An advanced multiple outlier detection algorithm for 3D similarity datum transformation. (15th October 2020)

• Record Type: Journal Article
• Authors: Ma, YouQing; Liu, ShaoChuang; Li, QunZhi
• Is Part Of: Measurement. Volume 163(2020)
• Publication Date: 2020-10-15
• Subjects: 3D similarity datum transformation -- Multiple outlier detection -- Least squares -- Thresholding rule -- Least trimmed squares; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2020.107945
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: 3D similarity datum transformation. Multiple outlier detection in the generalized Errors-In-Variables (EIV) model. Reduction of the influence of the multiple outliers. Accurate and reliable transformation parameters. Abstract: Estimated transformation parameters based on the three-dimensional (3D) similarity datum transformation are affected or even severely distorted when the observed coordinates are contaminated by gross errors or outliers. In our paper, the problem of 3D similarity datum transformation is described as a generalized Errors-In-Variables (EIV) model based on a least squares solution. Then, an advanced multiple outlier detection algorithm that uses the L 1 penalty function on the mean-shift model is proposed in the generalized EIV model. A general thresholding rule and a least trimmed squares estimator are invoked in the proposed algorithm. The results of the real and simulated experiments indicate that the proposed algorithm can effectively reduce the influence of multiple outliers and yield reliable transformation parameters compared with the iteratively reweighted total least squares, data snooping and robust outlier removal approaches.