摘要
针对当前利用移动设备检测坑槽不能很好满足数据来源充分性与实时性的问题，提出一种基于群智感知的道路坑槽检测系统。系统利用主动参与者的Android手机中的当前时间、当前网络IP、加速度计和GPS传感器收集路面信息，并上传至中心服务器。中心服务器上对上传的数据采用统计学的方法提取相关特征，采用k-近邻算法与k-means聚类算法对坑槽进行了实时轻量级检测。实验结果表明，系统能够检测坑槽的破损等级，准确性达到82．48％。
Nowadays, using mobile device is not enough to collect sufficient and real time road surface information. A mobile crowd sensing system for pothole detection is proposed in this paper. Active participators use their Android smartphones to collect the road surface trajectory by the smartphone＇s current time, current internet IP, accelerometer and GPS sensors data in this system. All collected data will be transmitted to the central server. Via careful selection of signal features using statistical methods, k-nearest neighbor algorithm and k-means clustering algorithm are used to implement lightweight real time pothole detection. The experimental results show this system can detect the damage level of pothole, and the detection accuracy is up to 82.48% .
出处
《广西大学学报（自然科学版）》
CAS
北大核心
2015年第2期436-443,共8页
Journal of Guangxi University（Natural Science Edition）
基金
国家自然科学基金资助项目（61462007）
关键词
坑槽检测
加速度计
机器学习
聚类算法
pothole detection
accelerometer
machine learning
clustering algorithm
