The invention discloses an automatic driving deviation processing method based on a ReliefF-DBN. The method comprises the following steps: step 1, acquiring a road surface driving image in real time; step 2, acquiring simulated noise environment data; step 3, extracting a road lane line image; step 4, training a Relief F-DBN lane line tracking model; step 5, calibrating vehicle speed and angle parameters; and step 6, starting a vehicle abnormity automatic processing system. According to the invention, the severe environment during driving is simulated, salt and pepper noise is superposed on the basis of the road image, the automatic driving tracking function of the patrol vehicle is realized, the stability, reliability and robustness of the vehicle tracking are enhanced, the edge line of the lane line is detected by using the YOLOv3 algorithm, the lane line can be accurately detected. Important lane line data is provided for the tracking model, meanwhile, the ReliefF-DBN model is used for training the lane line tracking model, the input dimension of the lane line data is reduced, and the running speed and accuracy of the tracking model are improved.
一种基于ReliefF‑DBN的自动驾驶偏离处理方法,该方法包括以下步骤:步骤1,实时采集路面行驶图像;步骤2,获取模拟噪声环境数据;步骤3,提取路面车道线图像;步骤4,训练ReliefF‑DBN车道线循迹模型;步骤5,校准车辆速度和角度参数;步骤6,开启车辆异常自动处理系统。本发明模拟了行车时的恶劣环境,在路面图像的基础上叠加了椒盐噪声,实现巡车辆的自动驾驶循迹功能,增强了车辆循迹的稳定性、可靠性和鲁棒性,并利用YOLOv3算法检测车道线的边缘线,可以精确的检测车道线,为循迹模型提供重要的车道线数据,同时使用ReliefF‑DBN模型训练车道线循迹模型,减少了车道线数据的输入维度,提高了循迹模型的运行速度和精确度。
Automatic driving deviation processing method based on ReliefF-DBN
一种基于ReliefF-DBN的自动驾驶偏离处理方法
2021-05-07
Patent
Electronic Resource
Chinese
IPC: | G06K Erkennen von Daten , RECOGNITION OF DATA / B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
Deviation correction control method and system for automatic driving
European Patent Office | 2025
|ReliefF-Based Feature Ranking and Feature Selection for Monitoring Induction Motors
BASE | 2018
|LANE DEVIATION DRIVING ALERT SYSTEM AND LANE DEVIATION DRIVING ALERT METHOD
European Patent Office | 2020
|Research of Red Tide Algae Images Feature Selection Method Based on ReliefF and SBS
British Library Conference Proceedings | 2014
|A course deviation detecting method for the automatic driving of an automobile
Tema Archive | 1976
|