Speaker
Description
Small satellite (SmallSat) technologies have enhanced the potential and feasibility of geodesic missions, through simplification of design and decreased costs allowing for more frequent launches. These aspects are especially relevant for the development of space programs in Africa, where satellite-based Earth surveys can enhance local science, industry, and communities. SmallSat data acquisition systems can benefit from the implementation of real-time machine learning, for better performance and greater efficiency on tasks such as image processing or feature extraction. However, SmallSat technologies come with limitations on power consumption of electronics, infrequent access, and the bandwidth of data that can be sent back to Earth, along with the need to function in a high radiation environment. This work presents convolutional autoencoders for implementation on the payload of SmallSats, designed to achieve dual functionality of data compression for more efficient off-satellite transmission, and anomaly detection to inform satellite data-taking. This capability is demonstrated for a use case of disaster monitoring using aerial image datasets of the African continent, expanding the landscape of space technology and artificial intelligence in Africa.
| Abstract Category | Computing & 4IR |
|---|