Due to the complex and interdependent world, underwater acoustic source detection is a difficult challenge. It can be extremely difficult to perform Underwater Acoustic Target Recognition (UATR) utilizing ship-radiated noise due to the complicated aquatic organisms. This paper presents a new deep neural network called Audio Perspective based RCNN (APRCNN), learned by initial underwater acoustic waves with Depth Wise Separable (DWS) Convolutional Deep Neural Network. The APRCNN would then be implemented for underwater acoustic signal prediction and classification. The proposed harvester learning features & integration layers of the SPDNN system were motivated by perceptual processing, and the time-dilated convolution was influenced by machine translation. In this article, the researchers provide a methodology for identifying underwater targets using a thick DNN model. The proposed framework was created to reallocate all previous feature maps to maximize the categorization rate while minimizing the calculation burden in various degraded scenarios. Additionally, the proposed technique allows using the actual audio signal in the transfer function as the right network input information rather than using time-frequency spectrum analyzer images. Based on research findings assessed on a real-world collection of passive sonar, our classification method beats other cutting-edge DNN systems & classic ML methods, achieving an accuracy level of 98.4% at 0-dB SNR.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.