Investigations into deep-sea ecosystems are essential for elucidating the origins of life on Earth. Nevertheless, constraints in marine biotechnology have limited our understanding of biological processes occurring in these environments. In this study, we developed and evaluated an in situ cultivation system specifically designed for deep-sea environments. This system enables continuous induction cultivation directly on the seabed and facilitates comprehensive monitoring of the entire cultivation process via an imaging technique. The system is composed of three distinct modules: a cultivation and observation module, a methane slow-release module, and a power management module. This system enables controlled, slow release of energy and materials, allowing for the cultivation and sampling of communities on the seafloor, and supports long-term sequential monitoring through imaging. Through three deployment trials, we successfully established an artificial methane-driven deep-sea ecosystem on the seafloor of the South China Sea. Furthermore, by modifying the types of energy and materials supplied, the system can be adapted to address diverse scientific objectives, offering a robust tool for advancing research on deep-sea life.

Deep-sea mining generates plumes containing heavy metals harmful to marine ecosystems. Assessing their impact on benthic organisms is essential. Traditional monitoring methods commonly employ acoustic, optical, and electrical approaches to investigate collective behaviors of organisms, while direct acquisition and in situ preservation of biological specimens during deployment remain challenging. To address this limitation, a novel sequential near-bottom mobile organisms sampling equipment equipped with multiple chambers was developed. Sequential sampling of near-bottom mobile organisms was conducted to obtain samples at different time points for individual-level biological assessment. This study investigates the alcohol diffusion dynamics at various injection volumes. The results show that during injection, alcohol is concentrated in the upper section of the chamber. When 5.0 L of alcohol is introduced, the alcohol mass fraction within the sampling container rapidly reaches and stabilizes at 87%. The sampling equipment’s performance was evaluated through dock experiments and deep-sea trials at a depth of 1571 m in the South China Sea. Field tests successfully captured five jellyfish, with alcohol mass fractions of 85%, 87%, and 87% achieved in the respective chambers. These results demonstrate the feasibility of the equipment for sequential biological sampling and in situ preservation in deep-sea environments.

The spatiotemporal characterization of plume sedimentation and microorganisms is critical for developing plume ecological monitoring model. To address the limitations of traditional methods in obtaining high-quality sediment, a novel sampling system with 6000 m operational capability and three-month endurance was developed. It is equipped with three sediment samplers, a set of formaldehyde preservation solution injection devices. The system is controlled by a low-power, timing-triggered controllers. To investigate low-disturbance rheological mechanisms, gap controlled rheological tests were conducted to optimize the structural design of the sampling and sealing assembly. Stress-controlled shear rheological tests were employed to investigate the mechanisms governing yield stress in sediments under varying temperature conditions and boundary roughness. Additionally, the coupled Eulerian-Lagrangian (CEL) method and sediment rheological constitutive models were employed to simulate tube-soil interaction dynamics and sediment disturbance. The radial heterogeneity of sediment disturbance and friction variation of the sampling tube were revealed. The tube was completely “plugged” at a penetration depth of 261 mm, providing critical data support to the penetration depth parameters. The deep-sea pressure test and South China Sea field trials demonstrated the system’s capability to collect and preserve quantitative time-series sediment samples with high fidelity.

With the growing demand for marine resource development, the detection of marine benthic organisms has become increasingly important for ecological protection, biodiversity, and resource management. However, traditional detection algorithms face challenges in complex marine environments, and advancements in deep learning technology provide new solutions. To enhance the precision and performance of detecting benthos in marine settings, this paper introduces an improved version of the Faster R-CNN model. The key enhancements include using ResNet50 for feature extraction, incorporating a multi-scale channel attention mechanism to highlight critical features and accommodate multi-scale information, replacing RoIPooling with RoIAlign to reduce quantization errors, and substituting the fully connected layer with an extreme learning machine classifier to enhance generalization and accuracy. Through extensive ablation experiments and comparative analyses, this paper demonstrates that the proposed improved model achieves higher detection accuracy compared to current mainstream detection models. Specifically, on the URPC2020 dataset, the model achieved approximately a 5

The spatial distribution and assemblages of fish larvae related to monsoon in the northern South China Sea were investigated in this study. A total of 1710 fish larvae were collected at 19 stations during May (intermonsoon) and August (southwest monsoon). The samples contained 171 distinct taxa belonging to 97 genera and 72 families. Cluster analysis revealed that the spatial structure of larvae in the northern South China Sea is distinctly divided into two groups: the shelf and the slope group, and these differences correspond well with the South China Sea slope. Zooplankton, chlorophyll a, surface sea temperature, sea salinity at 10 m, and dissolved oxygen concentration were the key environmental factors affecting the distribution of ichthyoplankton in the northern South China Sea. From the intermonsoon period to the southwest monsoon period, a notable shift occurred in the latitudinal distribution trend of fish larvae on the slope. Concurrently, the dominant zooplankton groups that shape the distribution pattern of fish larvae transited from chaetognaths and cnidaria to smaller calanoids.

