Techniques for automatic large scale change analysis of temporal multispectral imagery

Show simple item record

dc.contributor.advisor Messinger, David
dc.contributor.author Mercovich, Ryan
dc.date.accessioned 2012-07-18T19:32:14Z
dc.date.available 2012-07-18T19:32:14Z
dc.date.issued 2012-01-20
dc.identifier.uri http://hdl.handle.net/1850/15235
dc.description.abstract Change detection in remotely sensed imagery is a multi-faceted problem with a wide variety of desired solutions. Automatic change detection and analysis to assist in the coverage of large areas at high resolution is a popular area of research in the remote sensing community. Beyond basic change detection, the analysis of change is essential to provide results that positively impact an image analyst's job when examining potentially changed areas. Present change detection algorithms are geared toward low resolution imagery, and require analyst input to provide anything more than a simple pixel level map of the magnitude of change that has occurred. One major problem with this approach is that change occurs in such large volume at small spatial scales that a simple change map is no longer useful. This research strives to create an algorithm based on a set of metrics that performs a large area search for change in high resolution multispectral image sequences and utilizes a variety of methods to identify different types of change. Rather than simply mapping the magnitude of any change in the scene, the goal of this research is to create a useful display of the different types of change in the image. The techniques presented in this dissertation are used to interpret large area images and provide useful information to an analyst about small regions that have undergone specific types of change while retaining image context to make further manual interpretation easier. This analyst cueing to reduce information overload in a large area search environment will have an impact in the areas of disaster recovery, search and rescue situations, and land use surveys among others. By utilizing a feature based approach founded on applying existing statistical methods and new and existing topological methods to high resolution temporal multispectral imagery, a novel change detection methodology is produced that can automatically provide useful information about the change occurring in large area and high resolution image sequences. The change detection and analysis algorithm developed could be adapted to many potential image change scenarios to perform automatic large scale analysis of change. en_US
dc.language.iso en_US en_US
dc.subject Automatic clustering en_US
dc.subject Change analysis en_US
dc.subject Change detection en_US
dc.subject Graph theory en_US
dc.subject Multispectral algorithms en_US
dc.subject Unsupervised classification en_US
dc.subject.lcc TA1637 .M47 2012
dc.subject.lcsh Image analysis en_US
dc.subject.lcsh Remote sensing--Data processing en_US
dc.subject.lcsh Multispectral photography--Data processing en_US
dc.subject.lcsh Image processing--Digital techniques en_US
dc.title Techniques for automatic large scale change analysis of temporal multispectral imagery en_US
dc.type Dissertation en_US
dc.description.college College of Science en_US
dc.description.department Chester F. Carlson Center for Imaging Science en_US

Files in this item

Files Size Format View
RMercovichDissertation1-20-2012.pdf 50.15Mb PDF View/Open

This item appears in the following Collection(s)

Show simple item record

Search RIT DML


Advanced Search

Browse