An Open Source Desktop Application for Classification of Remote Sensing Data

In this paper, we propose a desktop application for the analysis and classification of multi and hyperspectral remote sensing images through a simple and intuitive interface. The user can load the data input in raw or MATLAB formats and then a detailed inspection of the different areas or bands of the image can be performed as well as a comparison of pixels based on their spectra. Regarding the classification ability, different spatial-spectral classification schemes have been implemented by using a pipeline consisting of five configurable stages. The modular implementation makes it easy to introduce in the future new classification schemes, new stages to the existing ones, and new algorithms for each stage. It also allows the easy comparison among different classification schemes based on numerical and graphical results. The application is licensefree, runs on the Linux operating system and has been developed in C language using the GTK library to build the graphical user interfaces, as well as other free libraries.

keywords: remote sensing, spatial-spectral classification, segmentation, desktop application, hyperspectral data