The invention provides a method and system for automatically correcting and supplementing collected data, and the method comprises the steps: carrying out the extraction of traffic flow and road information of an automobile data recorder video in a real scene through deep learning target detection, multi-target tracking, lane line detection, image classification and other algorithms, and then automatically comparing with collected csv data; error information in the data is corrected, acquisition information which does not exist in the data is supplemented, so that errors of acquisition equipment such as a radar are made up, the accuracy of the acquired data is integrally improved, and finally the accuracy of a constructed scene library or a simulation scene is improved.
本发明提供一种采集数据自动修正和补充方法及系统,该方法采用深度学习目标检测、多目标追踪、车道线检测、图像分类等算法对真实场景行车记录仪视频进行交通流及道路信息的提取,再自动跟采集的csv数据进行对比,纠正数据当中的错误信息,补充数据当中没有的采集信息,从而弥补雷达等采集设备的误差,整体上提升采集数据精度,最终实现构建的场景库或仿真场景精度的提升。
Method and system for automatically correcting and supplementing collected data
一种采集数据自动修正和补充方法及系统
2023-05-09
Patent
Electronic Resource
Chinese
Pneumatic tire capable of automatically supplementing air
European Patent Office | 2025
|Crane pulley device capable of automatically supplementing oil
European Patent Office | 2021
|Beidou positioning-based farming feeding system for automatically supplementing feed
European Patent Office | 2020
|Ship anchor weighing device capable of automatically supplementing sand
European Patent Office | 2020
|System for automatically supplementing image of blind area of rearview mirror
European Patent Office | 2022
|