An approach to identify multiple outliers based on sequential likelihood ratio tests. Issue 357 (2nd November 2017)

• Record Type: Journal Article
• Authors: Klein, I.; Matsuoka, M. T.; Guzatto, M. P.; Nievinski, F. G.
• Is Part Of: Survey review. Volume 49:Issue 357(2017)
• Page Start: 449
• Page End: 457
• Publication Date: 2017-11-02
• Subjects: Multiple outliers -- Sequential testing procedure -- Maximum likelihood ratio -- Geodetic networks; Surveying -- Periodicals; Great Britain -- Surveys -- Periodicals; 526.9
• DOI: 10.1080/00396265.2016.1212970
• Languages: English
• ISSNs: 0039-6265

Abstract

Abstract : One of the main challenges in the quality control of geodetic measurements is the reliable identification of multiple outliers. Within this context, the goal of this paper is to present a procedure designated here as Sequential Likelihood Ratio Tests for Multiple Outliers (SLRTMO). To verify its performance, a levelling network was simulated involving one, two and three (simultaneous) outliers. Also a GNSS network involving one and two (simultaneous) outliers was analysed. Results showed that SLRTMO is efficient for single and multiple outliers, simulated with magnitude greater than five standard deviations, with a mean success rate of 79.6% for these cases . Furthermore, the maximum number of outliers to be tested has to be defined according to the redundancy of the network so as to ensure the performance of SLRTMO.