License plate recognition system

This solution uses NuMicro® M480 microcontroller, connected to an external CMOS sensor, to capture the license plate, and then use Convolutional Neural Network (Convolutional Neural Network) to identify it. It uses Arm® Cortex®-M4 as the core and uses DSP instructions to accelerate the algorithm. After identification, the information It can be transmitted to the cloud through Wi-Fi or NB-IoT to realize endpoint image recognition function.

Solution features
- Support hardware/software driven CMOS sensor
- QVGA: 320 * 240
- Support deep neural network and convolutional neural network
- Provide a variety of image processing algorithms
- Wireless transmission and cloud-connected

MCU Features
M480 (32-bit Arm® Cortex ®-M4F)
- Operating frequency up to 192 MHz, built-in DSP, can complete image recognition in 0.5 seconds
- Hardware image capture and hardware binarization function
- Up to 160 KB SRAM, can support QVGA image capture