Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

Convolutional Autoencoders for Data Compression and Anomaly Detection in Small Satellite Technologies

Created by
  • Haebom

Author

Dishanand Jayeprokash, Julia Gonski

Outline

This paper presents a method for implementing a convolutional autoencoder on a small satellite payload to achieve the dual functions of data compression and in-situ anomaly detection. Advances in small satellite technology have increased the potential for geodetic missions, and machine learning (ML) can enhance the efficiency of satellite data processing. This study demonstrates these capabilities through a disaster monitoring use case using an aerial image dataset of the African continent, suggesting a novel ML-based approach for small satellite applications and the potential for expanding space technology and artificial intelligence in Africa.

Takeaways, Limitations

•
Takeaways:
◦
Presenting an efficient ML-based solution for data compression and anomaly detection in small satellites.
◦
Contribute to the advancement of space technology and artificial intelligence in developing countries, including Africa.
◦
Expanding the potential of small satellites in various fields, including disaster monitoring.
◦
An efficient data processing method using convolutional autoencoders is presented.
•
Limitations:
◦
The high reliance on datasets from the African continent necessitates a review of generalizability to other regions.
◦
Lack of performance verification in actual satellite environments.
◦
Generalization to various types of anomaly detection is needed.
◦
Further research is needed on hyperparameter optimization of convolutional autoencoders.
👍
Made with Slashpage