Volume 9, Issue 1 (2005)
MJSP 2005, 9(1): 97-110 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Zeaiean P, Alimohamad A, Rabiei H R. Modeling Uncertainty in Change Detection Based on Classification of Remote Sensing Data. MJSP 2005; 9 (1) :97-110
URL: http://hsmsp.modares.ac.ir/article-21-2983-en.html
URL: http://hsmsp.modares.ac.ir/article-21-2983-en.html
1- Assistant Professor, Department of Geography, Tarbiat Moallem University
2- Assistant Professor, Faculty of Mapping, KNT University
3- M.Sc. Graduate, Department of Remote Sensing, Tarbiat Modarres University
2- Assistant Professor, Faculty of Mapping, KNT University
3- M.Sc. Graduate, Department of Remote Sensing, Tarbiat Modarres University
Abstract: (5115 Views)
Land use/cover change map production is one of the basic needs for environmental monitoring and management. Since the change maps are usually used in planning and decision-making, certainty and reliability of these maps can be very important in many applications. Unfortunately in many studies only probability values as obtained from MLC approach have been used for uncertainty estimation.
Here a new approach has been developed which is based on the probability information as well as spatial parameters including distance, neighborhood, extent and the type of change.
In this study, two Landsat TM images of Isfahan urban area provided in 1990 and 1998 have been co-registered using first order polynomial and nearest neighbor resampling approach. The registered images have been then classified to ten different land use/land cover classes using Maximum Likelihood Classification algorithm. Probabilistic measures generated by the MLC have been used for modeling uncertainty. Using different spatial analysis functions for modeling the change of agricultural areas to residential areas, the relevant spatial parameters have been extracted. Based on logistic regression approach, probabilistic parameters and spatial parameters have been integrated to generate a layer, which shows uncertainty of change of agricultural areas to residential areas. The Relative Operating Characteristics (ROC) index has been used for validation of the model and it has been estimated to be 0.9944, which is an indicative of very good model fitting. As a final conclusion, development of this model is suggested for quantitative evaluation of uncertainty in change detection.
Received: 2004/02/8 | Accepted: 2004/09/20 | Published: 2010/04/27
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |