2015 CSCE Annual Conference Regina - Building on our Growth Opportunities Conference
Mr. David Cody Bradley, O'Kane Consultants Inc. Fredericton, NB, Canada
Dr. Jeff Rankin, UNB
Dr. Lloyd Waugh, UNB
Researchers at the University of New Brunswick provide a means to document and display the progress of construction projects through panoramic images. The introduction of change detection within panoramas facilitates progress monitoring by identifying areas within the panoramas which have changed since the last capture date. This paper investigates change detection challenges and solutions to improve current panoramic change detection process. To reach this goal, a pilot study is conducted which identifies and evaluates photography and post-processing change detection variables. Adobe Photoshop is used to perform all post-processing. Areas of change within panoramas are identified using pixel based intensity differencing. The resulting change image is overlaid on the original image, highlighting changed objects and providing insight on construction progress. Change detection begins with two panoramas taken from the same location on different dates with enough time elapsed so that visible changes are evident. Research reveals that while panoramic change detection techniques produce acceptable results, they still require a significant amount of input and judgment from the user. Therefore, expanding the current status of panoramic monitoring systems through pixel based change detection is still impractical, and there is an opportunity for further research in this area.