Bathymetric map production by using satellite imagery: A case study of Kozlu Fishery Harbour

The coastal zones are one of the continually changing regions of the world. The examination and the measurement of these changes are very important for the construction of coastal areas, ports, pipelines and other critical infrastructures. One of the most important parameters in the related studies is the information obtained about the water depth. Bathymetry is defined as the study or technique of determining the topography of the seabed. The bathymetric methods used by human beings since ancient times continue to develop in parallel with technological developments. In the early times, while the bathymetric data were obtained by classical methods, they were later replaced with modern methods such as Echo Sounding and LIDAR (Laser Imaging Detection and Ranging) Bathymetry. Nowadays, in addition to these methods, studies are carried out to obtain and improve the bathymetric data using satellites with various sensors. SDB (Satellite-derived Bathymetry), which has recently introduced a new future in the hydrographic surveying industry, is a technique that obtains bathymetric data from multispectral satellite images. In this thesis, it is aimed to obtain bathymetry data using satellite images and determine the accuracy of these data by interpreting the results produced by LRT (Log Ratio Transformation) method on an area where bathymetry data were previously obtained. For this purpose, Landsat 8 satellite imagery of the Zonguldak Kozlu Fishing Shelter were processed using ENVI Classic remote sensing software by using various band combinations and filtering techniques and relative water depths were calculated. The obtained data were compared with the reference depth data derived from the current bathymetry data of Zonguldak Kozlu Fishing Shelter and error values were determined for each band combination. B1-B6, B2-B6 and B2-B5 band combinations, which give the lowest root mean square (RMS) error values for the whole region, were also evaluated in terms of model errors. It is concluded that the working area of the B2-B5 band combination with ± 1.54m root mean square error was the most appropriate evaluation. Bathymetric map of the study area was produced by using multiple modeling software with the data obtained from this evaluation.