Image processing for sediment classification
Cristian Constantin Molder, Florin Şerban, Iulian Constantin Vizitiu, Mihai Stanciu · Computational intelligence · 2008
Sediment analysis using data obtained from a profiler helps geologists to interpret and classify different structural layer of the sea and ocean floor. Different sonar can be used (the monobeam, the multibeam, or the side-scan) and several sonar specific image profiles are recorder. The process of interpretation is made exclusively by a human expert analysis of a paper print or a display image, therefore leading to misinterpretations and a long time required to process a great amount of recorder data. There is a immediate need to create an automatic or semiautomatic system in order to speed up the interpretation process, allowing more time to data refinement that to brute interpretation. In this article we propose the use of two image processing techniques which characterize the recorder data using sediment layer contour extraction and inter-layer texture analysis. The results based on real data are compared with the ones obtained from a geologist and confirms the proposed approach.