A measure of average error variance of line features

Eryong Liu, Wenzhong Shi

Research output: Journal article publicationJournal articleAcademic researchpeer-review

1 Citation (Scopus)

Abstract

This article presents a new development in measuring the positional error of line features in Geographic Information Systems (GIS), in the form of a new measure for estimating the average error variance of line features, including line segment, polyline, polygon, and curved lines. This average error measure is represented in the form of a covariance matrix derived by an analytical approach. Corresponding error indicators are derived from this matrix. The error of line features mainly results from two factors: (1) an error propagated from the original component points of line features and (2) a model error of interpolation between these points. In this study, a method of average error estimation has been derived regarding the first type error of line features that are interpolated by either linear or cubic interpolation methods. The main contribution of the research is the provision of an error measure to assess the quality of spatial data in application settings. The proposed error models for estimating average error variance of line features in a GIS are illustrated by both simulated and practical experiments. The results show that the line accuracy from a linear interpolation is better than a line interpolated using a cubic model.
Original languageEnglish
Pages (from-to)321-327
Number of pages7
JournalCartography and Geographic Information Science
Volume43
Issue number4
DOIs
Publication statusPublished - 7 Aug 2016

Keywords

  • accuracy of lines
  • interpolation models along line features
  • line features
  • Positional error

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Geography, Planning and Development
  • Management of Technology and Innovation

Fingerprint

Dive into the research topics of 'A measure of average error variance of line features'. Together they form a unique fingerprint.

Cite this