Demo of “Vehicle counting with Ultra96 FPGA using the DNNDK”. We have used the Xilinx-DNNDK platform for machine learning acceleration. The classification and detection network is implemented on the DNNDK platform while the counting algorithm run on multi-core CPU of MPSoC FPGA of Ultra96.
Standalone Ultra96 FPGA can give real time FPS on the tracking and counting application based on Machine Learning. This implementation can count the vehicles, pedestrians and motorcycle. For the bidirectional and other traffic scenario [example: flat road], we will release demo video soon. The same algorithm with updates work on those scenario to. This video demo is an early demo of our upcoming product.
Ultra96 FPGA is $249 FPGA board having Multiple ARM CPU and FPGA fabric. Due to its MPSoC architecture, it can accelerate the Machine Learning task for different applications as ADAS, Surveillance and Computer Vision. For details of Ultra96 can be fount at Xilinx or Avnet’s website.
Watch the Demo at our YouTube Channel: