Research On Information Enhancement And Fusion Of Underwater Perception Networks

With the increasing scarcity of land resources and the escalating risk of marine environmental pollution,underwater sensing networks that support various marine activities have become a research hotspot.As an essential tool for developing and utilizing marine resources,underwater sensing networks rely on acoustic signals for wireless connectivity.These networks randomly deploy mobile units capable of data acquisition,such as autonomous robots and submarines,along with cost-effective fixed sensor devices(e.g.,buoys and monitors)on the seabed or in areas of interest.This deployment approach forms a comprehensive network that is multi-hop,decentralized,node-rich,and covers a wide area,enabling in-situ monitoring,collection,processing,and aggregation of underwater information.However,compared to terrestrial wireless networks,underwater sensing networks face more challenges.Firstly,the complex and variable acoustic conditions in the underwater environment pose a threat to network stability.Secondly,the uncontrollable drifting phenomenon of underwater nodes caused by water currents also brings difficulties to the stable operation of the network.Furthermore,the harsh underwater environment greatly increases the difficulty of information acquisition.Moreover,underwater sensing networks typically employ distributed sensing methods based on different information sources,which presents a significant challenge for real-time panoramic perception.Addressing the limited effective information rate caused by equipment limitations and environmental variability,as well as providing longer-lasting effective node detection and sensing,are crucial issues faced by underwater sensing networks.Therefore,to meet the requirements of high-precision and high-efficiency marine observation,detection,and information acquisition over a wide area,this paper focuses on three aspects: underwater information enhancement,acousticoptical information fusion,and network topology optimization.The main challenges of information enhancement,fusion,and network topology optimization are analyzed,and corresponding solutions are proposed to address the aforementioned problems.The main contributions of this paper are as follows:(1)Information Enhancement Method Based on Frequency Separation and Image Fusion:The information acquisition in underwater sensing networks requires high-precision information capture of targets.Compared to terrestrial environments,the underwater environment suffers from low effective information acquisition rates due to equipment volume limitations,scattering effects,and biological camouflage protection characteristics.Therefore,this paper focuses on information enhancement tasks in underwater sensing networks.To address the issues of texture detail loss and color distortion caused by low resolution of visual information acquisition devices,a frequency separation and image fusion-based information enhancement method is proposed.This method can reconstruct richer texture details in underwater visual information,fuse pixels to restore color and information quality,and solve the problem of low effective information extraction caused by low-cost acquisition devices and harsh environments in sensing networks.(2)Rapid Perception Method Based on Acoustic-Optical Fusion: Effective detection and identification are key prerequisites for conducting interdisciplinary and cross-domain research and operations.Due to the differences in carrying capacity,detection means,and cognitive regions of underwater mobile units,the execution efficiency of clusters in tasks such as security monitoring,location guidance,and target pursuit is greatly affected.To improve the detection efficiency of sensing networks,considering the complementary characteristics of acoustic and optical devices in terms of sensing range,sensing accuracy,and sensing cost,this paper proposes a rapid perception method based on acoustic-optical fusion.This method utilizes acoustic technology for rapid target identification and localization,enhancing the precision and accuracy of sonar information extraction and perception.It also employs optical technology for three-dimensional reconstruction of targets,improving the accuracy and efficiency of target 3D reconstruction.The multi-source perception approach based on acoustic-optical fusion will have significant reference value for underwater activities and operations.(3)Network Topology Optimization Algorithm Based on Dynamic Prediction: As the mobility of underwater nodes inevitably increases the difficulty,deployment cost,and complexity of network topology optimization,this paper focuses on network topology optimization.Addressing the complex and variable marine environment and energy-constrained issues,a dynamic prediction-based network topology optimization strategy is proposed.Firstly,a prediction model for node movement under water currents is constructed,comprehensively considering the current position and predicted position of nodes for optimization decision-making,resulting in an optimal topology structure.Unlike existing methods,the optimization strategy proposed in this paper can consider the impact of multiple factors such as node movement under water currents and network end-to-end delay,ensuring the reliability,connectivity,and maximum coverage of sensing network nodes in both temporal and spatial dimensions.