Artificial intelligence and image processing
Real-time detection and classification algorithms on embedded platforms; camera interfaces and accelerated inference.
Image-processing and learning-based algorithms have to run in real time on embedded hardware in the field. Model selection, quantisation and the use of accelerators (GPU, NPU, FPGA) are decided by the power and latency budget of the target platform.
Detection and classification algorithms on embedded platforms
Porting object detection and classification models to embedded GPU / NPU platforms (TensorRT and similar inference engines), balancing speed and accuracy through quantisation, OpenCV-based pre-processing and camera interfaces (GigE, MIPI CSI, USB3). Infrared imaging and time alignment of multi-sensor data fall within this scope.
Related products and services
Embedded Software and Application Development
- Firmware design for FPGAs and microcontrollers
- Computer interface software design
Laser Rangefinder
- 905 nm wavelength
- 3 km measurement range
- 0.6 m measurement accuracy
- 6× zoom
Have a requirement in this area? Tell us about it.
Get in touch