A Deep Learning Framework for Hyperspectral Image Classification using PCA and Spectral LSTM Networks

Deep learning techniques, like Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, have demonstrated remarkable performance in hyperspectral image (HSI) classification in recent years. This paper proposes a novel framework for the classification of hyperspectral images using a Long Short-Term Memory Networks (LSTM) and Principal Component Analysis (PCA). PCA is basically used to minimize the high-dimensional image data, and then it send to LSTM for classification. LSTM is a RNN that is designed to perform better on sequential analysis of hyperspectral images. This proposed research and analysis of HIS data was performed on the standardized dataset of Pavia University and Kennedy Space Center (KSC). The proposed research framework effectively classifies the HIS images and shows its efficency in comparison to existing state-of-the-art classifiers. Overall, the proposed framework presents a promising approach to the classification of hyperspectral images by combining the strengths of LSTM and PCA. The paper contributes to the literature on hyperspectral image classification by demonstrating the effectiveness of deep learning models and the importance of considering the sequential nature of hyperspectral data.