|
An Investigation into Fuzzy-Statistical Outlier Detection Applied in LASER Scanner Cloud Points
|
R. Arabsheibani   , Y. Kananisadat , A. Abedini , A. Zoghian |
|
|
|
Abstract: (5160 Views) |
| Typically the random errors exist in every observed parameter as an undesirable part of measured observation which following the normal distribution function (NDF) in statistical components analysis as a satisfactory assumption. Generally those random errors straying from NDF characteristics are being considered as blunders or outliers throughout the set of observations which would effect on the sensitivity analysis negatively. This paper presents application and comparison of Baarda method as a conventional statistical methodology and Fuzzy approach in order to determine the outliers in two set of data including three dimension (3D) points of ground Laser Scanner so-called cloud points as well as a set of random simulated 3D cloud points. Although there are several methods to analyze the random errors and outliers but it is still necessary to find more optimized solutions especially for 3D cloud points such as ground Laser Scanner observations and or performing an improvement to the previous ones in terms of accurateness and preciseness. However Baarda method eliminates the outliers as soon as finding them but Fuzzy method deals mostly with the residuals and observed errors in the adjustment computations procedures which depends critically on statistical tests outputs. The results show that the detected outliers using Baarda method sequel to or more than what Fuzzy method whether for the real cloud points of Laser Scanner or the random 3D simulated cloud points discovered. Moreover, Baarda method is acting quite similar to the Fuzzy method’s results in low number of points approximately while the comparison presented that Fuzzy method approached to the acceptable value of 1 through ameliorating adjustment computations in larger set of points quantity step by step. Besides, the future capabilities of Fuzzy approaches to detect the outliers in different levels of observation quantity are discussed |
|
| Keywords: Fuzzy-statistical methods, Baarda methods, cloud points, outlier |
|
|
Full-Text [PDF 276 kb]
(1064 Downloads)
|
|
Type of Study: Research |
Subject:
Geo&Hydro
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
