| 008 |
|
130529n^^^^^^^^xx^||||^o^^^^^|||^0^eng^d |
| 245 |
00 |
|a Urban damage assessment from remotely sensed images |h [electronic resource] |y English. |
| 260 |
|
|a [S.l.] : |b ultidisciplinary Center for Earthquake Engineering Research (MCEER), |c 2001. |
| 490 |
|
|a Student Research Accomplishments |n 8 |y English. |
| 506 |
|
|a Refer to main document/publisher for use rights. |
| 510 |
|
|a Rejaie, A., Shinozuka, M. (2001). Urban damage assessment from remotely sensed images. International Institute of Innovative Risk Reduction Research on Civil Infrastructure Systems and Department of Civil Engineering, University of Southern California. |
| 520 |
3 |
|a This document presents a change detection methodology for analyzing co-registered pre- and post-disaster remotely sensed images. The images were subjected to auto-correlation and principal component analysis to ascertain which technique would be preferable. It was determined that key problems implementing automated analyses can be rectified through PCA. The resulting method produces highlighted regions on the image to visually isolate the changed pixels from those for which no change was detected. Two sets of aerial photos were used as test material. The initial set consisted of pre- and post-earthquake images of Kobe, Japan. However, due to variations with respect to altitude, illumination, viewpoint, and camera, they were substituted with 3 sets of aerial images of 3-D models representing multiple structures. Each image set contained variations in either illumination, number of structures, or both. The PCA algorithm was executed with MATLAB software on a UNIX operating system. PCA produced superior results to that of automated correlation, although some false positives were produced. False detections occurred primarily in shadowed regions, but were also seen in very small, isolated patterns. These issues were rectified for the shadowed regions by masking and threshold biasing the images. The authors speculate that errors involving small clusters might be rectified by incorporating elevation data into the analysis. This technique requires that prospective images be co-registered. Since such images might prove difficult to obtain, future research will involve analyses using slightly mis-registered images in an effort to develop techniques that would compensate for the resulting errors. |
| 533 |
|
|a Electronic reproduction. |c Florida International University, |d 2013. |f (dpSobek) |n Mode of access: World Wide Web. |n System requirements: Internet connectivity; Web browser software. |
| 650 |
1 |
|a Remote sensing |x Technologies. |
| 700 |
1 |
|a Rejaie, Ali |g Ph.D. Candidate |u International Institute if Innovative Risk Reduction Research on Civil Infraestructure Systems and Department of Civil Engineering, University of California. |
| 710 |
2 |
|a Disaster Risk Reduction Program, Florida International University (DRR/FIU), |e summary contributor. |
| 856 |
40 |
|u http://dpanther.fiu.edu/dpService/dpPurlService/purl/FI13042531/00001 |y Click here for full text |
| 992 |
04 |
|a http://dpanther.fiu.edu/sobek/content/FI/13/04/25/31/00001/FI13042531_thm.jpg |
