Full text loading...
Through this study, a novel methodology is proposed where the far field signature (FFS) is estimated using energy wise sorted near field hydrophone (NFH) with the help of Artificial Neural Network (ANN). The NFH of a survey are fed as input and the FFS is the output of the ANN model. The trained model is applied to a separate dataset of NFH from a marine survey and the FFS is estimated. The accuracy achieved for validation and test data is above 95%. The predicted FFS on validation showed excellent correlation (∼97%) with the known FFS. Further research is warranted in this project to make the algorithm more robust, however, the proposed method has shown very good results to encourage more study in this area.