Underwater target detection with hyperspectral imagery for search and rescue operations

Thesis Type: Postgraduate

Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Graduate School of Informatics, Information Systems, Turkey

Approval Date: 2017




Fast and precise target detection is an issue in underwater object detection applications such as search and rescue, mine detection, archeological remains, etc. Hyperspectral imaging could be effective in determining different components in a selected region thanks to its ability to provide detailed spectral information. The detection of underwater objects with their above-water reflectance signature may not be efficient because of the reflectance changes due to water column. The detection should be conducted with the reflectance signature affected by the water column. To simulate object reflectance under a specific water column, water column correction algorithms can be taken into consideration. Once underwater spectral reflectance is obtained, their effect on the detection performance can be observed. In this thesis, the effect of water column to an object reflectance is investigated. The water column correction algorithms are utilized to create underwater object spectrum from above-water object reflectance. The effect of underwater target spectrum to object detection is also investigated. The materials used for the study are related to search and rescue operations to be evaluated according to their capability of detection in such scenarios. To increase reaction speed in search and rescue operations, the algorithms used for water column correction and target detection are applied to each material to evaluate the optimum water column correction – target detection algorithm pair for material types.