Fusion of hyperspectral and depth data using morphological image processing for pixel-based classification
| Author(s) | Goker, Abdullah | |
| Date Accessioned | 2019-01-16T14:31:28Z | |
| Date Available | 2019-01-16T14:31:28Z | |
| Publication Date | 2018 | |
| SWORD Update | 2018-10-18T16:02:15Z | |
| Abstract | The availability of multi-sensor data from the same field of view has increased drastically with recent developments in sensor technologies. There are many image processing algorithms to extract different features of objects from sensors, but no single-sensor technology is sufficient to provide dependable classification. Extracting features from multiple sources with morphological operations gives rise to problems like the curse of dimensionality, which degrades the performance of the classifier and considerably increases the computational time. In order to overcome this problem, in this project the features are fused in a lower dimensional space, while as much information as possible about the features of the pixels is preserved. In this way, the classification performance of the given system can be enhanced. | en_US |
| Advisor | Arce, Gonzalo R. | |
| Degree | M.S. | |
| Department | University of Delaware, Department of Electrical and Computer Engineering | |
| Unique Identifier | 1082357096 | |
| URL | http://udspace.udel.edu/handle/19716/24043 | |
| Language | en | |
| Publisher | University of Delaware | en_US |
| URI | https://search.proquest.com/docview/2128406865?accountid=10457 | |
| Title | Fusion of hyperspectral and depth data using morphological image processing for pixel-based classification | en_US |
| Type | Thesis | en_US |