Automatic outlier detection in multibeam bathymetric data using robust LTS estimation

The bathymetric data volumes produced by Multibeam echo-sounders are increasingly large and inevitably contain outliers which must be detected and eliminated. In this paper, an algorithm based on Least Trimmed Squares (LTS) estimator is presented for outlier detection and elimination. LTS estimator is a robust estimator with high Break-down Point (BP). It can efficiently reduce the influences of outliers on surface fitting and obtain a relatively correct seabed trend surface. The difference between the real measured value and the estimated value is then calculated so that the outliers can be detected and removed. The algorithm is tested using both synthetic data and real bathymetric data acquired by multibeam echo-sounders. The results show that the algorithm is robust and able to detect existing clusters and discrete outliers in data sets effectively.