Side-scan sonar seabed sand wave detection method based on multi-scale convolution and pooling strategy

Abstract

The present invention provides a side-scan sonar seabed sand wave detection method based on the multi-scale convolution and pooling strategy, seabed sand wave identification is carried out based on deep learning, waveform matching analysis is combined, and the side-scan sonar seabed sand wave detection method is applied to seabed sand wave detection. Deep semantic information of a seabed sand wave image is extracted based on an accumulative learning bilateral branch network, and a pyramid convolution module and strip pooling are introduced according to characteristics of seabed sand waves to carry out feature adaptive optimization and global information enhancement, so that the problems of unbalanced number distribution of seabed sand wave samples, diverse morphological characteristics of sand waves in a complex seabed environment and serious interference of ocean reverberation noise are solved, the seabed sand waves are positioned by taking echo intensity as a research object through waveform matching, and the morphological characteristics such as wavelength, wave height and asymmetric index of the seabed sand waves and the migration trend of the seabed sand waves are analyzed. Therefore, a foundation is laid for further investigation and analysis of seabed sand waves.

Description

technical field [0001] The invention specifically relates to a side-scan sonar seabed sand wave detection method based on a multi-scale convolution and pooling strategy, and belongs to the technical fields of marine intelligent information processing and underwater target detection and recognition. Background technique [0002] Marine exploration is mainly used to detect various geophysical field characteristics, geological structures and mineral resources on the seabed, and to detect the distribution characteristics of seabed placer and sand waves. Among them, seabed sand wave is a landform that is generally distributed in various sea areas formed under the action of various environmental factors, such as sediment particle size, tidal characteristics and related residual currents. Seabed sand wave identification is an important way to investigate seabed sediments, and it is an important research content in the fields of marine sediment bathymetry, marine geomorphology, mari...

AI Technical Summary

Problems solved by technology

[0007] Finally, whether it is a straight-line sand wave or a crescent-shaped sand wave, its shape has the commonality of long strips, and the ripples present a discrete distribution with intervals. The regular square pooling kernel limits the context flexibility of sand wave feature capture and does not Conducive to the effective extraction of sand wave features, the traditional square pooling kernel is likely to introduce unnecessary connections such as noise to interfere with the identification of seabed sand waves

Images