地面机器人多尺度道路感知方法研究  

Research on Multi-scale Road Perception Method Of Ground Robot

在线阅读下载全文

作  者:李政委 LI Zhengwei(Taishan College of Science and Technology,Tai,an,Shandong 271000,China)

机构地区:[1]泰山科技学院,山东泰安271000

出  处:《移动信息》2024年第8期257-260,共4页MOBILE INFORMATION

摘  要:由于道路状态的差异化,采用统一的尺度标准对其进行感知会产生相对较大的误差。文中以地面机器人为研究对象,提出了一种多尺度道路感知方法。该方法以基于USB串行总线的便携式外置彩色图像采集盒——USB21A为视觉图像采集装置,在对称双目模式下,根据道路环境的实际光线条件及道路规格来调整传感器的拍摄角度和焦距,获取清晰的道路边界图像信息。在道路感知阶段,以对称双目视觉传感器采集到的RGB图像为基础,对其进行以道路边缘为核心的截取处理,并结合以视觉传感器为中心的坐标系中的道路边缘图像中的像素点位置信息,确定道路边缘信息。在测试结果中,地面机器人沿不同路段中心线行驶的偏移量始终稳定在0.1 m以内,最大值仅为0.08 m(弯道路段),其余路段的偏移量基本稳定在0.05 m以内。Due to the differentiation of road states,using a unified scale standard to perceive them will produce relatively large errors.In this paper,a multi-scale road perception method is proposed with a ground robot as the research object.The method uses a portable external color image acquisition box based on USB serial bus-USB21A as the visual image acquisition device.In the symmetrical binocular mode,the shooting angle and focal length of the sensor are adjusted according to the actual light conditions of the road environment and road specifications to obtain clear road boundary image information.In the road perception stage,based on the RGB image collected by the symmetrical binocular vision sensor,it is intercepted with the road edge as the core,and combined with the pixel position information in the road edge image in the coordinate system centered on the visual sensor,the road edge information is determined.In the test results,the offset of the ground robot traveling along the centerline of different road sections is always stable within 0.1 m,the maximum value is only 0.08 m(curved road sections),and the offset of the rest of the road sections is basically stable within 0.05 m..

关 键 词:地面机器人 道路感知 USB21A视觉图像采集装置 对称双目模式 道路边界图像 截取处理 像素点位置 

分 类 号:TP242.6[自动化与计算机技术—检测技术与自动化装置]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象