Search

CN-122015989-A - AI image enhanced acoustic chromatography multi-mode river flow monitoring system

CN122015989ACN 122015989 ACN122015989 ACN 122015989ACN-122015989-A

Abstract

The invention discloses an AI image enhanced acoustic chromatography multi-mode river flow monitoring system, which relates to the technical field of graphic data sensing and river flow monitoring, and comprises an acoustic chromatography monitoring module, a monitoring module and a monitoring module, wherein the acoustic chromatography monitoring module adopts M sequence spread spectrum to measure acoustic data; the system comprises an image acquisition and AI enhancement module, a multi-mode data synchronous control module, a multi-source data fusion module, an AI intelligent flow calculation module, a data storage and visualization display module, a remote control and parameter configuration module and an Internet of things interface, wherein the image acquisition and AI enhancement module is used for reconstructing and outputting an enhanced image through denoising and superdivision, the multi-mode data synchronous control module is used for realizing millisecond level alignment of acoustic and visual data, the multi-source data fusion module is used for extracting and normalizing multi-dimensional feature vectors by using a cross-mode attention mechanism, the AI intelligent flow calculation module is used for fitting a mapping relation and calculating flow through a neural network, the data storage and visualization display module is used for storing data through an edge and cloud dual architecture, and the remote control and parameter configuration module is used for realizing bidirectional communication through the Internet of things interface. The invention integrates the acoustic chromatography and AI enhanced image technology, has comprehensive monitoring coverage, low power consumption and stable operation, and provides high-efficiency and reliable support for water resource management.

Inventors

  • SHI ZHAOLIANG
  • SU HANSONG
  • WU BO

Assignees

  • 北京中海技创科技发展有限公司

Dates

Publication Date
20260512
Application Date
20260128

Claims (10)

  1. 1. An AI image enhanced acoustic tomography multi-modal river flow monitoring system, comprising: The acoustic chromatography monitoring module consists of at least four acoustic station nodes arranged on two sides of a river section, each acoustic station node is integrated with an omnidirectional underwater acoustic transducer, a signal transmitting unit, a signal receiving unit and a demodulation module, an M-sequence spread spectrum modulation and matched filtering algorithm is adopted to measure the transmission time and speed of sound waves, and acoustic propagation data of the whole area of the river section are synchronously acquired; the image acquisition and AI enhancement module is used for carrying a high-definition network camera and an image preprocessing unit on a river cross section, acquiring visual image data, optimizing an image through a self-adaptive denoising algorithm, a super-resolution reconstruction technology and an edge enhancement algorithm, and outputting an enhanced image; the multi-mode data synchronization control module is internally provided with a millisecond-level time stamp synchronization unit and a data alignment engine, receives acoustic and visual data and marks time stamps for the two types of data through Beidou/GPS time service signals; The multi-source data fusion module adopts a cross-modal attention mechanism to extract acoustic and visual characteristics, constructs multi-dimensional characteristic vectors containing acoustic parameters, image characteristics and environmental parameters, performs normalization processing, and transmits the fused characteristic vectors to the AI resolving module; The AI intelligent flow resolving module is internally provided with a pre-training flow prediction model based on deep learning, inputs multi-mode feature vectors and combines river section data, and calculates instantaneous flow and average flow by fitting the mapping relation of acoustic propagation speed, image flow velocity texture and actual flow through a deep neural network algorithm; the data storage and visual display module is used for storing recently acquired original data, enhanced images and calculation results at the edge end, storing monitoring data and historical records in a cloud database for a long time, and displaying flow data, acoustic propagation curves, enhanced images and flow change trends by carrying a visual interface; And the remote control and parameter configuration module is used for realizing two-way communication with a remote control center through the communication interface of the Internet of things and receiving a remote control instruction to finish parameter updating, state inquiry and fault alarm processing.
  2. 2. The AI-image-enhanced acoustic tomography multi-modal river flow monitoring system of claim 1, further comprising an image-enhancement quality quantitative assessment module configured to construct a multi-dimensional image quality assessment model for quantifying the availability of enhanced images by a composite index expressed as ; Wherein the method comprises the steps of For the image enhancement quality composite score, For the texture sharpness weighting factor, In order to enhance the texture sharpness value of the post-image, For the texture sharpness threshold of a standard-definition image, As the contrast weighting coefficient(s), In order to enhance the contrast value of the post-image, Is the contrast threshold value of the standard image, As the weighting coefficient of the signal to noise ratio, In order to enhance the signal-to-noise ratio of the post-image, For a preset signal-to-noise threshold, Only the enhanced image with the score reaching the set standard can participate in multi-mode data fusion, and an AI self-adaptive enhanced parameter adjusting unit is mounted at the same time, so that the denoising strength, the super-resolution reconstruction iteration number and the edge enhancement coefficient are dynamically optimized according to river environment real-time monitoring data.
  3. 3. The AI-image-enhanced acoustic tomography multi-modal river flow monitoring system of claim 1, further comprising a multi-modal feature fusion weight dynamic adjustment module that adaptively assigns fusion weights based on real-time quality status of acoustic data and image data, the contribution duty ratio of the enhanced data expressed as ; Wherein the method comprises the steps of As a fusion weight of the acoustic data, For the fusion weights of the visual image data, To score the confidence level of the acoustic data, For the signal integrity of the acoustic data, For the acoustic weight adjustment factor to be used, For the characteristic integrity of the image data, For the visual weight adjustment coefficient, a dynamic weight distribution mechanism adjusts the fusion proportion according to the real-time quality of two types of data, when the quality of certain type of data is reduced, the weight proportion is automatically reduced, a characteristic quality real-time detection unit is additionally arranged, the quality grades of the two types of data are judged in real time through an acoustic signal-to-noise ratio analysis, image texture definition detection and characteristic dimension integrity verification triple mechanism, a weight adjustment buffer mechanism is established, and a weight change rate threshold is set.
  4. 4. The AI image-enhanced acoustic chromatography multi-mode river flow monitoring system according to claim 1, further comprising a dynamic correction module for geometrical parameters of river cross sections, wherein the dynamic correction module is used for calculating geometrical parameters in real time by utilizing an image measurement algorithm in combination with preset cross section reference parameters, automatically correcting cross section parameter deviation caused by water level change and riverbed siltation, synchronizing the corrected geometrical parameters to the AI intelligent flow calculation module in real time, integrating a multi-view image fusion unit, acquiring multi-view images by a high-definition camera deployed on the river cross sections, generating a cross section three-dimensional point cloud model by utilizing a stereoscopic vision algorithm, introducing a historical cross section data trend correction mechanism, and filtering correction errors caused by short-term abnormal disturbance by analyzing cross section geometrical parameter change rules accumulated for a long time.
  5. 5. The AI image-enhanced acoustic chromatography multi-mode river flow monitoring system according to claim 1, further comprising an acoustic signal anti-interference optimization module, wherein an adaptive band-pass filter is adopted to filter an environmental interference signal, an interference frequency interval is identified through signal power spectrum analysis, filtering parameters are adjusted, meanwhile, the coding length and modulation mode of an M-sequence spread spectrum signal are optimized, full-scale section width measurement is supported, an interference source intelligent identification unit is additionally arranged, an interference signal feature library is constructed based on a machine learning algorithm, the types of the interference sources are automatically identified through analysis features, a differential anti-interference strategy is formulated for different interference sources, an acoustic signal transmission power dynamic adjustment unit is carried, and the transmission power is adjusted in real time according to the river section width and water turbidity.
  6. 6. The AI image-enhanced acoustic chromatography multi-mode river flow monitoring system according to claim 1, further comprising a flow data anomaly detection and correction module, wherein an anomaly data identification model is built in, the anomaly data type is identified by analyzing the time sequence change rule of flow data and the consistency of acoustic and image characteristics, the anomaly data is corrected by adopting an interpolation algorithm based on historical data and a multi-mode characteristic backtracking verification method, the correction record is synchronously stored in a data storage module, a multi-mode cross verification unit is additionally arranged, the complementarity of acoustic data and image data is utilized for bidirectional verification, an anomaly cause classification model is simultaneously constructed, anomaly causes are automatically distinguished, an anomaly data classification processing mechanism is established, and a correction strategy is automatically selected or an alarm is triggered according to the anomaly severity.
  7. 7. The AI image-enhanced acoustic chromatography multi-mode river flow monitoring system according to claim 1, further comprising a low-power consumption intelligent control module, wherein a dual-power supply mode of solar power supply and storage battery backup is adopted, a built-in power consumption monitoring unit and an intelligent dormancy mechanism are adopted, a non-core module is controlled to enter a low-power consumption dormancy state in a data acquisition gap, a sampling frequency and a data transmission interval are dynamically adjusted according to a monitoring task priority, a power consumption dynamic allocation unit is additionally arranged, power supply resources are allocated according to module importance and real-time task requirements, a solar power prediction unit is simultaneously carried, solar power supply capacity in a future period is predicted by combining local weather forecast data and historical solar power supply records, and system operation parameters are adjusted in advance.
  8. 8. The AI-image-enhanced acoustic tomography multi-mode river flow monitoring system of claim 1, further comprising an equipment fault self-diagnosis module for monitoring the operating state of the core equipment in real time, identifying the type of equipment fault by detecting signal transmission power, image acquisition success rate, communication link stability, generating a fault diagnosis report and pushing to a control center by remote communication, and labeling the position and maintenance advice of the faulty equipment.
  9. 9. The AI image enhanced acoustic chromatography multi-mode river flow monitoring system of claim 1, further comprising a historical data mining and trend prediction module, wherein based on long-term stored monitoring data, the regular features of flow are mined through a time sequence analysis algorithm, a flow trend prediction model is built by combining weather forecast data and watershed hydrologic data, a prediction result is displayed through a visual interface, multi-source data is integrated and incorporated into the prediction model, nonlinear association relation between each factor and river flow is mined through a deep learning algorithm, meanwhile, a personalized prediction configuration unit is additionally arranged, customized prediction time scale and precision requirements are supported, a customized prediction report is generated, and a model early warning threshold is optimized by combining historical abnormal flow event data.
  10. 10. The AI image enhanced acoustic chromatography multi-mode river flow monitoring system of claim 1, further comprising a multi-platform data interaction interface module, a reserved standardized data communication interface, a data encryption transmission unit, an interface self-adaptive adaptation unit and a protocol conversion and format adaptation technology, wherein the standardized data communication interface is used for supporting the butt joint with a third-party platform, realizing the real-time sharing of data reports by adopting a unified data format, simultaneously supporting the receiving of control instructions of the third-party platform, the data encryption transmission unit is additionally arranged, the encryption processing is carried out on transmission data by adopting a national-density symmetric encryption algorithm, the interface self-adaptive adaptation unit is carried, the data protocol and the format requirements of the third-party platform are automatically identified, the seamless butt joint is realized by the protocol conversion and format adaptation technology, the two transmission modes of batch data synchronization and real-time pushing are supported, and the data interaction log is recorded.

