Fishery resource dynamic monitoring method and system based on underwater detection equipment

Abstract

The application provides a fishery resource dynamic monitoring method and system based on an underwater detection device, relates to the technical field of dynamic monitoring analysis, and realizes fine management of water area monitoring by collecting fish group density data in real time, constructing a fish group density matrix, and combining grid element division and geographic hash coding. The gradient value of adjacent grid elements is calculated based on the density matrix to generate a gradient matrix, which can accurately identify the fish group density mutation area, mark the abnormal area, and then dynamically adjust the sonar sampling frequency of the underwater detection device to improve the monitoring accuracy of the abnormal area. Compared with the prior art, the application has been significantly optimized in terms of monitoring accuracy, abnormal detection response capability and resource utilization efficiency, especially in terms of fish group migration trend prediction and dynamic adjustment of grid division, which improves the adaptability and prediction capability of the monitoring system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dynamic monitoring, and more particularly, to a fishery resource dynamic monitoring method and system based on underwater detection equipment. BACKGROUND

[0002] Underwater fishery resource monitoring technology plays an important role in modern fishery management and ecological protection. With the development of underwater detection equipment, especially the application of sonar technology, multispectral imaging and underwater autonomous underwater vehicle (AUV), the real-time monitoring accuracy of fishery resources is continuously improved. Traditional fishery monitoring mainly relies on fishermen's experience or fixed-point sampling methods, such as trawl surveys, fixed-point hydrophone monitoring and satellite remote sensing data analysis, etc. These methods have certain limitations in real-time monitoring and fine management in large-scale waters. In recent years, fish school density analysis based on sonar detection has gradually become the mainstream. Through the analysis of echo intensity, scattering characteristics and motion trajectory, the distribution of underwater fish schools can be accurately identified. However, the existing technology still has deficiencies in the spatial division, dynamic adjustment and prediction ability of density data, especially when facing the rapid migration of fish schools and local abnormal concentration, the monitoring accuracy and data updating efficiency are difficult to meet the needs of fine management of fishery resources.

[0003] The main problems of the existing technology are as follows: First, the spatial organization of monitoring data is relatively extensive. Traditional methods often use fixed grids or regularized zoning, but do not fully consider the dynamic changes of fish school distribution, resulting in monitoring results lagging behind the actual situation. Second, the sampling frequency of existing sonar detection equipment is fixed, which is difficult to dynamically adjust to local abnormal areas, resulting in large detection error of fish school density in abnormal areas and affecting the accuracy of monitoring. In addition, in predicting the migration rules of fish schools, the existing methods are mostly based on statistical models or historical data analysis, lacking feedback and optimization of real-time monitoring data, and unable to form an efficient adaptive monitoring mechanism.

[0004] Therefore, there is a need for a fishery resource dynamic monitoring scheme based on underwater detection equipment. SUMMARY

[0005] In order to solve the above technical problems, the present application provides a fishery resource dynamic monitoring method and system based on underwater detection equipment.

[0006] According to one aspect of the present application, a fishery resource dynamic monitoring method based on underwater detection equipment is provided, which comprises: collecting fish population density in real time to form a fish population density matrix; dividing the monitoring water area into grid units, each of which is assigned a unique geographic hash code; calculating the density gradient value of adjacent grid units based on the density matrix to generate a gradient matrix and mark abnormal areas; adjusting the sonar sampling frequency of the underwater detection equipment for the abnormal areas; determining the migration rule based on historical fish population migration data, combining the data collected after adjusting the sonar sampling frequency, predicting the fish population density change trend, and dynamically correcting the fish population density matrix to adjust the division of the grid units in the abnormal areas.

[0007] Further, the division of the grid units comprises: determining the size threshold of the grid units based on the geographic range of the monitoring water area, the fish population density and the underwater environmental characteristics, and dividing the monitoring water area into grids according to the size threshold.

[0008] Further, the geographic hash code comprises the grid center coordinates, the depth level and the preset resource density threshold.

