轮式移动机器人单目视觉的目标测距方法

发布时间：2018-04-04 21:49

本文选题：轮式移动机器人　切入点：摄像机标定　出处：《计算机工程》2017年04期

【摘要】：为减小特征匹配以及摄像机光轴倾斜对单目测距造成的误差,提出一种基于单目视觉的轮式移动机器人目标测距方法,将测距目标从平面物体推广到立体物体,在无须校正的情况下提高测量精度。通过对摄像机进行标定测量其内外参数,建立小孔平面成像模型得到世界坐标系与像素坐标系的对应关系。针对坐标变换矩阵奇异的情况,引入面积这一特性进行求解,推导特定矩阵奇异情况下目标距离与像素面积的关系。实验结果表明,该方法可将综合误差率控制在0.7%以内,能够满足轮式移动机器人单目视觉测距实时性和可靠性的要求。
[Abstract]:In order to reduce the error caused by the feature matching and the tilt of the optical axis of the camera, a method of ranging the target of wheeled mobile robot based on monocular vision is proposed, which extends the ranging target from a plane object to a solid object.Improve measurement accuracy without correction.By calibrating and measuring the internal and external parameters of the camera, a plane imaging model with small holes is established to obtain the corresponding relationship between the world coordinate system and the pixel coordinate system.In order to solve the singularity of coordinate transformation matrix, the characteristic of area is introduced, and the relation between the distance of target and the area of pixel is deduced in the case of singularity of a given matrix.The experimental results show that the proposed method can control the comprehensive error rate within 0.7%, and can meet the requirements of real-time and reliability of the wheeled mobile robot's monocular vision ranging.
【作者单位】：东南大学自动化学院复杂工程系统测量与控制教育部重点实验室;
【基金】：国家自然科学基金(61473082)
【分类号】：TP391.41;TP242

【相似文献】