Molluscs living in dynamic deep-sea cold seep environments have evolved distinct feeding strategies for survival. Here, we present the chromosome-level genomes of two sympatric mollusc species with distinct feeding strategies, a symbiosis-dependent mussel Gigantidas haimaensis and a predatory snail Phymorhynchus buccinoides . Comparative genomic analysis revealed gene family expansions related to the bacterial component degradation (e.g., b4GalT s) in G. haimaensis , suggesting an adaptation to symbiotic life. Conversely, P. buccinoides exhibited gene family expansions associated with appetite regulation (e.g., ox2r ) and the digestive system (e.g., sult1 and chst ), indicating genetic modifications for deep-sea predation. Furthermore, we conducted an in situ experiment mimicking a scenario in which ocean warming and sea-level rise resulted in a mass methane leakage in deep-sea cold seeps. Interestingly, G. haimaensis increased its metabolic rate and exhibited transcriptional responses. However, P. buccinoides suppressed energy production and responses at translational and posttranslational levels, which is compatible with their distinct feeding strategies. Collectively, our results provide insights on the evolutionary basis and resilience mechanisms related to energy management, which may facilitate methane tolerance of molluscs in the deep-sea cold seeps threatened by climate change.

It is important to marine ecology research that plankton samples are collected without damage, especially for time series samples. Usually, most fixed-point plankton samplers are made using a pump with paddle blades in order to increase the flow rate. But it can easily injure soft plankton. In this paper, a shaftless hollow sampling pump is designed, which can provide a highly efficient driving component for the plankton sampler. The numerical model of the sampling pump is established, and the flow rate of the sampling pump at different rotational speeds is simulated by the computational fluid dynamics method. In order to obtain a higher flow rate, the influence of internal and external cavity size, blade angle, and blade number on the flow rate of the sampling pump with a constant rotational speed of the blade was simulated and discussed. The results show that the flow rate at the internal cavity is positively correlated with the inlet and outlet pressure differences of the internal cavity, and the greater the negative pressure at the outlet of the internal cavity, the greater the flow rate. When the internal and external cavity sizes are h = 14 mm, d = 52 mm, blade angle θ = 45°, and number of blades s = 5, the flow rate of the sampling pump internal cavity reaches the maximum. Finally, the feasibility of the shaftless hollow sampling pump is verified by experiments. The shaftless hollow sampling pump can realize non-destructive sampling of plankton. This paper presents a theoretical design foundation for a new non-destructive siphon sampling method for marine plankton, which is of great significance for marine plankton sampling and subsequent research.

Aquatic organisms serve as crucial indicators of ecosystem health and water quality conditions. Accurate classification and monitoring of aquatic organisms facilitate the timely detection of ecological environmental changes, which is of significant importance for the conservation of biodiversity and environmental protection. Pointing at the issues of fuzzy images and difficult classification caused by the poor lighting conditions of aquatic organisms and many impurities around, we put forward a novel classification model of aquatic organisms with convolution neural network, feedforward network and optimization algorithm. The model first abstracts the traits of aquatic organisms, and then classifies the abstracted features. Through migration learning of ResNet101 neural network on Fish4Knowledge dataset and ISLVRC2012 and some data sets related to aquatic organisms, a network that will be used to abstract features is obtained. The model put forward in the paper firstly uses ResNet101 to abstract the aquatic organism image’s characteristics, and then uses the Golden Eagle optimization algorithm (GEO) optimized by chaos mapping optimization and the RVFL neural network optimized by L2,1 regularization to classify aquatic organisms. The algorithm is compared with other RVFL algorithms optimized by Ant Lion Optimizer (ALO), Grey Wolf Optimization (GWO), Particle Swarm Optimization (PSO), Harris Hawks Optimization (HHO) and Dragonfly Algorithm (DA). The findings indicate that the methodology introduced in this paper exhibits an improvement over the original ResNet101, with increases of 1.58

