Deep Learning Hyperspectral Image Classification using Multiple Class-Based Denoising Autoencoders, Mixed Pixel Training Augmentation, and Morphological Operations

Herein, we present a system for hyperspectral image segmentation that utilizes multiple class-based denoising autoen-coders which are efficiently trained. Moreover, we present a novel hyperspectral data augmentation method for labelled HSI data using linear mixtures of pixels from each class, which helps the system with edge pixels which are almost always mixed pixels. Finally, we utilize a deep neural network and morphological hole-filling to provide robust image classification. Results run on the Salinas dataset verify the high performance of the proposed algorithm.

I. Introduction

Deep learning has achieved remarkable results in the areas of computer vision and object recognition. More recently, in the field of remote sensing, hyperspectral image (HSI) classification has had significant performance improvements using deep learning [1]. There are, however, many issues that plague HSI deep learning networks. Among these are the lack of training data, noise, and very high data dimensionality [1].