Unmanned vehicle visual processing method and system based on road shoulder characteristics and unmanned vehicle equipment
Unmanned vehicle visual processing method and system based on road shoulder characteristics and unmanned vehicle equipment
- Neue Suche nach: WANG XIAOYU
- Neue Suche nach: WANG CHENGXIONG
- Neue Suche nach: WANG XINYANG
- Neue Suche nach: PENG LIANQING
- Neue Suche nach: HE SIYI
- Neue Suche nach: WANG XIAOYU
- Neue Suche nach: WANG CHENGXIONG
- Neue Suche nach: WANG XINYANG
- Neue Suche nach: PENG LIANQING
- Neue Suche nach: HE SIYI
The invention discloses an unmanned vehicle visual processing method and system based on road shoulder characteristics and unmanned vehicle equipment. The method comprises the steps that S1, a depth image of a road shoulder is collected during the driving period of an unmanned vehicle; s2, analyzing the depth image to obtain visual identification data, wherein the visual identification data comprises a deflection angle of the unmanned vehicle and a distance between the unmanned vehicle and a road shoulder; s3, judging whether errors exist in the obtained current visual identification data or not, if no errors exist, storing the current visual identification data as current processing data, and if errors exist, abandoning the current visual identification data, and calling the previous stored visual identification data as the current processing data; and S4, calculating an adjustment instruction for adjusting the drift angle of the unmanned vehicle according to the current processing data so as to enable the distance between the unmanned vehicle and the road shoulder to approach the target distance. By introducing an error discrimination mechanism, the situation that the whole control process is unstable due to data errors of a certain frame or several frames can be avoided, so that the reliability and the stability of control are improved.
本发明公开了一种基于路肩特征的无人车视觉处理方法及其系统以及无人车设备，其方法包括：步骤S1：在无人车行驶期间采集路肩的深度图像；步骤S2：分析深度图像得到视觉识别数据，视觉识别数据包括无人车的偏角以及与路肩的间距；步骤S3：判断获取当前视觉识别数据是否存在错误，若无错误，则作为当前处理数据进行存储，若错误，则舍弃当前视觉识别数据，调取前一次所存储的视觉识别数据作为当前处理数据；步骤S4：根据当前处理数据计算调节无人车偏角的调节指令以使无人车与路肩的间距趋近目标间距。通过引入错误判别机制，可以避免因为某一帧或某几帧数据出错而导致整个控制过程失稳的情况，从而提高控制的可靠性和稳定性。
Dokumentinformationen
Unmanned vehicle visual processing method and system based on road shoulder characteristics and unmanned vehicle equipment
基于路肩特征的无人车视觉处理方法、系统及无人车设备
2022-07-08
Patent
Elektronische Ressource
Chinesisch
Ähnliche Titel
Unmanned deicing vehicle road surface clearing method and unmanned deicing vehicle
| Europäisches Patentamt | 2020
Unmanned vehicle control system, unmanned vehicle, and unmanned vehicle control method
| Europäisches Patentamt | 2024
UNMANNED VEHICLE CONTROL SYSTEM, UNMANNED VEHICLE, AND UNMANNED VEHICLE CONTROL METHOD
| Europäisches Patentamt | 2023
Unmanned vehicle control system, unmanned vehicle, and unmanned vehicle control method
| Europäisches Patentamt | 2022