Superpixel-Oriented Classification of PolSAR Images Using Complex-Valued Convolutional Neural Network Driven by Hybrid Data
Xianxiang Qin,Huanxin Zou,Wangsheng Yu,Peng Wang
DOI: https://doi.org/10.1109/tgrs.2020.3040452
IF: 8.2
2021-12-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:Recently, convolutional neural networks (CNNs) have been successfully developed and used in the classification of polarimetric synthetic aperture radar (PolSAR) images. However, they often suffer from some problems, such as time-consuming, unsatisfactory detail-preservation, and bad effectiveness given limited training samples. Focusing on these problems, we propose a complex-valued CNN (CV-CNN)-based algorithm for PolSAR image classification in this article. On the one hand, a superpixel-oriented (SPO) scheme is employed to reduce the computational cost of the algorithm and preserve image details simultaneously, which takes superpixels instead of single pixels as classification units. In particular, to meet the input requirement of CV-CNN, three alternative methods of superpixel regularization are designed and compared. On the other hand, considering that both measured data (MD) and manually designed polarimetric features (PFs) have their own advantages, the hybrid data (HD) combining them is employed to drive CV-CNN, which is helpful to improve the effectiveness of the algorithm. We perform experiments on three actual PolSAR image data sets acquired by AIRSAR and Radarsat-2 systems as well as a semisimulated data set. The experimental results demonstrate that, compared to conventional pixel-oriented methods, the proposed SPO scheme is much more time-efficient and is also beneficial to detail preservation. Moreover, the CV-CNN driven by HD generally obtains consistently better classification results than that driven by pure MD or manually designed PFs.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics