Unmanned underwater operations using remotely operated vehicles or unmanned surface vehicles are increasing in recent times, and this guarantees human safety and work efficiency. Optical cameras and multi-beam sonars are generally used as imaging sensors in underwater environments. However, the obtained underwater images are difficult to understand intuitively, owing to noise and distortion. In this study, we developed an optical and sonar image fusion system that integrates the color and distance information from two different images. The enhanced optical and sonar images were fused using calibrated transformation matrices, and the underwater image quality measure (UIQM) and underwater color image quality evaluation (UCIQE) were used as metrics to evaluate the performance of the proposed system. Compared with the original underwater image, image fusion increased the mean UIQM and UCIQE by 94% and 27%, respectively. The contrast-to-noise ratio was increased six times after applying the median filter and gamma correction. The fused image in sonar image coordinates showed qualitatively good spatial agreement and the average IoU was 75% between the optical and sonar pixels in the fused images. The optical-sonar fusion system will help to visualize and understand well underwater situations with color and distance information for unmanned works.

Integrating underwater optical wireless communication (UOWC) with marine applications such as underwater sensors, buoys, and marine surface vehicles (MSVs), requires the aligning and maintaining of the optical beam between the transmitter and receiver for point-to-point (P2P) UOWC during the data transmission. An additional issue is the difficulty in exchanging batteries for marine applications because of the relatively high costs and risks compared with battery exchanging in terrestrial applications. This study seeks to resolve these issues via joint optimization of the beam divergence angle and transmission power level in an underwater sensor (i.e., transmit node) to maintain a seamless connection with an MSV (i.e., receive node) while minimizing the battery consumption of the sensor. In this regard, we adopt a hybrid underwater acoustic-optical communication system, where acoustic and optical communications are used for low-rate control data transmission and high-rate sensing data transmission, respectively. Under this framework, we propose a two-phase deep reinforcement learning (TPDRL) algorithm considering two agents (inner and outer) that determine different actions using an underwater sensor. Specifically, the primary role of the outer agent is to choose a transmission power level based on the long-term signal-to-noise ratio (SNR) between the underwater sensor and MSV. Next, the inner agent finds the beam divergence angle for the given transmission power (selected from the outer agent) based on the short-term instantaneous SNR. Simulation results demonstrate that the proposed TPDRL algorithm enables seamless and energy-efficient P2P UOWC, performing better than the algorithm with only the inner agent and other existing algorithms.

Red tides caused by Margalefidinium polykrikoides occur continuously along the southern coast of Korea, where there are many aquaculture cages, and therefore, prompt monitoring of bloom water is required to prevent considerable damage. Satellite-based ocean-color sensors are widely used for detecting red tide blooms, but their low spatial resolution restricts coastal observations. Contrarily, terrestrial sensors with a high spatial resolution are good candidate sensors, despite the lack of spectral resolution and bands for red tide detection. In this study, we developed a U-Net deep learning model for detecting M. polykrikoides blooms along the southern coast of Korea from PlanetScope imagery with a high spatial resolution of 3 m. The U-Net model was trained with four different datasets that were constructed with randomly or non-randomly chosen patches consisting of different ratios of red tide and non-red tide pixels. The qualitative and quantitative assessments of the conventional red tide index (RTI) and four U-Net models suggest that the U-Net model, which was trained with a dataset of non-randomly chosen patches including non-red tide patches, outperformed RTI in terms of sensitivity, precision, and F-measure level, accounting for an increase of 19.84%, 44.84%, and 28.52%, respectively. The M. polykrikoides map derived from U-Net provides the most reasonable red tide patterns in all water areas. Combining high spatial resolution images and deep learning approaches represents a good solution for the monitoring of red tides over coastal regions.

Underwater optical images face various limitations that degrade the image quality compared with optical images taken in our atmosphere. Attenuation according to the wavelength of light and reflection by very small floating objects cause low contrast, blurry clarity, and color degradation in underwater images. We constructed an image data of the Korean sea and enhanced it by learning the characteristics of underwater images using the deep learning techniques of CycleGAN (cycle-consistent adversarial network), UGAN (underwater GAN), FUnIE-GAN (fast underwater image enhancement GAN). In addition, the underwater optical image was enhanced using the image processing technique of Image Fusion. For a quantitative performance comparison, UIQM (underwater image quality measure), which evaluates the performance of the enhancement in terms of colorfulness, sharpness, and contrast, and UCIQE (underwater color image quality evaluation), which evaluates the performance in terms of chroma, luminance, and saturation were calculated. For 100 underwater images taken in Korean seas, the average UIQMs of CycleGAN, UGAN, and FUnIE-GAN were 3.91, 3.42, and 2.66, respectively, and the average UCIQEs were measured to be 29.9, 26.77, and 22.88, respectively. The average UIQM and UCIQE of Image Fusion were 3.63 and 23.59, respectively. CycleGAN and UGAN qualitatively and quantitatively improved the image quality in various underwater environments, and FUnIE-GAN had performance differences depending on the underwater environment. Image Fusion showed good performance in terms of color correction and sharpness enhancement. It is expected that this method can be used for monitoring underwater works and the autonomous operation of unmanned vehicles by improving the visibility of underwater situations more accurately.

Prompt in-situ monitoring of sea radiation contamination is crucial to prevent the sea from spreading radioactive materials to other parts of the world and the hazard of internal exposures in marine life and the human through food chain. In this paper, we evaluated the performance of a new ocean radiation monitoring robotic boat prototype system. The system featured real-time and on-site gamma spectroscopy with a portable detector of a CsI(Tl) crystal and a PIN diode carried by an unmanned surface vehicle. The system performed good energy linearity and energy resolution (21% at 662 keV). The efficiency and the minimum detectable activity (24 h acquisition) for underwater ¹³⁷Cs detection were 1.334 × 10⁻⁵ and 2.5 Bq/kg, respectively. We conducted a Monte Carlo simulation to estimate the detection performance with uniform and point source distributions. In the water tank experiments, we measured the count rates with differing depths of a ¹³⁷Cs point source. The maximum detectable distance was 50 cm according to both experiment and simulation results. Additionally, on-site experiment was conducted to test the operability of the system in a real sea. The overall study showed the feasibility of the system for in-situ ocean radiation monitoring.

Large-sized crystals and state-of-the-art photosensors are desirable to cope with low environmental radioactivity (e.g., 1-2 Bq∙m^-3137Cs in surface seawater) for homeland security purposes. We compared the performances of two different gamma-ray detector assemblies, GAGG crystal + silicon photomultiplier (SiPM) and NaI(Tl) crystal + photomultiplier tube, for our mobile in-situ ocean radiation monitoring system. We performed energy calibration, followed by water tank experiments with varying the depth of a¹³⁷Cs point source. Experimental energy spectra were compared with MCNP-simulated spectra with identical setup and the consistency was validated. We finally assessed the detection efficiency and minimum detectable activity (MDA) of the detectors. Both GAGG and NaI detectors exhibited favorable energy resolutions (7.98 ± 0.13% and 7.01 ± 0.58% at 662 keV, respectively) and MDAs (33.1 ± 0.0645 and 13.5 ± 0.0327 Bq∙m^-3 for 24-h ¹³⁷Cs measurement, respectively). Matching the geometry of the GAGG crystal with that of the NaI crystal, the GAGG detector outperformed the NaI detector. The results demonstrated that the GAGG detector is potentially advantageous over the NaI detector in detection efficiency and compactness.