The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data

The accuracy of a map is dependent on the reference dataset used in its construction. Classification analyses used in thematic mapping can, for example, be sensitive to a range of sampling and data quality concerns. With particular focus on the latter, the effects of reference data quality on land c...

Full description

Bibliographic Details
Main Authors: Foody, Giles M., Pal, Mahesh, Rocchini, Duccio, Garzon-Lopez, Carol X., Bastin, Lucy
Format: Article
Published: MDPI 2016
Subjects:
Online Access:https://eprints.nottingham.ac.uk/38443/
_version_ 1848795612768632832
author Foody, Giles M.
Pal, Mahesh
Rocchini, Duccio
Garzon-Lopez, Carol X.
Bastin, Lucy
author_facet Foody, Giles M.
Pal, Mahesh
Rocchini, Duccio
Garzon-Lopez, Carol X.
Bastin, Lucy
author_sort Foody, Giles M.
building Nottingham Research Data Repository
collection Online Access
description The accuracy of a map is dependent on the reference dataset used in its construction. Classification analyses used in thematic mapping can, for example, be sensitive to a range of sampling and data quality concerns. With particular focus on the latter, the effects of reference data quality on land cover classifications from airborne thematic mapper data are explored. Variations in sampling intensity and effort are highlighted in a dataset that is widely used in mapping and modelling studies; these may need accounting for in analyses. The quality of the labelling in the reference dataset was also a key variable influencing mapping accuracy. Accuracy varied with the amount and nature of mislabelled training cases with the nature of the effects varying between classifiers. The largest impacts on accuracy occurred when mislabelling involved confusion between similar classes. Accuracy was also typically negatively related to the magnitude of mislabelled cases and the support vector machine (SVM), which has been claimed to be relatively insensitive to training data error, was the most sensitive of the set of classifiers investigated, with overall classification accuracy declining by 8% (significant at 95% level of confidence) with the use of a training set containing 20% mislabelled cases.
first_indexed 2025-11-14T19:34:52Z
format Article
id nottingham-38443
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:34:52Z
publishDate 2016
publisher MDPI
recordtype eprints
repository_type Digital Repository
spelling nottingham-384432020-05-04T18:14:03Z https://eprints.nottingham.ac.uk/38443/ The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data Foody, Giles M. Pal, Mahesh Rocchini, Duccio Garzon-Lopez, Carol X. Bastin, Lucy The accuracy of a map is dependent on the reference dataset used in its construction. Classification analyses used in thematic mapping can, for example, be sensitive to a range of sampling and data quality concerns. With particular focus on the latter, the effects of reference data quality on land cover classifications from airborne thematic mapper data are explored. Variations in sampling intensity and effort are highlighted in a dataset that is widely used in mapping and modelling studies; these may need accounting for in analyses. The quality of the labelling in the reference dataset was also a key variable influencing mapping accuracy. Accuracy varied with the amount and nature of mislabelled training cases with the nature of the effects varying between classifiers. The largest impacts on accuracy occurred when mislabelling involved confusion between similar classes. Accuracy was also typically negatively related to the magnitude of mislabelled cases and the support vector machine (SVM), which has been claimed to be relatively insensitive to training data error, was the most sensitive of the set of classifiers investigated, with overall classification accuracy declining by 8% (significant at 95% level of confidence) with the use of a training set containing 20% mislabelled cases. MDPI 2016-11-01 Article PeerReviewed Foody, Giles M., Pal, Mahesh, Rocchini, Duccio, Garzon-Lopez, Carol X. and Bastin, Lucy (2016) The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data. ISPRS International Journal of Geo-Information, 5 (11). 199/1-199/20. ISSN 2220-9964 classification; training; error; accuracy; remote sensing; land cover http://www.mdpi.com/2220-9964/5/11/199 doi:10.3390/ijgi5110199 doi:10.3390/ijgi5110199
spellingShingle classification; training; error; accuracy; remote sensing; land cover
Foody, Giles M.
Pal, Mahesh
Rocchini, Duccio
Garzon-Lopez, Carol X.
Bastin, Lucy
The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title_full The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title_fullStr The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title_full_unstemmed The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title_short The sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
title_sort sensitivity of mapping methods to reference data quality: training supervised image classifications with imperfect reference data
topic classification; training; error; accuracy; remote sensing; land cover
url https://eprints.nottingham.ac.uk/38443/
https://eprints.nottingham.ac.uk/38443/
https://eprints.nottingham.ac.uk/38443/