Research On Acoustical Detection Of Seabed Sediment With Multi-beam Sonar

The investigation on seabed sediment classification has wide application in the field of military and civil purpose. In nowadays, the only exact way to classify various sediment types is sampling from seabed. But it costs a lot of time and need great effort. The type of seabed sediment has relationship with its acoustic parameters; The acoustic method to classify seabed sediment can improve the efficiency of the task. Multi-beam sonar can offer date source with high precision and high efficiency, and neural network can improve the classification task's precision.In this dissertation, a project of seabed sediment classification based on multi-beam sonar was designed. Firstly, four acoustic parameters were introduced. The need to differentiate between various sediment types with multi-beam sonar and neural network was discussed. And then, the processing of seabed backscatter strength data based on Multi-beam sonar data was investigated. By interchanging the roles of transmitting array and receiving array with each other, the beam footprints rotate 90°,which called rotated multi-beam detected project, through this the sediment in the same place can be detected in multi-angle time after time. The date from rotated multi-beam detective project can improve the precision of multi-beam seabed backscatter strength data processing. According to the idiographic request of sediment classification task, fuzzy ART neural network was chosen to differentiate various sediment types and simulations were made to validate the correctness of classification arithmetic.