摘 要：本文基于商用车的应用场景，针对车道线被遮挡和光线变化的特定环境问题，提出一种改进的UFLD算法。该算法能够较好地检测出完整车道线，在保证快速检测的同时，对多种不同场景的检测准确率高于其他同类网络，能更好地适用于商用车的智能驾驶环境，对乘用车运行环境也具有一定的移植和应用价值。
关键词：智能驾驶；深度学习；车道线检测；行分类；均值滤波
中图分类号：U469.72；TP391.4 文献标志码：A DOI：10.15917/j.cnki.1006-3331.2025.03.001
Research on Intelligent Sensing Lane Detection Algorithm for Commercial Vehicles
CHEN Zhen,LIANG Fengshou,LAN Lu,CHEN Qiyuan,ZHANG Shiyu
Abstract: Aiming at specific environmental problems such as lane occlusion and light change,this paper proposes an improved UFLD algorithm based on the application scenario of commercial vehicles.This algorithm can detect complete lane lines well,ensuring fast detection while maintaining higher accuracy than similar networks in detecting different scenarios.It can better adapt to the intelligent driving environment of commercial vehicles and have transplant and application value for passenger car operating environments.
Key words: intelligent driving; deep learning; lane detection; row classification; mean value filtering
