A Method Based on FFT and Invariable Moment Features for UTAN

Traditionally, as the echo sounder is deployed in underwater terrain aided navigation system (UTAN), navigation and positioning accuracy is very low, even a large number of mismatches appear.In this paper, the multi-beam sounding has been introduced into UTAN, the input gray-scale image is transformed based on FFT from spatial domain to frequency domain. The Fourier phase correlation technique is used to estimate the candidate translations in order to decrease searching space. Hu invariable moments are then applied to obtain the best matching from candidate translations obtained from phase correlation. Based on different experiments to obtain the measured seabed topography, navigation data and multi-beam sounding data measured, the experimental result show the proposed method can acquire higher accuracy of the navigation and positioning for the underwater vehicle, the INS error is reduced to about 10% of the original error, especially in the case of large original errors (less than 3km) for INS the accuracy is improved obviously.