[0009] Further, the calculation of the density gradient value of adjacent grid units to generate the gradient matrix comprises: sequentially calculating the fish population density difference between adjacent grid units based on the fish population density matrix, and calculating the density gradient value of each grid unit to form a gradient data set; assigning a gradient weight to each grid unit according to the density gradient value, and generating a gradient matrix according to the spatial distribution relationship of the grid units, so that it can represent the spatial trend of the change of fish population density in the water area.

[0010] Further, when calculating the density gradient value of each grid unit, the geographic hash code is used as an index to ensure that the gradient information of each grid unit can be accurately corresponded to the spatial position.

[0011] Further, the determination of the abnormal area comprises: comparing the data in the gradient matrix with the density gradient threshold, marking the grid units that exceed the density gradient threshold, and marking the geographic hash codes thereof.

[0012] Further, the prediction of the fish population density change trend comprises: extracting the historical fish population migration data matching the current abnormal area from the historical database according to the marked geographic hash codes; using the time series data alignment method to match the sonar data of the abnormal area with the data of the corresponding historical period; using the trajectory comparison analysis method to calculate the moving direction, density increase / decrease trend of the fish population, and compare with the historical fish population migration path to predict the future diffusion or aggregation of the fish population.

[0013] Further, the dynamic correction of the fish population density matrix comprises: updating the time dimension data of the fish population density matrix according to the prediction result of the fish population density change trend.

[0014] Further, the adjustment of the division of the grid unit comprises: calculating the density change rate of the grid unit on the basis of the corrected fish population density matrix; using dynamic time warping or spatial weighted correlation analysis to evaluate whether the density change patterns of adjacent grids are similar; if the density change patterns of multiple adjacent grids are highly similar, marking as a density change related area; if the density change rate of a certain grid unit is continuously higher than the change threshold in multiple time windows, and the density distribution is uneven, triggering grid splitting; if the density change rate of a certain grid unit is continuously lower than the change threshold in multiple time windows, and the long-term change is stable, triggering grid merging; performing merging or splitting operation in the density change related area.

[0015] According to another aspect of the present application, a fishery resource dynamic monitoring system based on underwater detection equipment is provided, which comprises: a grid coding module that collects fish population density in real time to form a fish population density matrix; divides the monitored water area into grid units, and assigns a unique geographic hash code to each grid unit; a gradient analysis module that calculates the density gradient value of adjacent grid units based on the density matrix, generates a gradient matrix, and marks abnormal areas; a sonar control module that adjusts the sonar sampling frequency of the underwater detection equipment in the abnormal area; a migration prediction and adjustment module that determines the migration rule based on historical fish population migration data, combines the data collected after adjusting the sonar sampling frequency, predicts the fish population density change trend, dynamically corrects the fish population density matrix, and adjusts the division of the grid unit in the abnormal area.

[0016] Compared with the prior art, the fishery resource dynamic monitoring method and system based on underwater detection equipment provided by the present application realizes fine management of water area monitoring by collecting fish population density data in real time, constructing a fish population density matrix, and combining grid unit division and geographic hash coding. The gradient value of adjacent grid units is calculated based on the density matrix to generate a gradient matrix, which can accurately identify fish population density mutation areas and mark abnormal areas, and then dynamically adjust the sonar sampling frequency of the underwater detection equipment to improve the monitoring accuracy of the abnormal area. Compared with the prior art, the present application has been significantly optimized in terms of monitoring accuracy, abnormal detection response capability and resource utilization efficiency, especially in terms of fish population migration trend prediction and dynamic adjustment of grid division, which improves the adaptability and prediction ability of the monitoring system. In summary, the present application provides a more intelligent and efficient technical means for dynamic monitoring of fishery resources, and provides important support for precise fishery management and sustainable fishery development. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to make the technical solutions of the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor. In the drawings:

[0018] Figure 1 Flow chart of the fishery resource dynamic monitoring method based on underwater detection equipment according to the embodiment of the present application.

