无人驾驶环境中基于半监督模型的道路裂痕检测算法  

Road Crack Detection Algorithm Based on Semi Supervised Model in Autonomous Driving Environment

作　　者：蔡臻[1] 李锋[1] 

机构地区：[1]广东交通职业技术学院信息学院,广东广州510650

出　　处：《工业控制计算机》2025年第2期109-111,共3页Industrial Control Computer

基　　金：2022年广东省普通高校重点科研平台和项目(2022ZDZX1058);2023年广东省普通高校特色创新类项目(2023KTSCX264);2022年广东省科技创新战略专项资金项目(pdjh2022b0857)。

摘　　要：如何更高精度识别道路裂痕是无人驾驶采取紧急避让决策的重要依据,也是待需解决的关键问题。针对该问题,提出一种基于半监督模型的道路裂痕检测算法,在道路图像无标签、缺少标签的情况下,模型能有效地学习道路裂痕的特征。最终的实验结果表明,半监督模型算法同其他模型对比,表现出较佳的结果,达到了实际应用的要求。The construction and maintenance of urban pavements are an integral part of road construction work.Identifying road cracks with higher accuracy is the key to ensuring road quality,and how to solve the problem of extracting features and identifying cracks is also a fundamental problem to be solved.In this paper,a semi-supervised model-based road crack detection algorithm is proposed,which can effectively learn the features of road cracks in the case of unlabelled or unlabelled road images.The final experimental results show that the semi-supervised model algorithm shows better results than other models and meets the requirements of practical applications.

关 键 词：半监督 裂痕检测 计算机视觉 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] U463.6[自动化与计算机技术—计算机科学与技术]