Machine Learning solutions at Space Apps
Artificial Intelligence
Artificial Intelligence at Space Apps
Statistical analysis
Data science methods
We use the toolkit of traditional statistical mathematics to solve challenging categorisation and data processing tasks.
Machine learning
ANNs CNNs RNNs and more
For image recognition, video analysis, time series analysis and many other use cases.
Machine Learning Process
AI processing in the SA Hybrid Cloud
1Data Acquisition
The data comes from a sensor network or is produced by ESA satellites.
2Pre-Processing
Cleaning the data is needed to prepare for deeper analysis.
3AI Algorithms
The selected analysis will be performed on the data on the fly, or on a regular basis.
4Database
The results will be present in the database for further processing.
Edge Processing
In limited and special cases, processing has to happen before the data enters the cloud.
Typically Agricultural IoT, High volume data handling and other remote locations.
System on Chip processing
Small neural nets can be deployed on tiny microcontrollers. For many use cases they are the best solution to transmit only a few bytes of results.
- Low power
- Low data
- Cost effective
System on Module processing
Medium processing capacity is available on larger deployable systems. They provide much larger capacities and services. The user can have the freedom to log into the device to see processes himself.
- Medium processing capacity
- Continuous operation
Tensor Processing Units
For many services, the continuous data stream requires a dedicated processing unit to speed up neural networks. These are useful solutions for video streams, where the network is limited, and real time evaluation is needed.
- Fast processing
- Low bandwidth
- High power consumption
- Contact Us
If you are interested in our AI services, reach out to us!
© 2023