A Spectral Sequence-Based Nonlocal Long Short-Term Memory Network for Hyperspectral Image Classification

Abstract

Efficient classification for hyperspectral image (HSI), which assigns each pixel of the image into a specific category, has been a critical research topic in the HSI analysis area. Under the supervised classification settings, the deep learning approaches are very useful for label prediction. However, most deep learning modeling methods can't get the utmost out of spectral information, which is critically important for object interpretation. Consequently, a spectral sequence-based nonlocal long short-term memory (LSTM) network for HSI classification is proposed in this paper. To boost the dominant role of spectral information, the long short-term memory network for sequential data processing is employed. Furthermore, the nonlocal diverse regions are exploited to learn contextual features for stronger discriminative ability. Meanwhile, the attention mechanism is adopted to increase the classification performance. Experiments on Salinas, Indian Pines (IndianP), Pavia University (PaviaU), and Houston University (HoustonU) datasets are implemented. Benefiting from the idea of processing spectral bands as sequence data and nonlocal diverse regions, spectral-spatial information is fully fused to achieve accurate pixel-level classification. It can be demonstrated that the proposed method can achieve better quantitative and qualitative classification results compared with the other state-of-the-art classification methods.

Hyperspectral Imaging Algorithms and Applications: A Review

“…Deep Learning [8][40], including Artificial Neural Networks (ANNs) [15][69], Long Short-Term Memory (LSTM) [43] networks, and Convolutional Neural Networks (CNNs) [11][22][40] [42][59], has revolutionized HSI analysis. These deep learning models can extract intricate spectral and spatial features from hyperspectral data.…”

“…Long Short-Term Memory (LSTM) Networks: LSTMs [43] are a type of recurrent neural network (RNN) designed to capture sequential dependencies. Mathematically, LSTMs include a cell state, hidden state, and gates (forget, input, output), allowing them to remember and forget information…”

“…This method effectively addresses spectral aliasing and class mixing challenges, outperforming other methods with an overall accuracy of 82.11% and demonstrating its value in classifying complex coastal wetland scenarios. In [43], the authors present a novel hyperspectral image classification approach that leverages a spectral sequence-based nonlocal long short-term memory (LSTM) network. This method effectively utilizes spectral information and contextual features, resulting in superior pixel-level classification accuracy compared to existing methods, as demonstrated through experiments on multiple datasets.…”

Efficient residual network using hyperspectral images for corn variety identification

“…Therefore, hyperspectral imaging technology is widely employed in non-destructive testing of crop seed varieties, quality assessment, and vigor analysis ( Ma et al., 2020 ; Zhang et al., 2023a ; Zhang et al., 2024a ). Nevertheless, the high-dimensional nature of hyperspectral data, complex features, noise, and variations in illumination poses challenges for traditional image processing and classification techniques in recognizing hyperspectral corn seed images ( Zhang et al., 2021a ; Ghaderizadeh et al., 2022 ; Huang et al., 2022 ). Hence, this article aims to enhance corn seed hyperspectral image recognition accuracy and efficiency using the efficient residual network (ERNet).…”

Spectral–Spatial Graph Convolutional Network with Dynamic-Synchronized Multiscale Features for Few-Shot Hyperspectral Image Classification

“…By exploiting deeper features with the rich discriminative information, deep learning (DL) has been widely applied for HSI classification [4]. The representative models include the stacked autoencoder (SAE) [17], recurrent neural network (RNN) [18,19], convolutional neural network (CNN) [20], deep belief network (DBN) [21], generative adversarial networks (GAN) [22], and long short-term memory (LSTM) [23].…”