Submarine hydrothermal plumes play an important role in the process of material and energy exchange in the deep sea. Due to the influence of complex environmental factors (multiple vents, deep-sea stratified environment) and the limitation of observation capabilities, the dynamic characteristics of submarine hydrothermal plumes have not been fully understood, and it is urgent to study the diffusion and evolution process of submarine hydrothermal plumes by the computational fluid dynamics (CFD) model. In this paper, a CFD model was established to effectively simulate the diffusion and evolution process of two-vent submarine hydrothermal plumes under the deep-sea stratified environment, obtained the three-dimensional plume flow field structure of two-vent submarine hydrothermal plumes, and obtained the spatial distribution of dynamic characteristics parameters such as velocity, turbulent viscosity, turbulent kinetic energy and turbulence dissipation rate of two-vent submarine hydrothermal plumes. The equations for calculating the maximum plume rise height and the neutrally buoyant plume height of a two-vent submarine hydrothermal plume were deduced and established. In addition, a comparative analysis of the single-vent submarine hydrothermal plume revealed the mechanism of the influence of the key characteristics of the two-vent submarine hydrothermal plumes such as the maximum plume rise height. It was applied to analyze the dynamic characteristics of the plume occurred in Longqi hydrothermal field in the southwest Indian Ocean, and the results were verified by the field observation data. The spatial distribution of mass concentration of hydrothermal plume particles was further analyzed, and the difference of mass concentration in different hydrothermal plume structures was compared to study the exchange mechanism of material and energy between submarine hydrothermal plume and surrounding seawater. The results are of guiding significance for the tracing of submarine hydrothermal vents, the study of marine material exchange and energy transport processes, and the exploration and environmental evaluation of submarine polymetallic sulfide resources.

Downwelling aeration has become a widely applied approach to cope with the water eutrophication in stratified reservoirs, rivers and lakes. The aeration parameters involving flow rate, flow locations and working periodicity and their impacts on the temperature and dissolved oxygen (DO) distributions of water have been largely unclarified, causing extra time and energy consumptions in practice. In this study, a home-built water tank and an aeration pump are used to model the downwelling aeration processes in stratified water. Temporal influences of aeration parameters on the water stratifications and eutrophicated elements are systemically investigated, with the purpose of searching parametric configurations to enhance the anti-eutrophication efficiency. It is found that the variation rates of temperature destratification and DO distribution in the water body could be saturated and strongly correlated with the flow rate. Based on such experimental saturation rates, we find an optimized working condition from the aspect of energy saving: a 300 rpm pump speed and a 15 cm distance between the flow exit and the sediment surface. In such conditions, the total nitrogen and phosphorus dissolved in the bottom layer of water decrease exponentially with aeration time, and can be reduced by 53.8 and 86% in the first 6 h of aerations, respectively, taking full advantage of the microbial bonding to the sedimentations. The present work provides better understandings for efficient implementations of downwelling aerations.

水下云台是深海装备的重要组成部件.本文基于干式密封结构设计,研制了一套直流伺服电机驱动的深海二自由度云台,其具备更优的转向控制精度,同时可实现大负载、高转速和实时角度反馈功能.云台主体采用陶瓷角轴承运动承力、蜗轮蜗杆自锁、格莱圈旋转密封等设计,实现云台的俯仰旋转和水平旋转,并对云台关键部件进行强度校核,确保水下工作的稳定性.干式云台可内置磁编码器及主控电路,实现深海移动观测和目标跟踪所需的高精度反馈控制.转向测试与压力测试结果证明,所设计的云台可以实现5 kg负载下30(°)/s的高速精准转向,工作水深可达2 000 m.

Due to the complexity of the underwater environment, tracking underwater targets via traditional particle filters is a challenging task. To resolve the problem that the tracking accuracy of a traditional particle filter is low due to the sample impoverishment caused by resampling, in this paper, a new tracking algorithm using Harris-hawks-optimized particle filters (HHOPF) is proposed. At the same time, the problem of particle filter underwater target feature construction and underwater target scale transformation is addressed, the corrected background-weighted histogram method is introduced into underwater target feature recognition, and the scale filter is combined to realize target scaling transformation during tracking. In addition, to enhance the computational speed of underwater target tracking, this paper constructs a nonlinear escape energy using the Harris hawks algorithm in order to balance the exploration and exploitation processes. Based on the proposed HHOPF tracker, we performed detection and evaluation using the Underwater Object Tracking (UOT100) vision database. The proposed method is compared with evolution-based tracking algorithms and particle filters, as well as with recent tracker-based correlation filters and some other state-of-the-art tracking methods. By comparing the results of tracking using the test data sets, it is determined that the presented algorithm improves the overlap accuracy and tracking accuracy by 11% compared with other algorithms. The experiments demonstrate that the presented HHOPF visual tracking provides better tracking results.

