Unsupervised Quantification of Under-and Over-Segmentation for Object-Based Remote Sensing Image Analysis

Andres Troya-Galvis, Pierre Gancarski, Nicolas Passat, Laure Berti-Equille

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Object-based image analysis (OBIA) has been widely adopted as a common paradigm to deal with very high-resolution remote sensing images. Nevertheless, OBIA methods strongly depend on the results of image segmentation. Many segmentation quality metrics have been proposed. Supervised metrics give accurate quality estimation but require a ground-truth segmentation as reference. Unsupervised metrics only make use of intrinsic image and segment properties; yet most of them strongly depend on the application and do not deal well with the variability of objects in remote sensing images. Furthermore, the few metrics developed in a remote sensing context mainly focus on global evaluation. In this paper, we propose a novel unsupervised metric, which evaluates local quality (per segment) by analyzing segment neighborhood, thus quantifying under-and over-segmentation given a certain homogeneity criterion. Additionally, we propose two variants of this metric, for estimating global quality of remote sensing image segmentation by the aggregation of local quality scores. Finally, we analyze the behavior of the proposed metrics and validate their applicability for finding segmentation results having good tradeoff between both kinds of errors.

Original languageEnglish
Article number7112093
Pages (from-to)1936-1945
Number of pages10
JournalIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Volume8
Issue number5
DOIs
Publication statusPublished - 1 May 2015

Fingerprint

image analysis
segmentation
Image analysis
Remote sensing
remote sensing
Image segmentation
Agglomeration
homogeneity

Keywords

  • image region analysis
  • Image segmentation
  • object oriented methods
  • quality control

ASJC Scopus subject areas

  • Computers in Earth Sciences
  • Atmospheric Science

Cite this

Unsupervised Quantification of Under-and Over-Segmentation for Object-Based Remote Sensing Image Analysis. / Troya-Galvis, Andres; Gancarski, Pierre; Passat, Nicolas; Berti-Equille, Laure.

In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 8, No. 5, 7112093, 01.05.2015, p. 1936-1945.

Research output: Contribution to journalArticle

@article{8b97415825744ca28168f45da44bea12,
title = "Unsupervised Quantification of Under-and Over-Segmentation for Object-Based Remote Sensing Image Analysis",
abstract = "Object-based image analysis (OBIA) has been widely adopted as a common paradigm to deal with very high-resolution remote sensing images. Nevertheless, OBIA methods strongly depend on the results of image segmentation. Many segmentation quality metrics have been proposed. Supervised metrics give accurate quality estimation but require a ground-truth segmentation as reference. Unsupervised metrics only make use of intrinsic image and segment properties; yet most of them strongly depend on the application and do not deal well with the variability of objects in remote sensing images. Furthermore, the few metrics developed in a remote sensing context mainly focus on global evaluation. In this paper, we propose a novel unsupervised metric, which evaluates local quality (per segment) by analyzing segment neighborhood, thus quantifying under-and over-segmentation given a certain homogeneity criterion. Additionally, we propose two variants of this metric, for estimating global quality of remote sensing image segmentation by the aggregation of local quality scores. Finally, we analyze the behavior of the proposed metrics and validate their applicability for finding segmentation results having good tradeoff between both kinds of errors.",
keywords = "image region analysis, Image segmentation, object oriented methods, quality control",
author = "Andres Troya-Galvis and Pierre Gancarski and Nicolas Passat and Laure Berti-Equille",
year = "2015",
month = "5",
day = "1",
doi = "10.1109/JSTARS.2015.2424457",
language = "English",
volume = "8",
pages = "1936--1945",
journal = "IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing",
issn = "1939-1404",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Unsupervised Quantification of Under-and Over-Segmentation for Object-Based Remote Sensing Image Analysis

AU - Troya-Galvis, Andres

AU - Gancarski, Pierre

AU - Passat, Nicolas

AU - Berti-Equille, Laure

PY - 2015/5/1

Y1 - 2015/5/1

N2 - Object-based image analysis (OBIA) has been widely adopted as a common paradigm to deal with very high-resolution remote sensing images. Nevertheless, OBIA methods strongly depend on the results of image segmentation. Many segmentation quality metrics have been proposed. Supervised metrics give accurate quality estimation but require a ground-truth segmentation as reference. Unsupervised metrics only make use of intrinsic image and segment properties; yet most of them strongly depend on the application and do not deal well with the variability of objects in remote sensing images. Furthermore, the few metrics developed in a remote sensing context mainly focus on global evaluation. In this paper, we propose a novel unsupervised metric, which evaluates local quality (per segment) by analyzing segment neighborhood, thus quantifying under-and over-segmentation given a certain homogeneity criterion. Additionally, we propose two variants of this metric, for estimating global quality of remote sensing image segmentation by the aggregation of local quality scores. Finally, we analyze the behavior of the proposed metrics and validate their applicability for finding segmentation results having good tradeoff between both kinds of errors.

AB - Object-based image analysis (OBIA) has been widely adopted as a common paradigm to deal with very high-resolution remote sensing images. Nevertheless, OBIA methods strongly depend on the results of image segmentation. Many segmentation quality metrics have been proposed. Supervised metrics give accurate quality estimation but require a ground-truth segmentation as reference. Unsupervised metrics only make use of intrinsic image and segment properties; yet most of them strongly depend on the application and do not deal well with the variability of objects in remote sensing images. Furthermore, the few metrics developed in a remote sensing context mainly focus on global evaluation. In this paper, we propose a novel unsupervised metric, which evaluates local quality (per segment) by analyzing segment neighborhood, thus quantifying under-and over-segmentation given a certain homogeneity criterion. Additionally, we propose two variants of this metric, for estimating global quality of remote sensing image segmentation by the aggregation of local quality scores. Finally, we analyze the behavior of the proposed metrics and validate their applicability for finding segmentation results having good tradeoff between both kinds of errors.

KW - image region analysis

KW - Image segmentation

KW - object oriented methods

KW - quality control

UR - http://www.scopus.com/inward/record.url?scp=85027927135&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027927135&partnerID=8YFLogxK

U2 - 10.1109/JSTARS.2015.2424457

DO - 10.1109/JSTARS.2015.2424457

M3 - Article

VL - 8

SP - 1936

EP - 1945

JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

SN - 1939-1404

IS - 5

M1 - 7112093

ER -