The invention relates to a road speed calculation method based on deep learning anomaly correction, and mainly solves two kinds of calculation bottleneck problems of low-speed, static and abnormal data caused by satellite positioning data drift and low-speed non-characterization real road conditions such as boarding and alighting. The method comprises the following steps of: 1, acquiring vehicle data of a corresponding road, judging the state of a vehicle according to the acquired data, preprocessing the acquired data, and removing the corresponding data; secondly, after corresponding data are removed, the vehicle running speed of the corresponding road is calculated, and initial speed data of the corresponding road are obtained; 3, correcting the initial speed data of the corresponding road; and 4, comprehensive speed calculation: calculating the comprehensive speed based on the weighted bagging thought and obtaining the comprehensive speed of the corresponding road section.
本发明涉及一种基于深度学习异常矫正的道路速度计算方法,本发明重点解决由于卫星定位数据漂移导致低速、静止、异常的数据出现,以及如上下客等非表征真实路况低速的这两类计算瓶颈问题。第一步骤:获取相应道路的车辆数据,根据获取数据判断车辆的状态,对获取的数据进行预处理,剔除相应的数据;第二步骤:剔除相应的数据后,计算相应道路的车辆行驶速度,获取相应道路的初始速度数据;第三步骤:对相应道路的初始速度数据进行修正;第四步骤:综合速度计算,基于加权bagging思想的综合速度计算并获取相应路段的综合速速。
Road speed calculation method based on deep learning anomaly correction
一种基于深度学习异常矫正的道路速度计算方法
2024-05-28
Patent
Electronic Resource
Chinese
Road Traffic Anomaly Detection based on Deep Learning Technology
TIBKAT | 2021
|Road network speed prediction and anomaly identification method based on damping Holt model
European Patent Office | 2022
|Urban main road speed prediction method based on big data and deep learning
European Patent Office | 2021
|Road grading speed limiting method and device based on deep learning and storage medium
European Patent Office | 2025
|