To overcome the challenges posed by the underwater environment and restore the true colors of marine objects' surfaces, a novel underwater image illumination estimation model, termed the iterative chaotic improved arithmetic optimization algorithm for deep extreme learning machines (IAOA-DELM), is proposed. In this study, the gray edge framework is utilized to extract color features from underwater images, which are employed as input vectors. To address the issue of unstable prediction results caused by the random selection of parameters in DELM, the arithmetic optimization algorithm (AOA) is integrated, and the search segment mapping method is optimized by using hidden layer biases and input layer weights. Furthermore, an iterative chaotic mapping initialization strategy is incorporated to provide AOA with a better initial search proxy. The IAOA-DELM model computes illumination information based on the input color vectors. Experimental evaluations conducted on actual underwater images demonstrate that the proposed IAOA-DELM illumination correction model achieves an accuracy of 96.07%. When compared to the ORELM, ELM, RVFL, and BP models, the IAOA-DELM model exhibits improvements of 6.96%, 7.54%, 8.00%, and 8.89%, respectively, making it the most effective among the compared illumination correction models.

In this paper, the numerical models are selected to simulate the hydrothermal plume based on the water temperature observation data of the Longqi hydrothermal field in the Southwest Indian Ridge (SWIR). Then, the unsteady Reynolds-averaged Navier–Stokes equations are solved to evaluate the performance of the Realizable k-ε (rke) model and the SST k-ω (sst) model in hydrothermal plume simulation. By comparing the calculated results with the Conductivity Temperature Depth (CTD) observation data and the literature results, the difference in prediction performance between the two models is evaluated. Before the numerical simulation, the optimal mesh parameters are determined by considering the grid independence test. The results show that the relative difference of the maximum plume height calculated by the two models is within 5%. Compared with the CTD 05-2, the rke model calculates the root mean square error of the velocity is 0.5081, which is smaller than that of the sst model. In terms of turbulent viscosity, the rke model is in good agreement with reference value in predicting turbulent viscosity. Therefore, the turbulent viscosity distribution calculated by the rke model is more consistent with the plume development process than that calculated by the sst model. In addition, the two models have the same effect on the prediction of turbulent kinetic energy and plume temperature.

Aiming at the problem of poor image illumination correction accuracy, a kernel extreme learning machine optimized based on the differential evolution-improved marine predators algorithm is proposed to estimate the scene illumination information and restore the image. First, in order to solve the problem of randomization of the initial population of the marine predators algorithm, differential evolutions is used to provide a set of suitable initial populations to the algorithm. Then, an improved algorithm is used to optimize the weight and bias of the kernel extreme learning machine to solve the problem of the randomness of the weight and bias, so that the learning machine converges to the global optimal value and avoids the unstable predictions. Finally, after obtaining the predicted illumination information, the image model is corrected to the image effect under standard illumination through diagonal transformation. It can be seen from the experimental results that the best predictive value of the differential evolution marine predator algorithm learning machine is 0.0135915, and the quasi-bias is only 0.001687. Compared with traditional illumination estimation algorithms such as random vector functional link and extreme learning machine, the differential evolution marine predator algorithm learning machine algorithm proposed in this article has better stability, higher accuracy of calculated predicted values, and better image restoration effect.

To overcome the limitations of the traditional sliding mode control (SMC) method, including steady-state errors, low tracking accuracy, external disturbances, and difficulties in estimating uncertain parameters, a nonlinear integral SMC method with an adaptive extreme learning machine (ELM) and a robust control term is developed. First, the ELM is used to approximate the uncertain parameters in the manipulator dynamics model to improve the tracking accuracy of the manipulator. In this process of using an ELM to approximate uncertain parameters, a method of adaptively updating the output weight is applied to improve the stability and closed-loop tracking accuracy of the system and ensure the real-time performance of manipulator control. A new nonlinear integral sliding mode function is designed to reduce the steady-state error of the system, avoid the issue of system instability caused by large initial errors of the system, and enable the system to track the desired trajectory quickly. Moreover, a robust control term is added to the SMC law to compensate for the error of ELM approximation, which increases the robustness of the manipulator and reduces the fluctuation amplitude of the control input in the presence of external disturbances. Finally, a simulation analysis is performed and stable convergence of the proposed method using Lyapunov stability functions is demonstrated.

