基于非线性最小二乘法的二维码视觉定位算法研究

Research on QR code visual positioning algorithm based on nonlinear least squares method

针对AGV视觉定位算法中存在附加约束的问题,将附加约束的相机坐标系相对于全局坐标系位姿求解问题归纳为非线性最小二乘问题,并采用高斯-牛顿法进行求解。论文所研究的AGV定位算法采用二维码作为路标进行视觉定位,算法首先通过识别经过图像预处理后的二维码,得到该二维码所包含的相对于全局坐标系中的位姿信息;其次建立包含附加约束条件的非线性最优化问题模型,接着通过高斯-牛顿法求出满足包括附加约束条件在内的约束优化问题最优解;最后通过实验验证算法能够解决存在附加约束条件的视觉定位问题,验证了算法的可行性。

To address the problem of additional constraints in the AGV visual positioning algorithm,the problem of solving the camera coordinate system with additional constraints relative to the global coordinate system positional solution is reduced to a nonline-ar least squares problem and solved using the Gauss-Newton method.In this paper,the AGV positioning algorithm uses QR codes as a road sign for visual positioning.Firstly,the algorithm identifies the QR code after image preprocessing and obtains the positional information contained in the QR code relative to the global coordinate system.Secondly,it establishes a nonlinear optimization problem model including additional constraints and then solves this problem with additional constraints by the Gauss-Newton method.Finally,it is verified by experiments that the algorithm in this paper can solve the problems with addi-tional constraints and the feasibility of the algorithm is verified.