Background normalization algorithms attempt to suppress the ambient and self-noise during the measurements of sonar, which enhance the detection performance and the display effect of weak signals. Conventional background normalization methods are usually sensitive to the accuracy of prior set filtering interval and threshold, while significant noise is still detected in low frequency. In this paper, an improved background normalization algorithm is proposed by thresholding the processing interval between several local peak values and local valley values. Compared to the existing scenarios, the proposed approach automatically calculates the filtering interval and threshold, with substantial resilience to the noise level in low frequency. Experimental results illustrate the effectiveness of our algorithm.

Sound source recognition is a very important application of passive sonar. How to distinguish between surface and underwater acoustic sources has always been a challenge. Due to the mixing of underwater target radiated noise and marine environmental noise, especially in shallow water environments where multipath effects exist, it is difficult to distinguish them. To solve the surface and underwater acoustic source recognition problem, this paper proposes a multi-channel joint detection method based on machine learning. First, the simulation data are generated using the normal model KRAKEN setting in the same environment as the SACLANT 1993 experiment, which uses a vertical linear array of 48 hydrophones. Secondly, the GBDT classifier and LightGBM classifier are trained separately, and then the model is evaluated using precision, recall, F1, and accuracy. Finally, four ML models (kNN, random subspace kNN, GBDT, and LightGBM) are used to analyze all 48 channels of hydrophone data. For each model, two kinds of feature extraction methods (module features, real and imaginary features) are applied. Generally, the results show that both GBDT and LightGBM models have better performance than both kNN and random subspace kNN ones. For both GBDT and LightGBM models, the results using module features have better performance than using real and imaginary features.

The absence of natural sources of circularly polarized light has created a critical demand for high-performance polarization converters in emerging technologies such as quantum information processing, biomedical diagnostics, and advanced magneto-optical recording systems. Despite significant efforts, linear-to-circular polarization converters that combine high efficiency with tunability remain scarce. In this paper, we present a multifunctional and tunable device with 99.9% efficiency for converting linear to circular polarization in the telecom wavelength range. It can also function as a phase-only compensator with a controllable phase range of 1.93$\pi$ while maintaining a reflectance above 99%. The angular sensitivity analysis indicates that the device can tolerate incidence angles ranging from 8 to 10.4 degrees. This high-efficiency polarization conversion device offers significant advantages in integrated optics due to its multifunctionality and tunability and has promising applications in optical phased-array radars and quantum communication.

Ocean acoustic-tomography (OAT) inversion depends on relative clock offsets, transducer positions, the sound-speed profile, and the system time offset, which are unknown and coupled. We propose a segmented Bayesian method for their unified calibration. Reciprocal travel-time half-differences are used in a linear inversion of relative clock offsets, whereas half-sums are used in a nonlinear joint inversion of transducer positions, a low-order modal sound-speed profile, and the system time offset, with posterior uncertainties quantified. Under the applicability conditions, the joint posterior factorizes by the two parameter groups, supporting separate two-stage estimation. Field data from a shallow-water test site show that, within the principal 4–14 m ray-sampling range, the posterior profile reduces the root-mean-square error (RMSE) against an independent measured profile by approximately 41% (at the adopted data weighting) relative to the prior, mainly by correcting depth-mean sound speed. In temporal extrapolation, parameters calibrated over the first 40 h predict one-way travel times over the following 11 h; prediction residuals and systematic bias are substantially reduced. Ratios of post-calibration residual RMSE to short-term channel variability are 0.87–1.70, indicating residuals remain of the same order as intrinsic channel variability. The method calibrates OAT system parameters and improves prediction for data excluded from the inversion.