26 October 2011 Object-based change detection: dimension of damage in residential areas of Abu Suruj, Sudan
Author Affiliations +
Abstract
Given the importance of Change Detection, especially in the field of crisis management, this paper discusses the advantage of object-based Change Detection. This project and the used methods give an opportunity to coordinate relief actions strategically. The principal objective of this project was to develop an algorithm which allows to detect rapidly damaged and destroyed buildings in the area of Abu Suruj. This Sudanese village is located in West-Darfur and has become the victim of civil war. The software eCognition Developer was used to per-form an object-based Change Detection on two panchromatic Quickbird 2 images from two different time slots. The first image shows the area before, the second image shows the area after the massacres in this region. Seeking a classification for the huts of the Sudanese town Abu Suruj was reached by first segmenting the huts and then classifying them on the basis of geo-metrical and brightness-related values. The huts were classified as "new", "destroyed" and "preserved" with the help of a automated algorithm. Finally the results were presented in the form of a map which displays the different conditions of the huts. The accuracy of the project is validated by an accuracy assessment resulting in an Overall Classification Accuracy of 90.50 percent. These change detection results allow aid organizations to provide quick and efficient help where it is needed the most.
© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Timo Demharter, Timo Demharter, Ulrich Michel, Ulrich Michel, Manfred Ehlers, Manfred Ehlers, Peter Reinartz, Peter Reinartz, } "Object-based change detection: dimension of damage in residential areas of Abu Suruj, Sudan", Proc. SPIE 8181, Earth Resources and Environmental Remote Sensing/GIS Applications II, 81811J (26 October 2011); doi: 10.1117/12.897589; https://doi.org/10.1117/12.897589
PROCEEDINGS
7 PAGES


SHARE
Back to Top