Hummingbird E203 RISC-V processor core-based traffic flow detection system design

With the development of economy, the number of vehicles on our roads is increasing. Detecting the traffic flow and improving the intelligent transportation system are inevitable choices to improve the traffic condition. A scheme is proposed and designed in the paper. The video is divided into image sequences and then converted into grayscale images, and the noise is removed by binarizing the grayscale images and then median filtering. A virtual line is set in front of the road where the vehicle is moving, and the traffic flow can be detected by judging the change of the gray value of the virtual line. The hardware acceleration unit is designed and deployed based on the idea of hardware-software cooperation, and the FPGA and CPU work together. The experimental results show that this scheme has fast computing and processing speed, high accuracy in detecting traffic flow, and can get more satisfactory results.