Active Sonar Target Classification with Power-Normalized Cepstral Coefficients and Convolutional Neural Network

Abstract

Detection and classification of unidentified underwater targets maneuvering in complex underwater environments are critical for active sonar systems. In previous studies, many detection methods were applied to separate targets from the clutter using signals that exceed a preset threshold determined by the sonar console operator. This is because the high signal-to-noise ratio target has enough feature vector components to separate. However, in a real environment, the signal-to-noise ratio of the received target does not always exceed the threshold. Therefore, a target detection algorithm for various target signal-to-noise ratio environments is required; strong clutter energy can lead to false detection, while weak target signals reduce the probability of detection. It also uses long pulse repetition intervals for long-range detection and high ambient noise, requiring classification processing for each ping without accumulating pings. In this study, a target classification algorithm is proposed that can be applied to signals in real underwater environments above the noise level without a threshold set by the sonar console operator, and the classification performance of the algorithm is verified. The active sonar for long-range target detection has low-resolution data; thus, feature vector extraction algorithms are required. Feature vectors are extracted from the experimental data using Power-Normalized Cepstral Coefficients for target classification. Feature vectors are also extracted with Mel-Frequency Cepstral Coefficients and compared with the proposed algorithm. A convolutional neural network was employed as the classifier. In addition, the proposed algorithm is to be compared with the result of target classification using a spectrogram and convolutional neural network. Experimental data were obtained using a hull-mounted active sonar system operating on a Korean naval ship in the East Sea of South Korea and a real maneuvering underwater target. From the experimental data with 29 pings, we extracted 361 target and 3351 clutter data. It is difficult to collect real underwater target data from the real sea environment. Therefore, the number of target data was increased using the data augmentation technique. Eighty percent of the data was used for training and the rest was used for testing. Accuracy value curves and classification rate tables are presented for performance analysis and discussion. Results showed that the proposed algorithm has a higher classification rate than Mel-Frequency Cepstral Coefficients without affecting the target classification by the signal level. Additionally, the obtained results showed that target classification is possible within one ping data without any ping accumulation.

Developing a supervised machine‐learning model capable of distinguishing fiber orientation of polymer composite samples nondestructively tested using active ultrasonics

“…The STFT was chosen as the first dynamic signal processing technique due to its widely used popularity in time-based signaling events, such as sonar targeting, impact echo testing, and speech recognition. [52][53][54][55][56] With a wide variety of applications in which the time-domain of the signal is of importance, and with the ability to display the main nodal frequency behavior at specific time segments, STFT seemed like an ideal signal processing technique. Due to STFT being an image-based approach for evaluating the acoustic signal, CNN was chosen as it has largely been paired with STFT for image-based classification applications.…”

“…Due to STFT being an image‐based approach for evaluating the acoustic signal, CNN was chosen as it has largely been paired with STFT for image‐based classification applications. [ 54,57–60 ]…”

“…Due to STFT being an image-based approach for evaluating the acoustic signal, CNN was chosen as it has largely been paired with STFT for image-based classification applications. [54,[57][58][59][60] The STFT built the frequency-time domain on small time windows for the entirety of the frequency domain. For the STFT produced images, the color of each pixel represents its amplitude value, while the location of a pixel corresponds to its frequency and time coordinates, from which an image was made as shown in Figure 2B.…”

Real-Time Underwater Acoustic Homing Weapon Target Recognition Based on a Stacking Technique of Ensemble Learning

“…Underwater acoustic target recognition methods based on AI can be divided into statistical machine learning methods and deep learning methods. Among them, the object recognition method based on statistical machine learning includes support vector machine [8,9] and other methods [10], while the object recognition method based on deep learning includes convolutional neural networks [11][12][13][14][15], long short-term memory [16], and other methods [17]. It can be observed that for passive underwater acoustic target recognition, scholars have conducted a substantial amount of research.…”

Study on Small Samples Active Sonar Target Recognition Based on Deep Learning

“…Bu M et al applied several pretrained deep convolutional neural networks (DCNN) to the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), and experimentally demonstrated that their DCNNs outperform classical machine learning methods in active sonar target recognition [9]. Seungwoo Lee used power-normalized cepstral coefficients (PNCC) for feature extraction; this method classifies real underwater target echoes and clutter with convolutional neural networks [10]. Karl Thomas Hjelmervik et al found several good hyperparameter values for DL classifiers in active sonar target classification using Bayesian optimization [11].…”