Detecting Outliers and Filtering Noises in Multi-Beam Data

This paper combines median filter with local variance estimation and wavelet analysis to remove outliers and noises.Because median filter can distort details in datasets, its result is not ultimate. However, the local variance is calculated through it.Then removing outliers relies on limit errors calculated by the local variance.The algorithm is verified by the real data and simulated data.It can process large quantities of data, rapid speed and robust etc.