# Space Object Identification and Classification from Hyperspectral Material Analysis

**Authors:** Massimiliano Vasile, Lewis Walker, Andrew Campbell, Simao Marto, Paul Murray, Stephen Marshall, Vasili Savitski  
**arXiv:** [2308.07481](https://arxiv.org/abs/2308.07481)  
**Date:** 2023-08-14  
**Summary page:** https://www.summarizepaper.com/en/paper/2308.07481v1/

## Key points

- Data processing pipeline designed to extract information from hyperspectral signature of unknown space objects
- Methodology proposed for determining material composition of space objects using single pixel images
- Two techniques used for material identification and classification: machine learning and least square match with library of known spectra
- Utilization of supervised machine learning algorithm to classify object into specific categories based on detected materials
- Investigation of behavior of material classification methods under non-ideal circumstances such as weathered materials or lacking certain materials in training library
- Preliminary results on identification and classification of space objects using this approach
- Implications for various applications including space exploration and surveillance
- Authors include Massimiliano Vasile, Lewis Walker, Andrew Campbell, Simao Marto, Paul Murray, Stephen Marshall, and Vasili Savitski

## Summary

This paper presents a data processing pipeline designed to extract information from the hyperspectral signature of unknown space objects. The authors propose a methodology that determines the material composition of space objects using single pixel images. They employ two techniques for material identification and classification: one based on machine learning and the other based on a least square match with a library of known spectra. To classify the object into specific categories, a supervised machine learning algorithm is utilized which detects materials present on the object. The researchers also investigate how the material classification methods behave under non-ideal circumstances such as weathered materials or when the training library lacks certain materials that are present in the observed object. The paper provides preliminary results on the identification and classification of space objects using this approach and offers insights into how to accurately determine material composition from hyperspectral signatures even in challenging scenarios. The findings have implications for various applications including space exploration and surveillance. The authors of this study include Massimiliano Vasile, Lewis Walker, Andrew Campbell, Simao Marto, Paul Murray, Stephen Marshall and Vasili Savitski. This paper falls under categories such as astro-ph.IM (Astrophysics - Instrumentation and Methods for Astrophysics), cs.CV (Computer Science - Computer Vision and Pattern Recognition), eess.IV (Electrical Engineering and Systems Science - Image and Video Processing), physics.space-ph (Physics - Space Physics) as well as I.2.0; I2 10 (Artificial Intelligence - General; Vision and Scene Understanding).