APPLICATION OF DEEP LEARNING METHOD TO THE NOISE REMOVAL FOR MULTIBEAM ECHO SOUNDING DURING DREDGING WORKS

Abstract

 The Ministry of Land, Infrastructure, Transport and Tourism (MLIT) promotes “i-Construction” which drastically improves productivity in all construction processes. As a part of “i-Construction”, multibeam echo sounding is used to measure bathymetry for dredging works. Although multibeam echo sounding can measure many point data at one time, it takes considerable time and labor to remove the noise from multibeam echo sounding data.

 This research applies the deep learning method to the noise removal for multibeam echo sounding. The training data uses the multibeam echo sounding data observed at the actual dredging works. The accuracy and work efficiency of the developed denoising model using the deep learning method are examined by other data collected at the actual dredging works. The results show that the proposed model can remove the noise with a certain degree of accuracy and that the working time can be reduced.