Guan Yunlan 1) ;Liu Shaotang 2) ;Zhou Shijian 1,3) ; Zhang Liting 1) ; and Lu Tieding 1)
1)School of Surveying and Mapping Engineering, East China Institute of Technology, Fuzhou 3440002)Department of Civil Engineering, Henan Institute of Engineering, Zhengzhou 4511913)Jiangxi Academy of Science, Nanchang 330029
Abstract:]In traditional plane fitting methods for point clouds, people don’t consider errors in data and in coefficients matrix simultaneously, which will result in incorrectness of plane parameters. In order to overcome this shortcoming, a new method for fitting local plane to point clouds was proposed. The method is based on total least squares. In consideration of the errors in all observation data, we tried to delete outliers from point clouds, and thus obtained a robust solution to plane fitting parameter. Analytical experiments based on simulated data and real data were conducted, and comparisons between the method and traditional methods such as least square method and eigenvalue method were also implemented. The results show that the method has the capability to overcome bad influence from outliers, and to increase the reliability of parameters estimation.