University of Tasmania

File(s) under permanent embargo

Change Detection from High Resolution Satellite Imagery

posted on 2023-05-26, 20:23 authored by Fu, Xiao Ming
This study analyses whether change detection techniques can identify vegetation changes in Macquarie Island, using high resolution Quickbird satellite imagery; it also takes fuzzy aspects of changes into account in order to explore and discover the uncertainty in the change detection outputs. As Macquarie Island has been severely affected by rabbit grazing, this study probes into the vegetation changes on it by different change detection techniques, with a view to help people make proper management strategies to protect Macquarie Island's unique environment. Besides, because of its geographically isolated location, monitoring changes on the Island is most efficiently achieved using satellite imagery. Traditional change detection techniques can distinguish change from no change in a binary way; incorporating fuzziness is a new approach for change detection as it extends the binary results to the fuzzy results by revealing the transitional nature of the natural environment in the most real way, thus a more understandable interpretation of the change results can be available. Therefore, fuzzy change detection techniques have been investigated in this study. Two Quickbrid images were used as the input in this study; one was acquired in March 2005 and another was acquired in March 2007. In this study, six change detection techniques( with three in each group) have been implemented, namely, normalized difference vegetation index (NDVI), change vector analysis (CVA); postclassification change detection, as one group; and fuzzy normalized difference vegetation index (fuzzy NDVI), fuzzy change vector analysis (fuzzy CVA), fuzzy post-classification change detection, as the other. For the fuzzy group, fuzzy NDVI and fuzzy CVA were generated by fuzzy membership functions, and fuzzy postclassification change detection was generated by comparing with fuzzy classification results. The support vector machine (SVM) was used as the classifier. The study reported that fuzzy change detection not only can identify the changes in the Macquarie Island, but can better reveal the gradual process of change in a specific time period. Nevertheless, it does not improve the accuracy, compared with the traditional methods, mainly because that the results of fuzzy change detection are based on that of the binary change detection.


Publication status

  • Unpublished

Rights statement

Copyright 2007 the Author - The University is continuing to endeavour to trace the copyright owner(s) and in the meantime this item has been reproduced here in good faith. We would be pleased to hear from the copyright owner(s). Includes bibliography

Repository Status

  • Restricted

Usage metrics

    Thesis collection


    No categories selected


    Ref. manager