Unsupervised Deep Feature Extraction for Remote Sensing Image Classification

TL;DR

This paper presents an unsupervised deep learning approach using convolutional networks for remote sensing image classification, effectively handling high-dimensional data with limited labels and outperforming traditional methods.

Contribution

It introduces a greedy layer-wise unsupervised pre-training method with sparse representations for remote sensing data, demonstrating superior performance over PCA and current state-of-the-art algorithms.

Findings

Deep architectures outperform single-layer networks in feature abstraction.

The proposed method outperforms PCA, kPCA, and existing aerial classification algorithms.

Single-layer networks are effective with neighboring pixel receptive fields.

Abstract

This paper introduces the use of single layer and deep convolutional networks for remote sensing data analysis. Direct application to multi- and hyper-spectral imagery of supervised (shallow or deep) convolutional networks is very challenging given the high input data dimensionality and the relatively small amount of available labeled data. Therefore, we propose the use of greedy layer-wise unsupervised pre-training coupled with a highly efficient algorithm for unsupervised learning of sparse features. The algorithm is rooted on sparse representations and enforces both population and lifetime sparsity of the extracted features, simultaneously. We successfully illustrate the expressive power of the extracted representations in several scenarios: classification of aerial scenes, as well as land-use classification in very high resolution (VHR), or land-cover classification from multi- and hyper-spectral images. The proposed algorithm clearly outperforms standard Principal Component Analysis (PCA) and its kernel counterpart (kPCA), as well as current state-of-the-art algorithms of aerial classification, while being extremely computationally efficient at learning representations of data. Results show that single layer convolutional networks can extract powerful discriminative features only when the receptive field accounts for neighboring pixels, and are preferred when the classification requires high resolution and detailed results. However, deep architectures significantly outperform single layers variants, capturing increasing levels of abstraction and complexity throughout the feature hierarchy.