[0019] Figure 2 Flow chart of the fishery resource dynamic monitoring method based on underwater detection equipment according to the embodiment of the present application.

[0020] Figure 3 Flow chart of the fishery resource dynamic monitoring method based on underwater detection equipment according to the embodiment of the present application.

[0021] Figure 4 Block diagram of the fishery resource dynamic monitoring system based on underwater detection equipment according to the embodiment of the present application. DETAILED DESCRIPTION

[0022] In the following, the example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0023] As described in the above background, the main problems of the prior art are concentrated in the following aspects: first, the spatial organization of the monitoring data is relatively extensive, and the traditional method often uses fixed grid or regular partition, but does not fully consider the dynamic change of fish distribution, resulting in that the monitoring result lags behind the actual situation. Secondly, the sampling frequency of the existing sonar detection equipment is fixed, and it is difficult to dynamically adjust the local abnormal area, resulting in large error in fish density detection of abnormal area, affecting the accuracy of monitoring. In addition, in the prediction of fish migration law, the existing method is mostly based on statistical model or historical data analysis, lacking feedback and optimization of real-time monitoring data, and unable to form an efficient adaptive monitoring mechanism. Therefore, a fishery resource dynamic monitoring scheme based on underwater detection equipment is needed.

[0024] Figure 1 Flow chart of the fishery resource dynamic monitoring method based on underwater detection equipment according to the embodiment of the present application. As Figure 1As shown, in the fishery resource dynamic monitoring method based on the underwater detection equipment, S1: real-time collection of fish population density to form a fish population density matrix; the monitoring water area is divided into grid units, and each grid unit is assigned a unique geographic hash code; S2: calculation of the density gradient value of adjacent grid units based on the density matrix to generate a gradient matrix, and marking of an abnormal area; S3: adjustment of the sonar sampling frequency of the underwater detection equipment for the abnormal area; S4: determination of migration rules based on historical fish population migration data, combination of data collected after adjustment of the sonar sampling frequency, prediction of the fish population density change trend, and dynamic correction of the fish population density matrix to adjust the division of the grid units of the abnormal area.

[0025] In the embodiment of the present application, the step S1 specifically comprises: real-time collection of fish population density data of each grid by an underwater sonar array to generate a fish population density matrix, each element of the fish population density matrix corresponding to a current density value of a grid unit. The monitoring water area is divided into dynamically variable grid units, and each grid unit is assigned a unique geographic hash code, the geographic hash code including a grid center coordinate, a depth level and a preset resource density threshold. It should be noted that in fish population density monitoring, the underwater sonar array is the core sensing device, which can realize underwater target detection through the propagation and echo analysis of ultrasonic pulses. Generally, the traditional single-point sonar detection equipment is limited by the detection range and resolution, and it is difficult to realize accurate fish population density monitoring in a large range of water. The present application adopts an underwater sonar array, that is, a plurality of distributed sonar detection units are arranged in the monitoring water area, a multi-node collaborative sensing network is constructed through a network communication system, so that the detection range is expanded and the spatial resolution is improved.

[0026] In specific operation, the sonar detection units are arranged at different regions of the monitoring water area at a preset interval, and each unit can independently emit a sound wave signal and receive echo data. The process of sonar data collection includes the following key steps: (1) the sonar transmitter regularly emits ultrasonic pulses to the water area, and the ultrasonic waves produce reflected echoes when encountering fish targets during propagation in the water body; (2) the receiver obtains the echo signal, and calculates the target distance through time delay, and determines the target density by combining the echo intensity; (3) a data fusion algorithm is used between the plurality of sonar detection units to remove redundant data, and target reconstruction is performed according to the directionality characteristics of the echo signal, so that more accurate fish population density information is obtained.

