Big Data Analysis: Hyperspectral Image Processing for Agriculture Applications

Abstract

Hyperspectral imaging is employed in a broad array of applications. The usual idea in all of these applications is the requirement for classification of a hyperspectral image data. Where Hyperspectral data consists of many bands - up to hundreds of bands - that cover the electromagnetic spectrum. This results in a hyperspectral data cube that contains approximately hundreds of bands - which means BIG DATA CHALLENGE. In this paper, unsupervised hyperspectral image classification algorithm, in particular, Iterative Self-Organizing Data Analysis Technique Algorithm (ISODATA) algorithm used to produce a classified image and extract agricultural information, using ENVI (Environment of Visualizing Images) that is a software application utilized to process and analyze geospatial imagery. The study area, which has been applied on is Florida, USA. Hyperspectral dataset of Florida was generated by the SAMSON sensor. In this paper, the performance was evaluated on the base of the accuracy assessment of the process after applying Principle Component Analysis (PCA) and ISODATA algorithm. The overall accuracy of the classification process is 75.6187%.