Hyperspectral Image Compression and Classification Using PCA and Deep Learning

Abstract

Hyperspectral imaging is a technology that uses a broad spectrum of light to study and evaluate a large volume of information in images, allowing for better data classification. These are high- dimensional data that contain voluminous information. These high-dimensional data include thousands of features in which many unrelated features might influence the quality and accuracy of data. The presence of these irrelevant data results in an increase in computational time, the density of the image, etc., Dimension reduction of Hyperspectral images is the process of removal of redundant and irrelevant data thus reducing the number of input variables used to improve the accuracy and reduce the training time of data. The idea is to implement a solution that compresses the high-dimensional data and classifies them for practical use.