[0027] In the present application, the data acquisition of the sonar array is not limited to single-point measurement, but combines the signals of multiple sonar units to realize joint calculation in the spatial dimension. For example, when the data of a single sonar detection unit in a certain area is affected by environmental noise, resulting in attenuation of the echo signal, the data information of the adjacent sonar detection unit can be used for compensation to ensure the stability of the monitoring data. This way effectively improves the robustness of the monitoring, avoids the monitoring blind area caused by environmental interference, and thus improves the accuracy and reliability of the fish density data.

[0028] Further, after obtaining the basic fish density data, the data needs to be spatially structured for subsequent analysis and dynamic adjustment. The present application adopts the fish density matrix method to divide the entire monitoring water area into multiple discrete grid cells, and the density data of each grid cell is stored in the corresponding element position of the matrix. The construction of the matrix mainly includes the following steps: dividing the monitoring water area into regular or irregular grid cells; calculating the fish density of each grid cell according to the data collected by the sonar; constructing a two-dimensional or three-dimensional matrix to store the density information of all grid cells.

[0029] Compared with the traditional point data storage method, the fish density matrix has the advantage of providing more intuitive spatial distribution information, making data processing and trend analysis more efficient. For example, through matrix row and column operations, the density gradient change in a certain area can be quickly calculated to identify the trend of fish aggregation or migration. In addition, the density matrix can serve as the basic data structure for subsequent anomaly area detection, prediction analysis and dynamic adjustment, ensuring the efficient operation of the entire monitoring system.

[0030] It can be known that the division of grid cells in the monitoring water area directly affects the accuracy of the fish density data and the monitoring efficiency. If the grid division is too large, it is difficult to accurately reflect the changes of the local fish population; if the grid division is too small, it will result in a sharp increase in data processing, reducing the real-time performance of the system. Therefore, the present application adopts a dynamic variable grid division strategy to determine a reasonable grid size threshold based on the geographical range of the monitoring water area, fish distribution data and underwater environmental characteristics, and to make adaptive adjustments.

[0031] Among them, the division of the grid cells of the present application includes: determining the size threshold of the initial grid cells based on the geographical range of the monitoring water area, fish distribution data and underwater environmental characteristics, and dividing the monitoring water area into grids according to the threshold.

[0032] For example, based on monitoring the geographical range of the water area, the depth distribution, the bottom type and the water flow rate of the water area are obtained, and combined with historical fish distribution data, the spatial range of the main activity area and the low activity area of the fish population is determined. According to the historical fish distribution data, the influence weight of different environmental characteristics (such as depth, bottom type, water flow rate) on fish population density is determined, and the difference in environmental characteristics between the high-density distribution area and the low-density distribution area is determined through weight analysis. The high-density distribution area is divided into smaller grid cells to improve the monitoring accuracy; the low-density or less environmental characteristic change area is divided into larger grid cells to reduce redundant calculation, thereby determining the size threshold of the initial grid cell.

[0033] Further, in order to ensure the uniqueness and traceability of the monitoring data, the present application adopts a geographic hash coding method to assign a unique identifier to each grid cell. This coding method not only contains the spatial position information of the grid, but also combines the depth level and the preset resource density threshold, so as to facilitate fine management.

[0034] Specifically, the composition of the geographic hash coding includes the following key fields: grid center coordinates, represented by latitude and longitude data, to ensure the accurate geographical position of each grid; depth level, considering the vertical distribution characteristics of the fish population, different coding levels are assigned to grid cells of different depths to reflect the three-dimensional structure of the water area; resource density threshold, i.e. the fish population density threshold value set according to historical monitoring data or ecological standards, used to distinguish high-density and low-density areas.

[0035] By using geographic hash coding, the present application can quickly index fish population density data and improve data processing efficiency. For example, when detecting abnormal areas, the system can directly query the corresponding grid data based on the geographic hash coding, without traversing the entire fish population density matrix, thereby significantly improving the calculation speed. In addition, geographic hash coding can also be used for data visualization, enabling monitoring personnel to more intuitively understand the spatial distribution of fish populations and facilitate further fisheries management decisions.

