Design Of Highway Vehicle Overload Control Device Based On HiSilicon Hi3516DV300

The current methods for overload control on highways are traditional and ineffective.resulting in low overall efficiency.In order to address these issues,this paper designs an allweather.full-coverage,high-efficiency,low-cost,and easy-to-maintain overweight control device with the high-performance image processing chip HiSilicon Hi3516DV300 as the core processor.This paper studies the method of developing media functions using the Linux operating system and the chip application programming interface of HiSilicon Hi3516DV300.achieving video encoding and recording capabilities.It also researches a real-time stitching method based on region of interest(ROI)for side image stitching of vehicles.Furthermore,it improves the YOLOv5 object detection algorithm and achieves 98.10%accuracy in axis recognition of vehicles.The paper also explores the combination of YOLOv5 and CRNN algorithms for license plate recognition,accurately identifying vehicle license plate numbers.The designed overweight control device was tested at the entrance lane of the expressway.The device actively captures front-end images,rear-end images,side stitching images,axis recognition results,and license plate recognition results when vehicles pass by.The complete data can be uploaded to the data receiver through the HTTP protocol.