Description

AI image enhanced acoustic chromatography multi-mode river flow monitoring system Technical Field The invention relates to the technical field of graphic data sensing and river flow monitoring, in particular to an AI image-enhanced acoustic chromatography multi-mode river flow monitoring system. Background River flow monitoring is a core foundation for water resource management, flood control, disaster reduction, ecological protection and other works, and the data accuracy and instantaneity of the river flow monitoring directly influence the scientificity of related decisions. The current mainstream monitoring means mainly comprises single point invasion type or in-situ observation, and measurement is realized through equipment such as an acoustic Doppler flow velocity profiler, a temperature and salt depth meter and the like. The method needs to directly deploy the instrument in the water body, only can cover the local flow velocity of the detection range of the probe, is difficult to reflect the average flow of the river section, and mostly adopts intermittent navigation type observation, so that the real-time monitoring requirement cannot be met. In the river course that the environment is complicated, instrument deployment is liable to be disturbed by fishery activity, shipping traffic, and implementation degree of difficulty is big, and manpower and materials input cost is high simultaneously, and long-term operation maintenance burden is heavier. The acoustic chromatography technology provides a new path for solving the problems, and the acoustic chromatography technology can acquire average temperature flow data by penetrating the whole water body area through sound waves, can realize large-scale monitoring without invading the water body, and has the potential advantages of real-time performance and high precision. However, the existing acoustic chromatographic monitoring system still has obvious technical shortboards, and is subject to disturbance of factors such as turbulence of water flow, turbidity of water quality, environmental noise and the like, so that signal transmission is unstable, and data acquisition accuracy is reduced due to the fact that the flow calculation is carried out by single acoustic data. Meanwhile, although some systems are matched with image acquisition equipment, the acquired images are easily influenced by environmental factors such as illumination change, water flow disturbance, strong reflection and the like, and the problems of more noise, low resolution, fuzzy key characteristics and the like exist, so that effective supplement cannot be provided for flow calculation. The lack of targeted image enhancement processing techniques makes it difficult for visual data to effectively complement acoustic data, and the advantages of multi-modal monitoring are not fully exploited. In addition, the existing multi-mode monitoring system generally has the problems of poor data synchronism, stiff fusion mechanism, insufficient anti-interference capability and the like. The method has the advantages that the time stamps of different mode data are inconsistent, so that fusion results are distorted, dynamic change of data quality is not considered during feature fusion, low-quality data are prone to interference with calculation results, and the method further influences monitoring accuracy in the face of complex interference sources such as shipping noise and electronic equipment interference, and lacks of differential anti-interference strategies. Meanwhile, the problems of difficult power supply in remote river areas, difficult quick investigation of equipment faults, insufficient data interaction safety and the like are solved, and the large-scale application and long-term stable operation of the conventional system are limited. Along with the continuous improvement of the requirements of water resource management on the accuracy, the instantaneity and the continuity of monitoring data, the prior art cannot meet the requirements of diversified scenes, and an integrated monitoring system integrating functions of AI image enhancement, multi-mode accurate fusion, strong anti-interference, low power consumption and the like is needed, so that the bottleneck of the prior art is broken. Disclosure of Invention The invention provides an AI image enhanced acoustic chromatography multi-mode river flow monitoring system, which aims to solve the problems in the prior art. In order to achieve the purpose, the invention adopts the following technical scheme that the AI image enhanced acoustic chromatography multi-mode river flow monitoring system comprises the following modules: The acoustic chromatography monitoring module consists of at least four acoustic station nodes arranged on two sides of a river section, each acoustic station node is integrated with an omnidirectional underwater acoustic transducer, a signal transmitting unit, a signal receiving unit and a demodulation module, an M-