[0036] As can be seen, the present application combines dynamic variable grid division with geographic hash coding index to accurately manage the monitoring water area, and combines real-time data acquisition of underwater sonar arrays to construct a high-precision fish population density matrix, achieving efficient modeling of fish population distribution information. Compared with the prior art, the present application breaks through the limitations of traditional fixed grid division, enabling the grid cell to dynamically adjust according to the fish activity characteristics and environmental changes, improving the spatial adaptability and calculation efficiency of data acquisition. In addition, the introduction of geographic hash coding ensures the uniqueness of the grid cell, enabling the information of different data sources to be quickly associated, improving the accuracy of data processing and analysis.

[0037] In fish population density monitoring, observing fish population distribution through the density matrix alone may not fully reflect the dynamic changes of the fish population, especially during the aggregation, diffusion or migration of the fish population, the density changes in the local area are often more valuable. Therefore, the present application reveals the change trend of the fish population density in the water area by calculating the density gradient change value of the adjacent grid cells.

[0038] In the embodiment of the present application, the step S2 specifically comprises: based on the fish population density matrix, calculating the density gradient change value of the adjacent grid cells to generate a density gradient matrix; if the density gradient change value of a certain grid cell exceeds a preset abnormal threshold, marking the grid cell as an abnormal area and assigning a red coded identifier; the density gradient change value is calculated by weighting the absolute value of the density difference between adjacent grids, and the weight is determined by the reciprocal of the grid spacing.

[0039] The calculation of the density gradient value of the adjacent grid cells to generate the gradient matrix comprises: based on the fish population density matrix, sequentially calculating the fish population density difference between adjacent grid cells, and calculating the density gradient value of each grid cell to form a gradient data set; according to the density gradient value, assigning a gradient weight to each grid cell, and generating a gradient matrix according to the spatial distribution relationship of the grid cells, so that it can represent the spatial trend of the fish population density change in the water area. When calculating the density gradient value of each grid cell, the geographic hash code is used as an index to ensure that the gradient information of each grid cell can be accurately corresponded to the spatial position.

[0040] Specifically, the calculation process of the density gradient change value comprises the following steps: first, according to the geographic hash code of the grid cell, determine the adjacent grid of each grid cell. The adjacency relationship can be four-adjacency (up, down, left and right) or eight-adjacency (including diagonal direction), and the present application adopts eight-adjacency mode to improve the fineness of the density gradient calculation. For each grid cell, calculate the fish population density difference between it and all adjacent grids; based on the density difference, calculate the density gradient change value, in order to consider the spatial factors, the present application adopts a weighted calculation method, wherein the weight coefficient value is determined by the reciprocal of the grid cell spacing, that is, the reciprocal of the geographic distance between the current grid cell and the adjacent grid, to ensure that the influence of the nearby grid on the gradient value is greater, while the influence of the distant grid is relatively small. Through the above calculation method, it can be ensured that the density gradient matrix can not only reflect the local changes of the fish population density in the water area, but also consider the spatial relationship between the grids, thereby improving the accuracy and practicability of the data.

[0041] After completing the density gradient calculation, the calculation results need to be stored in a structured manner for subsequent analysis and anomaly detection. The present application adopts the density gradient matrix mode to map the density gradient value of each grid cell to the corresponding spatial position to form a visual gradient data set.

[0042] The generation process of the density gradient matrix is as follows: a two-dimensional or three-dimensional matrix of the same size as the fish density matrix is constructed, all element values are initialized to zero, the gradient value of each grid cell is filled into the corresponding position of the gradient matrix according to the density gradient value calculation rule, the spatial consistency of the data is ensured, and for the case of local data anomaly, the gradient matrix is optimized by using a smoothing filter method to reduce the influence of environmental noise on the data and ensure the stability of the result.

