This study

based on machine vision technology and in line with the actual needs of intelligent construction site vehicle recognition and vehicle management

by creating a dataset of vehicle types and license plate data for the intelligent construction site

and using the YOLOv11 model

designs an intelligent construction site vehicle recognition algorithm and conducts targeted training. The algorithm accurately identifies vehicle types and license plate colors in video streams

and resolves the problems of multi-plate recognition and vehicle-plate association. Experiments show that in both daytime and nighttime scenes

the proposed algorithm can accurately identify the types of vehicles and multiple license plates in the video

count the number of vehicles in the video

and generate statistical reports.