针对标靶球定位缺少有效的确定协因数阵方法的问题,顾及点到平面的距离反映了点与平面的相关性及入射角对点云数据点位精度的影响,将两者推广到三维激光扫描标靶球定位中。以距离、先验入射角确定各点协因数值,并给出观测向量协因数阵及系数矩阵协因数阵,利用稳健加权总体最小二乘方法进行标靶球定位。实例表明,以距离确定协因数阵的稳健加权总体最小二乘方法解算标靶球参数估计比其他方法更精确。
We are concerned with the issue of lacking an effective method to determine the covariance matrix of coefficient matrix and observation vector in spherical target positioning.The distance from point to plane reflects the correlation of point and plane,and the incidence angle makes a difference to point clouds.We expand both of them to spherical target positioning.The covariance matrix of coefficient matrix and observation vector are provided.The robust weighted total least squares method is applied to spherical target positioning,which is based on weighted total least squares and setting certain criteria.Point clouds with outliers are conquered in this method.The experimental results show that robust weighted total least squares,which determined covariance by distance,is better than other methods in spherical target positioning.