[0043] After the generation of the density gradient matrix, the abnormal area needs to be identified and marked. The present application sets a preset abnormal threshold, if the density gradient value of a grid cell exceeds the preset abnormal threshold, it is determined that the grid cell has an abnormal situation, and a red coded identifier is assigned.

[0044] The specific steps of the abnormal area marking are as follows: traversing the density gradient matrix, judging whether the density gradient value of each grid cell exceeds the abnormal threshold, for the grid cell exceeding the threshold, setting its state as abnormal and assigning a red code, and connecting the geographic hash code of the abnormal area with the map system to intuitively display the spatial position of the abnormal area, so that the management personnel can respond quickly.

[0045] The abnormality detection mechanism of the present application can effectively identify the sudden changes of fish schools in the water area, such as sudden gathering, rapid migration or disappearance of fish schools, etc. Compared with the traditional abnormality detection method based on fixed threshold, the present application combines the density gradient information, so that the abnormality identification is more accurate and can adapt to the dynamic changes of different water areas.

[0046] In the traditional underwater detection system, the sampling frequency of the sonar is usually fixed, or only depends on simple preset rules for adjustment, which is difficult to flexibly adapt to complex underwater fish school activity patterns. The present application designs a dynamic sampling strategy based on the density change characteristics, so that the sonar detection system can more intelligently adapt to different fish school dynamic situations, and improve the pertinence and effectiveness of data acquisition.

[0047] In the embodiment of the present application, the step S3 specifically comprises: adjusting the sonar sampling frequency of the underwater detection device for the abnormal area marked with the red coded identifier.

[0048] In step S2, the fish density change abnormal area is successfully identified by calculating the density gradient value between grid cells, and is marked with a red color-coded identifier. In order to further optimize the monitoring of these areas, it is first necessary to quickly locate the abnormal area by geographic hash coding. Geographic hash coding can efficiently map spatial information to a one-dimensional index, so that in large-scale water monitoring data, abnormal grid cells can be quickly screened out, avoiding the computational overhead brought by global search. After retrieving the abnormal area, the corresponding fish density gradient information needs to be extracted, including but not limited to: density change rate - used to measure the fluctuation of fish density in a short period of time; gradient direction - used to determine whether the fish is gathering or dispersing to a certain area; change pattern - used to analyze the short-term moving trend of fish, such as drift, homing behavior affected by environmental changes. These information provides the basis for subsequent adjustment of the sonar sampling frequency, so that the adjustment strategy can be dynamically optimized according to the specific situation, rather than simply adjusting the fixed rule.

[0049] Specifically, according to the abnormal area marker set generated in step S2, the specific location of the abnormal area is quickly retrieved through the geographic hash coding of the grid cell, and the corresponding fish density gradient information is extracted. The sonar sampling frequency adjustment strategy is determined as follows: when the fish density in the abnormal area increases sharply in a short period of time, the sonar sampling frequency is increased to enhance the data resolution and ensure detailed monitoring of the high-density fish area; when the fish density in the abnormal area decreases sharply or the density gradient changes dramatically, the sampling frequency is appropriately increased to capture the fish movement trajectory and determine whether the fish migration is caused by external factors (such as sudden change of water temperature, fishing activities, etc.). By increasing the sampling frequency, dense data points can be obtained in a shorter period of time, so as to analyze the sudden movement pattern of the fish and determine whether it is a normal ecological behavior or affected by external interference, and then optimize the subsequent monitoring strategy. When the fish density change trend in the abnormal area tends to be stable, the sonar sampling frequency is reduced to reduce the computational burden and optimize resource allocation. According to the adjustment strategy, the underwater exploration equipment covering the abnormal area is adjusted in its sonar sampling frequency, while the index relationship of the geographic hash coding is maintained to ensure that the correspondence between the monitoring data and the spatial position remains unchanged. After adjusting the sonar sampling frequency, the abnormal area still needs to be continuously monitored to ensure that the adjusted strategy can effectively reflect the real changes of fish density. The adjusted data maintains the same geographic hash coding index as the original data, ensuring that the spatio-temporal correspondence of the data is not destroyed. If the density gradient change tends to be stable, the default sampling frequency is restored to ensure the stability of the entire system; if the density change is still dramatic, the sampling strategy is further optimized in combination with historical data, for example, using machine learning algorithms to predict the next density change pattern, and adjusting the sampling scheme accordingly to achieve more accurate monitoring.

