SAR Image Classification Using Greedy Hierarchical Learning With Unsupervised Stacked CAEs

作     者：Sun, Zhensheng Li, Juan Liu, Peng Cao, Weijia Yu, Tao Gu, Xingfa 

作者机构：Univ Chinese Acad Sci Sch Elect Elect & Commun Engn Aerosp Informat Res Inst Beijing 100049 Peoples R China Chinese Acad Sci Inst Remote Sensing & Digital Earth Aerosp Informat Res Inst Beijing 100101 Peoples R China 

出 版 物：《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 (IEEE地学与遥感汇刊)

年 卷 期：2021年第59卷第7期

页      面：5721-5739页

核心收录：

学科分类：0808[工学-电气工程] 1002[医学-临床医学] 08[工学] 0708[理学-地球物理学] 0816[工学-测绘科学与技术] 

基　　金：China's 13th Five-Year Plan Civil Space Pre-Research Project [Y7K00100KJ] 

主　　题：Synthetic aperture radar Remote sensing Feature extraction Training Convolution Speckle Machine learning Convolutional neural network (CNN) hierarchical parameter-oriented learning image classification local pattern recognition stacked convolutional autoencoders (SCAEs) synthetic aperture radar (SAR) 

摘      要：Synthetic aperture radar (SAR) can provide stable data source for earth observation due to its advantages of all day and night, all-weather, and strong penetration. SAR image classification as a fundamental procedure has been proved its great value in plenty of remote sensing applications. Conventional classification algorithms mainly rely on hand-designed features, which are susceptible to widespread coherent speckle noise and geometric distortion in high-resolution SAR images. Inspired by the recent impressive success in data mining and deep learning, a greedy hierarchical convolutional neural network (GHCNN) is developed. It aims at obtaining optimized feature representation, relieving the effect of speckle noise, and promoting the local pattern recognition of geometric distortion in single-polarized SAR image classification. First, a series of convolutional autoencoders (CAEs) is trained in the greedy layer-wise unsupervised strategy. This step provides an unbiased regularizer and a priori distribution derived from large volumes of unlabeled SAR patches. Then, to optimize multiple parameter subspaces globally, several CAEs are coupled together to form a deeper hierarchical structure in a stacked and unsupervised fashion. Afterward, a convolutional network with identical topology inherits the pretrained weights. After supervised finetuning, it realizes class prediction. Synchronously, t-distributed stochastic neighbor embedding (t-SNE) algorithm is applied to monitor the efficiency of feature representation during the training period. Experimental results demonstrate that the proposed method has competitive advantages over involved contrast methods.