A Hyperspectral Image Classification Method Based on Multi-Discriminator Generative Adversarial Networks.

Haiyun Liu,Hongmin Gao,Jiawei Wang,Mingxia Wang,Dan Yao,Zaijun Hua,Chenming Li

doi:10.3390/s19153269

Abstract

Hyperspectral remote sensing images (HSIs) have great research and application value. At present, deep learning has become an important method for studying image processing. The Generative Adversarial Network (GAN) model is a typical network of deep learning developed in recent years and the GAN model can also be used to classify HSIs. However, there are still some problems in the classification of HSIs. On the one hand, due to the existence of different objects with the same spectrum phenomenon, if only according to the original GAN model to generate samples from spectral samples, it will produce the wrong detailed characteristic information. On the other hand, the gradient disappears in the original GAN model and the scoring ability of a single discriminator limits the quality of the generated samples. In order to solve the above problems, we introduce the scoring mechanism of multi-discriminator collaboration and complete semi-supervised classification on three hyperspectral data sets. Compared with the original GAN model with a single discriminator, the adjusted criterion is more rigorous and accurate and the generated samples can show more accurate characteristics. Aiming at the pattern collapse and diversity deficiency of the original GAN generated by single discriminator, this paper proposes a multi-discriminator generative adversarial networks (MDGANs) and studies the influence of the number of discriminators on the classification results. The experimental results show that the introduction of multi-discriminator improves the judgment ability of the model, ensures the effect of generating samples, solves the problem of noise in generating spectral samples and can improve the classification effect of HSIs. At the same time, the number of discriminators has different effects on different data sets.

Highlights

Remote sensing technology is a kind of long-distance earth observation technology which was raised in the 1960s
In order to solve these problems and improve the performance of the model, this paper introduces the scoring mechanism of multi-discriminator cooperative work and proposes a multi-discriminator generative adversarial networks (MDGANs), which is applied to three hyperspectral data sets to complete semi-supervised classification
We study the general process of semi-supervised classification of hyperspectral images based on MDGANs

Summary

Introduction

Remote sensing technology is a kind of long-distance earth observation technology which was raised in the 1960s. Dropout, which is widely used in CNN to prevent model over-fitting, can be used in the integration of multiple discriminators [31]. A robust scoring system generally takes into account the impact of outliers (such as the lowest and highest scores) and removes them To this end, such a mechanism was introduced into the training process of MDGANs, that is, part of discriminators were filtered within each training cycle and the remaining discriminators participated in the voting score. In this way, in each training cycle, we dynamically integrate discriminators to guide the generator to generate spectral samples, so that MDGANs can learn a series of pattern features and avoid the phenomenon of pattern collapse in the training process. In order to make the model more generalized and comparative, the classifier we added after the discriminator did not choose CNN but the general softmax multi-classifier, so as to better compare with the original GAN and other classical classification algorithms

Methods

Results

Conclusion