[0050] It should be noted that in some cases, the fish population density change trend of the abnormal area tends to be stable, that is, although the density gradient anomaly is detected initially, the fish population activity tends to be stable and no longer fluctuates greatly in the subsequent period. In this case, appropriately reducing the sonar sampling frequency can reduce the data flow, optimize the data storage and computing resources; for long-term deployed underwater monitoring equipment, energy management is a key issue, and reducing unnecessary high-frequency sampling can prolong the service life of the equipment; the sampling resources can be reasonably allocated to other high-dynamic areas instead of continuously focusing on the area that has tended to be stable, so that the resource utilization rate of the entire monitoring network is optimized.

[0051] It can be seen that, compared with the traditional fixed sampling method, the sonar sampling frequency is dynamically adjusted based on the change trend of the fish population density gradient in this step, which can more accurately adapt to the changes of the fish population activity in the water area, make the monitoring data more accurate, and effectively reduce unnecessary computing resource consumption. Through flexible adjustment strategy, not only the capture ability of the abnormal area can be improved, but also the resource allocation can be optimized, so that the entire monitoring system is more efficient and intelligent.

[0052] In the embodiment of the present application, the step S4 specifically comprises: determining the migration rule based on the historical fish migration data, combining the data collected after the sonar sampling frequency is adjusted, predicting the fish population density change trend, and dynamically correcting the fish population density matrix to adjust the division of the abnormal area grid unit.

[0053] It is known that in fish ecology research and fishery monitoring, the migration rule of fish population is a crucial research direction. Due to the periodicity and environmental dependence of fish activity, the migration patterns of different fish species in different water areas may differ significantly. Therefore, analyzing the migration rule of fish population using historical data can provide a scientific basis for predicting the current change of fish population density. Generally, the movement path of fish population is closely related to multiple environmental variables, including water temperature, dissolved oxygen content, food supply, tidal flow direction, etc. Changes in these factors will directly affect the migration pattern of fish population. For example, in cold seasons, some fish species may move to deep water areas to avoid low temperature, while in areas with sufficient food supply, fish population may tend to stay. Therefore, relying solely on real-time monitoring data may not accurately predict the future behavior of fish population, and combining historical data for analysis can provide a more comprehensive reference for trend prediction.

[0054] Traditional fish migration rule research generally relies on long-term manual monitoring and statistics. Researchers need to sample at fixed monitoring points regularly and make trend judgments based on experience. This method can provide certain migration rule reference, but has the following limitations: the update cycle of manual monitoring data is long, which cannot reflect the dynamic changes of fish population in real time; only relying on visual observation and basic environmental parameters, lacking more accurate fish population behavior data, etc. The present application adopts the method of geographic hash coding to index the fish migration data in the historical database, so that when determining the current abnormal area, the fish migration data with similar environmental characteristics can be quickly extracted from the historical database.

[0055] In order to improve the prediction accuracy, the present application adopts a method combining time series data alignment and trajectory comparison analysis to predict the trend of fish population density change, as shown in Figure 2 The method includes: extracting historical fish migration data matching the current abnormal area from the historical database according to the marked geographic hash code; using time series data alignment method to match the sonar data of the abnormal area with the data of the corresponding historical period; using trajectory comparison analysis method to calculate the moving direction, density change trend of the fish population, and compare with the historical fish migration path to predict the future diffusion or aggregation of the fish population.

