This demo captures a vehicle registration plate, using machine learning neural network algorithms to identify. The Arm® Cortex®-M4 core supports the DSP instruction to speed up algorithms. Recognized information can be transmitted to the cloud through Wi-Fi or NB-IoT to realize the image recognition on the edge device.

• Features
  

  -	Hardware/ Software driven CMOS sensor
  -	QVGA: 320 * 240
  -	Supports deep neural network and convolutional neural network
  -	Provides multiple image processing algorithms
  -	Multiple wireless transmission methods (NB-IoT and Wi-Fi)

• Specifications
  

  M480 (32-bit Arm® Cortex®-M4F)
  -	Up to 192 MHz core speed, built-in DSP, image recognition in 0.5 second
  -	Camera capture interface supports hardware binarization
  -	Up to 160 KB SRAM for QVGA image capture

• System Block Diagram