Novel Semi-Supervised Hyperspectral Image Classification Based on a Superpixel Graph and Discrete Potential Method

Yifei Zhao,Fenzhen Su,Fengqin Yan

doi:10.3390/rs12091528

Abstract

Hyperspectral image (HSI) classification plays an important role in the automatic interpretation of the remotely sensed data. However, it is a non-trivial task to classify HSI accurately and rapidly due to its characteristics of having a large amount of data and massive noise points. To address this problem, in this work, a novel, semi-supervised, superpixel-level classification method for an HSI was proposed based on a graph and discrete potential (SSC-GDP). The key idea of the proposed scheme is the construction of the weighted connectivity graph and the division of the weighted graph. Based on the superpixel segmentation, a weighted connectivity graph is constructed usingthe weighted connection between a superpixel and its spatial neighbors. The generated graph is then divided into different communities/sub-graphs by using a discrete potential and the improved semi-supervised Wu–Huberman (ISWH) algorithm. Each community in the weighted connectivity graph represents a class in the HSI. The local connection strategy, together with the linear complexity of the ISWH algorithm, ensures the fast implementation of the suggested SSC-GDP method. To prove the effectiveness of the proposed spectral–spatial method, two public benchmarks, Indian Pines and Salinas, were utilized to test the performance of our proposal. The comparative test results confirmed that the proposed method was superior to several other state-of-the-art methods.

Highlights

With the rapid development of hyperspectral remote sensing, computers, and communication technology, a large number of hyperspectral data containing hundreds of narrow spectral bands and rich spatial information have been collected in the past few decades
To address the aforementioned problems, this study suggests a semi-supervised superpixel-level hyperspectral image (HSI) classification method based on a graph and discrete potential (SSC-GDP), aiming at classifying HSIs accurately and quickly
To evaluate the effectiveness of the proposed SSC-GDP algorithm, we took two Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) datasets, namely an Indian Pines image and a Salinas image, as examples in our experiments. These two benchmark images are widely used to test the performance of HSI classification methods

Summary

Introduction

With the rapid development of hyperspectral remote sensing, computers, and communication technology, a large number of hyperspectral data containing hundreds of narrow spectral bands and rich spatial information have been collected in the past few decades. A hyperspectral image (HSI) with detailed spectral information and abundant spatial texture significantly improves its identification ability for land cover and is widely applied to various fields, such as ocean exploration, environmental monitoring, urban planning, and others [1,2,3]. It is still a challenging problem regarding how to classify HSI quickly and accurately in the field of remote sensing because of its characteristics of high dimension, a large amount of data, and massive noise pixels. The volume of the HSI decreases sharply when reducing the dimensionality of pixels

Methods

Results

Discussion

Conclusion