Using mixed objects in the training of object-based image classifications

Image classification for thematic mapping is a very common application in remote sensing, which is sometimes realized through object-based image analysis. In these analyses, it is common for some of the objects to be mixed in their class composition and thus violate the commonly made assumption of o...

Full description

Bibliographic Details
Main Authors: Costa, Hugo, Foody, Giles M., Boyd, Doreen S.
Format: Article
Published: Elsevier 2017
Subjects:
Online Access:https://eprints.nottingham.ac.uk/39755/
_version_ 1848795905975648256
author Costa, Hugo
Foody, Giles M.
Boyd, Doreen S.
author_facet Costa, Hugo
Foody, Giles M.
Boyd, Doreen S.
author_sort Costa, Hugo
building Nottingham Research Data Repository
collection Online Access
description Image classification for thematic mapping is a very common application in remote sensing, which is sometimes realized through object-based image analysis. In these analyses, it is common for some of the objects to be mixed in their class composition and thus violate the commonly made assumption of object purity that is implicit in a conventional object-based image analysis. Mixed objects can be a problem throughout a classification analysis, but are particularly challenging in the training stage as they can result in degraded training statistics and act to reduce mapping accuracy. In this paper the potential of using mixed objects in training object-based image classifications is evaluated. Remotely sensed data were submitted to a series of segmentation analyses from which a range of under- to over-segmented outputs were intentionally produced. Training objects were then selected from the segmentation outputs, resulting in training data sets that varied in terms of size (i.e. number of objects) and proportion of mixed objects. These training data sets were then used with an artificial neural network and a generalized linear model, which can accommodate objects of mixed composition, to produce a series of land cover maps. The use of training statistics estimated based on both pure and mixed objects often increased classification accuracy by around 25% when compared with accuracies obtained from the use of only pure objects in training. So rather than the mixed objects being a problem, they can be an asset in classification and facilitate land cover mapping from remote sensing. It is, therefore, desirable to recognize the nature of the objects and possibly accommodate mixed objects directly in training. The results obtained here may also have implications for the common practice of seeking an optimal segmentation output, and also act to challenge the widespread view that object-based classification is superior to pixel-based classification.
first_indexed 2025-11-14T19:39:31Z
format Article
id nottingham-39755
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:39:31Z
publishDate 2017
publisher Elsevier
recordtype eprints
repository_type Digital Repository
spelling nottingham-397552020-05-04T18:32:17Z https://eprints.nottingham.ac.uk/39755/ Using mixed objects in the training of object-based image classifications Costa, Hugo Foody, Giles M. Boyd, Doreen S. Image classification for thematic mapping is a very common application in remote sensing, which is sometimes realized through object-based image analysis. In these analyses, it is common for some of the objects to be mixed in their class composition and thus violate the commonly made assumption of object purity that is implicit in a conventional object-based image analysis. Mixed objects can be a problem throughout a classification analysis, but are particularly challenging in the training stage as they can result in degraded training statistics and act to reduce mapping accuracy. In this paper the potential of using mixed objects in training object-based image classifications is evaluated. Remotely sensed data were submitted to a series of segmentation analyses from which a range of under- to over-segmented outputs were intentionally produced. Training objects were then selected from the segmentation outputs, resulting in training data sets that varied in terms of size (i.e. number of objects) and proportion of mixed objects. These training data sets were then used with an artificial neural network and a generalized linear model, which can accommodate objects of mixed composition, to produce a series of land cover maps. The use of training statistics estimated based on both pure and mixed objects often increased classification accuracy by around 25% when compared with accuracies obtained from the use of only pure objects in training. So rather than the mixed objects being a problem, they can be an asset in classification and facilitate land cover mapping from remote sensing. It is, therefore, desirable to recognize the nature of the objects and possibly accommodate mixed objects directly in training. The results obtained here may also have implications for the common practice of seeking an optimal segmentation output, and also act to challenge the widespread view that object-based classification is superior to pixel-based classification. Elsevier 2017-03-01 Article PeerReviewed Costa, Hugo, Foody, Giles M. and Boyd, Doreen S. (2017) Using mixed objects in the training of object-based image classifications. Remote Sensing of Environment, 190 . pp. 188-197. ISSN 1879-0704 OBIA; Mixed pixels; Under-segmentation; Over-segmentation; Scale parameter http://www.sciencedirect.com/science/article/pii/S0034425716304965 doi:10.1016/j.rse.2016.12.017 doi:10.1016/j.rse.2016.12.017
spellingShingle OBIA; Mixed pixels; Under-segmentation; Over-segmentation; Scale parameter
Costa, Hugo
Foody, Giles M.
Boyd, Doreen S.
Using mixed objects in the training of object-based image classifications
title Using mixed objects in the training of object-based image classifications
title_full Using mixed objects in the training of object-based image classifications
title_fullStr Using mixed objects in the training of object-based image classifications
title_full_unstemmed Using mixed objects in the training of object-based image classifications
title_short Using mixed objects in the training of object-based image classifications
title_sort using mixed objects in the training of object-based image classifications
topic OBIA; Mixed pixels; Under-segmentation; Over-segmentation; Scale parameter
url https://eprints.nottingham.ac.uk/39755/
https://eprints.nottingham.ac.uk/39755/
https://eprints.nottingham.ac.uk/39755/