The color of an object appears different from its true color when illuminated with light sources of different hues. To solve this problem, this article proposes a combination algorithm (SCA-GWO-LSSVR) based on the sine-cosine algorithm (SCA) and the gray wolf optimization (GWO) algorithm to optimize the regression prediction model of the least-squares support vector regression (LSSVR) algorithm. The performance of the traditional LSSVR is significantly affected by the penalty parameter (gamma) and the sig2 kernel function parameter. The proposed method uses the improved GWO algorithm to search the population to find the best LSSVR parameter solution. The proposed algorithm uses the SCA to create multiple random candidate solutions in population initialization to avoid blind initialization of the GWO algorithm. In the process of iterative optimization, the SCA is infiltrated, and its sine-cosine wave mathematical model is used to quickly identify the best outward or inward position of the gray wolf. Finally, the LSSVR combines the optimal sig2 kernel function parameters and penalty parameters (gamma) to obtain a highly versatile illumination correction model. The experimental results show that the fitting accuracy of the proposed method reaches 86.8%, which is 5% higher than that of the LSSVR algorithm alone.

To automatically eliminate the influence of external illumination on the acquired image, restore the true color information of the object, and provide true and stable color features for computer vision tasks, our work is based on the convolutional neural network (CNN) VGG19 for feature extraction of image information. Moreover, it proposes a model based on atom search optimization improved by chaotic-logistic maps to optimize the regularization random vector functional link (RRVFL). The optimized RRVFL replaces CNN's regression layer to estimate the image illumination and then restore the image. First, we used chaotic-logistic maps to optimize the initial value of search agent of the atom search optimization (CASO) algorithm and then used the search agent segment mapping method proposed in our study to simultaneously optimize the number of nodes in hidden layer, regularization factors, input weight, and the bias of RRVFL hidden layer. This avoids the problem of serious fluctuation of prediction results and low accuracy caused by the randomness of the RRVFL parameters. After the predicted illumination information was obtained with the optimized RRVFL, the image was restored using the diagonal transformation method. Comparative experiments showed that the average angular errors of the illumination estimation of the V19-CASO-RRVFL algorithm proposed in our study were 1.73, 0.30, 0.3406, and 0.0976 lower than those of the gray-edge-2, GE-CASO-RRVFL, DS-Net, and color constancy on deep residual learning algorithms, respectively. (C) 2022 Society of Photo-Optical Instrumentation Engineers (SPIE)

Variations in illumination lead to serious errors in the evaluation of chromatic aberration. We propose an illumination correction model based on the opposition-based learning improved whale optimization algorithm for support vector regression optimization, named "OBL-IWOA-SVR." Because the initial population quality of the whale optimization algorithm has a significant impact on the solution speed and accuracy, the Opposition-based learning strategy is adopted in this article to mix the original population and its opposite individuals and select the best as the new population, replacing the random initialization to generate a more suitable initial population. This increases the diversity of the population and thus overcomes the impact of the quality of the initial population on the algorithm performance. Secondly, the proposed algorithm adopts the adaptive weight and the convergence factor based on the variation of the cosine law to balance the algorithm's global exploration ability and local development ability and enhance the convergence accuracy. Finally, the algorithm utilizes good global searching ability to optimize the penalty factor and nuclear parameters and obtains the optimal support vector machine parameter combination to construct the illumination correction model OBL-IWOA-SVR accurately and robustly. Experimental results show that the illumination correction model proposed in this article is found superior to other models in significance analysis: the root mean square error of the proposed model is 0.0173, the smallest of all illumination correction models. Furthermore, the model exhibits good stability and high illumination estimation accuracy.

The multi-net visible fidelity zooplankton collector is designed to obtain near-bottom fidelity zooplankton. The collector is sent to the designated sampling location based on the information provided by the camera and altimeter. The host computer sends instructions to control the opening of the net port for sample collection and closing of the sampling cylinder cover after sampling. The collector contains three trawls so that three samples can be collected for each test, and environmental parameters can be collected simultaneously. After sampling, The sample maintains its fidelity, that is, maintaining the temperature and pressure of the seabed sample after sampling. Two experiments were carried out in the Western Pacific, and six bottles of zooplankton samples were successfully obtained. The development of a multi-net visible zooplankton collector is of great significance for the collection of near-bottom zooplankton.