Hyperspectral Classification Based on Lightweight 3-D-CNN With Transfer Learning

TL;DR

This paper introduces a lightweight 3-D CNN model with transfer learning strategies to improve hyperspectral image classification, especially with limited training samples, achieving competitive results across different sensors.

Contribution

The paper proposes a novel deep 3-D CNN architecture with fewer parameters and introduces two transfer learning strategies that do not require same-sensor data, enhancing hyperspectral classification.

Findings

The lightweight 3-D-CNN outperforms traditional models in classification accuracy.

Transfer learning strategies improve performance with limited samples.

Model achieves competitive results across multiple datasets.

Abstract

Recently, hyperspectral image (HSI) classification approaches based on deep learning (DL) models have been proposed and shown promising performance. However, because of very limited available training samples and massive model parameters, DL methods may suffer from overfitting. In this paper, we propose an end-to-end 3-D lightweight convolutional neural network (CNN) (abbreviated as 3-D-LWNet) for limited samples-based HSI classification. Compared with conventional 3-D-CNN models, the proposed 3-D-LWNet has a deeper network structure, less parameters, and lower computation cost, resulting in better classification performance. To further alleviate the small sample problem, we also propose two transfer learning strategies: 1) cross-sensor strategy, in which we pretrain a 3-D model in the source HSI data sets containing a greater number of labeled samples and then transfer it to the target HSI data sets and 2) cross-modal strategy, in which we pretrain a 3-D model in the 2-D RGB image data sets containing a large number of samples and then transfer it to the target HSI data sets. In contrast to previous approaches, we do not impose restrictions over the source data sets, in which they do not have to be collected by the same sensors as the target data sets. Experiments on three public HSI data sets captured by different sensors demonstrate that our model achieves competitive performance for HSI classification compared to several state-of-the-art methods