Real-time edge parking lot detection system
- Project
- 17001 I2PANEMA
- Type
- New system
- Description
An AI-based parking lot detection has been developed. Detection algorithms have been transformed to be used on an edge device, accelerated by a tensor processing unit (TPU). The edge device is wirelessly connected to the I2PANEMA backend server using LoRaWAN to provide parking lot occupancy data in real-time. The system can make energy-efficient yet accurate object detection.
- Contact
- University of Rostock
- frank.golatowski@uni-rostock.de
- Technical features
Tensor processing unit, LoRaWAN, Compressed AI models
- Integration constraints
Currently based on a public parking area
- Targeted customer(s)
Logistic Companies, Port owners
- Conditions for reuse
Adaptation to local conditions (camera position, angle, structure of parking area, targeted vehicles)
- Confidentiality
- Public
- Publication date
- 01-01-2022
- Involved partners
- Materna Information & Communications SE (DEU)
- University of Rostock (DEU)