Abstract
This paper outlines a method for segmentation and classification of ISAR images generated at Sub-THz frequencies for the purposes of space domain awareness. Image segmentation is achieved using statistical region merging. Simulated ISAR imagery is segmented into simple regions, which are used to train a machine learning model to predict the classes within a series of test images. The results indicate that the use of support vector machines for statistical inference has great potential as part of a broader classification process, able to use multiple predictors to draw distinctions between a number of classes.
| Original language | English |
|---|---|
| Title of host publication | 2024 International Radar Symposium (IRS) |
| Publisher | IEEE Computer Society Press |
| Pages | 233-238 |
| Number of pages | 6 |
| ISBN (Electronic) | 9788395602092 |
| ISBN (Print) | 9798350371109 |
| Publication status | Published - 28 Aug 2024 |
| Event | 2024 International Radar Symposium, IRS 2024 - Wroclaw, Poland Duration: 2 Jul 2024 → 4 Jul 2024 |
Publication series
| Name | Proceedings International Radar Symposium |
|---|---|
| Publisher | IEEE |
| ISSN (Print) | 2155-5745 |
| ISSN (Electronic) | 2155-5753 |
Conference
| Conference | 2024 International Radar Symposium, IRS 2024 |
|---|---|
| Country/Territory | Poland |
| City | Wroclaw |
| Period | 2/07/24 → 4/07/24 |
Bibliographical note
Publisher Copyright:© 2024 Warsaw University of Technology.
Keywords
- Classification
- ISAR
- machine learning
- segmentation
- simulation
- space domain awareness
- Sub-THz
- support vector machine
ASJC Scopus subject areas
- Computer Networks and Communications
- Computer Science Applications
- Signal Processing
- Electrical and Electronic Engineering
- Astronomy and Astrophysics
- Instrumentation