首页|基于共面靶标的线结构光传感器标定新方法

基于共面靶标的线结构光传感器标定新方法

扫码查看
现有的线结构光传感器标定方法中,结构光平面上标定点的计算精度将会直接影响到传感器最终的标定精度.提出了一种基于共面靶标的线结构光传感器标定新方法,该方法无需计算光平面上的标定点,也无需反复计算摄像机的外参数.多次移动共面靶标计算不同方向激光条纹直线的消隐点,并对其拟合直线得到光平面的消隐线,完成光平面法向的标定.根据交比不变原理计算共面靶标上标定点间的距离,并以其为约束来标定剩余参数.考虑到误差传递的影响,定义了优化目标函数,以已求得参数为初值进行非线性优化.对比实验表明了该方法具有较高的标定精度,测量误差均方根为0.0306mm,且标定过程简单,计算复杂度低,适合现场标定.
New Calibration Method for Line Structured Light Sensor Based on Planar Target
Among the existing calibration methods for line structured light sensor,the accuracy of calibration points in laser plane has direct influence on the calibration result of line structured light sensor.A new calibration method for line structured light sensor based on planar target is proposed.The proposed approach does not need to calculate the calibration points in laser plane and the extrinsic camera parameters repeatedly.Planar target is moved to capture laser stripe with different directions and compute the vanish points,and then the normal vector of laser plane is calibrated by the way of fitting straight using the vanish lines.After that,distances between each calibration points are computed based on the principle of cross ratio invariance,and the remaining parameters are calculated according to the distances.Considering the influence of error propagation,an optimized objective function is defined with the obtained parameters as initial value.The results of comparative experiment show that the proposed method has higher calibration accuracy,and the root mean square measurement error is 0.0306 mm.Besides that,the calibration process is simple with low computational complexity,and suitable for calibration on spot.

measurementline structured lightplanar targetcalibrationvanish pointvanish linecross ratio invariance

陈天飞、赵吉宾、吴翔

展开 >

河南工业大学电气工程学院,河南郑州450001

中国科学院沈阳自动化研究所,辽宁沈阳110016

测量 线结构光 共面靶标 标定 消隐点 消隐线 交比不变

国家自然科学基金辽宁省科技创新重大专项河南省教育厅科学技术研究重点项目河南工业大学高层次人才基金

5137547620130100214B4130012014BS008

2015

光学学报
中国光学学会 中国科学院上海光学精密机械研究所

光学学报

CSTPCDCSCD北大核心EI
影响因子:1.931
ISSN:0253-2239
年,卷(期):2015.35(1)
  • 25
  • 11