Browsing by Author "Filppula, Topi"

The production of depth information of navigation charts is, despite significant developments in bathymetric surveying technology in the last decades, still mostly based on traditional manual methods. Even today the depth contours are generally digitized by hand and the chart soundings are individually selected from the dense bathymetric point data. With decreasing human resources these labor intensive methods have created a bottleneck in the Finnish Transport Agency's navigation chart production. The main goal of this study was to examine two different depth model surface smoothing methods in order to enable the automated production of quality depth contours. The methods used to smooth the depth model surfaces were navigationally safe Laplace-interpolation and Rolling Coin, a method developed specifically for this study. The properties and differences of the smoothed depth model surfaces were examined using surface differences (separation surfaces), deviation metrics and different neighborhood based indices describing local variations in the depth model surface. The suitability of the smoothing methods for automated depth contour generation was evaluated by creating automatically generated depth contours from the smoothed depth model surfaces and comparing these contours to the contours on the current navigation chart products. The comparison of the contours focused on evaluation of the number and lengths of the contours. Cartographic evaluation of the contours was not included in this thesis. Both depth model surface smoothing methods proved to be navigationally safe and promising in terms of automation of depth contour generation. The Rolling Coin method proved to be especially interesting and functional even in the generally troublesome narrow and dredged channels. The Finnish Hydrographic Office is going to take the Rolling Coin method for production use.