Proceedings Article | 25 January 2018
KEYWORDS: RGB color model, Coastal modeling, Photography, Polishing, Digital photography, Image processing, Oceanography, Atmospheric propagation, Algorithm development, 3D modeling
In the shallow coastal zone (up to 4 m depth), the quality and accuracy of bathymetry data is insufficient because of the spatial variability of the seabed. For the Baltic polish coast bathymetry data are usually created in profiles based on echo sounding measurements. There are some trial applications of green laser – LIDAR known. It can be a method of shallow areas studies quality improvement. However, this method is still expensive and that is why we have decided to use the RGB digital aerial photographs to create a model for mapping the seabed of the shallow coastal zone.
So far, in the 60's, researchers in the USA (Musgrove, 1969) and Russia (Zdanowicz, 1963) developed the first method of bathymetry determining from aerial panchromatic (black-white) photographs. This method was adapted for the polish conditions by Furmanczyk in 1975 and in 2014 we returned to his concept using more advanced techniques of recording and image processing.
In our study, we propose to determine the bathymetry in shallow coastal zone of the Baltic Sea by using the digital color vertical aerial photographs (both single and multi-channel spectral). These photos are the high-resolution matrix (10 cm per pixel) containing values of the pixels in the individual spectral bands (RGB). This gives great possibilities to determine the bathymetry in order to analyze the changes in the marine coastal zone. Comparing the digital bathymetry maps - obtained by proposed method - in the following periods, you can develop differential maps, which reflect the movements of sea-bottom sediments. This can be used to indicate the most dynamic parts in the coastal zone.
The model is based on the pixel values and relative depths, measured in-situ in geo-located points (in selected checkpoints for three areas: beach, shoal and depth parts). As a result, the relation of the pixel brightness and sea depth (the algorithm) was defined in two ways: 1) from relation of two or three bands and 2) from ratio of two channels. Using the algorithm, depth calculations for the whole scene were done and high resolution bathymetric map created to obtain three-dimensional bathymetry visualization. However, the algorithm requires numbers of adjustments resulting from, e.g., the phenomenon of vignetting, light propagation in the atmosphere or light reflection from the sea surface. We have developed the algorithm with correction formulas and created a final model in MATLAB. This model enables to determine the bathymetry of the most dynamic areas in the marine coastal zone up to 3-4 meters depth with a relatively good accuracy. In addition, the possibility to take pictures from the drone instead of a plane significantly reduces the cost of the process. We have also tried to adapt our methodology to satellite data processing.
We will present the model and its results for the area of the Polish western Baltic coast.
1. Musgrove R,G., 1969. Photometry for interpretation. Photogrametric Engineering No. 10.
2. Furmańczyk K., 1975. Możliwości praktycznego zastosowania metody fotogrametrycznej do określania głębokości w strefie brzegowej morza. Gdańsk.
3. Zdanowicz W.G., 1963. Primienienije aerometodow dlia issledowanija moria. Leningrad.
