three-dimensional data set,which is rich in spectral information and capable of distinguishing different types of materials.They are widely used in various remote sensing surveying fields.However,with the rapid development of deep learning,hyperspectral image classification has made great progress,but still faces some difficulties.The annotation of hyperspectral images requires a significant amount of manpower,financial resources,and time.And the number of available labeled samples is limited,making it difficult to achieve accurate classification results through training.Therefore,the classification of hyperspectral images with only a small number of labeled samples is a challenge.Researching hyperspectral image classification in scenarios with few samples is of great practical significance for promoting the application of hyperspectral technology.Methods In recent years,Self-supervised Learning(SSL)has emerged as an effective approach to reduce the reliance on costly data annotation for hyperspectral image classification.SSL methods have achieved high classification accuracy in natural image classification by learning latent features that arise from different views of the same image.To explore the potential of SSL methods in hyperspectral image classification,a self-supervised hyperspectral image classification method under the Bootstrap Your Own Latent(BYOL)framework,referred to as BSSL,has been proposed.This method leverages the self-supervised image feature learning framework of BYOL,which can train the network and fine-tune parameters without the need for negative sample pairs,utilizing spatial-spectral similar pairs of the same category to extract more discriminative features.Specifically,the method mainly includes four parts:pre-training of BYOL,superpixel clustering,re-training of BYOL based on similar pairs,and final classification.In the BYOL model,the encoder employs a spectral-spatial transformer network to extract

关 键 词：高光谱图像 自监督分类 BYOL 超像素聚类 空-谱联合 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]