[0056] For example, the historical trajectory with the most similar environmental conditions in the current time period is extracted from the historical database; the matching degree of the current trajectory and the historical trajectory is calculated by using DTW or Hausdorff distance; the historical trajectory closest to the current trajectory is selected as the prediction reference sample. After trajectory matching, it is necessary to analyze whether the motion direction of the fish population changes. Centroid drift analysis or principal direction vector calculation can be used, wherein the centroid drift analysis calculates the change of the center point of the fish population in the time series to judge the overall moving trend of the fish population; the principal direction vector calculation calculates the principal direction vector by counting the motion direction of the fish population individuals to judge whether the fish population is gathering in a certain area or diffusing outward. For example, if the centroid of the fish population continuously moves to a certain area in multiple time windows, and the principal direction vector is concentrated in that direction, it can be judged that the fish population is gathering in that area; if the direction vector of the fish population individuals is divergent and the centroid change has no obvious rule, it may indicate that the fish population is diffusing. After obtaining the trajectory comparison analysis result, the trend prediction is carried out combined with the historical data: the trajectory extrapolation method is used to compare the similar fish population trajectories in history to predict the possible moving direction in the next stage; if the historical data shows that under similar environmental conditions, the fish population usually moves in a certain direction, the current fish population may also have the same trend.

[0057] It can be seen that the step has the advantages that the geographic hash coding can realize fast retrieval and improve data matching efficiency; the trend comparison method based on historical data can more accurately predict the future density change of the fish school; and the whole process can be automatically completed by the system, reducing manual intervention and improving real-time performance.

[0058] After obtaining the trend of the change in the fish school density, the fish school density matrix needs to be dynamically corrected to provide more accurate input data for subsequent grid division adjustment. The present application corrects the fish school density matrix in the following ways: updating the time dimension data of the fish school density matrix according to the prediction result of the trend of the change in the fish school density; smoothing the fish school density of different time windows by using an adaptive weighted filtering method to reduce the error caused by short-term fluctuations; and correcting the prediction result in combination with the influence of environmental factors to ensure that the corrected density matrix conforms to the actual ecological environment characteristics.

[0059] It should be noted that the core of dynamically correcting the density matrix lies in ensuring the continuity and rationality of the data. Since the activities of the fish school are greatly affected by the environment, directly using the prediction data may lead to error accumulation, so multi-level correction must be performed. For example, in sea areas where the tide has a significant impact, the change in the fish school density may have periodic fluctuations. In this case, if necessary, a Fourier transform can be used to analyze the periodic pattern, and a periodic adjustment factor can be introduced in the correction process to improve the stability of the prediction.

[0060] On the basis of the corrected fish school density matrix, the present application further dynamically adjusts the grid cells to optimize the monitoring accuracy of the fish school density, as shown in FIG. 8. Figure 3 As shown in FIG. 8, the division of the adjusted grid cells includes: on the basis of the corrected fish school density matrix, calculating the density change rate of the grid cell, which represents the dynamic characteristics of the fish school in the area; using dynamic time warping or spatial weighted correlation analysis to evaluate whether the density change patterns of adjacent grid cells are similar; if the density change patterns of multiple adjacent grid cells are highly similar, marking them as density change related areas, which are suitable for performing merging or splitting operations; if the density change rate of a certain grid cell continuously exceeds the change threshold in multiple time windows, and the density distribution is uneven, triggering grid splitting; if the density change rate of a certain grid cell continuously falls below the change threshold in multiple time windows, and the long-term change is stable, triggering grid merging; performing merging or splitting operations in the density change related areas. According to the grid structure adjusted by splitting and merging, updating the grid index table to ensure that data storage and query can adapt to the new grid form. Recalculating the geographic hash coding of the new grid cells to ensure that data matching will not be lost or incorrectly associated due to grid adjustment.

[0061] For example, the trigger grid splitting mechanism includes: based on the corrected fish density matrix, calculating the density difference value of each grid cell at the current time and the previous time, obtaining the density change rate; through local density difference, the density change in each grid cell is locally counted, and the grid area with large density change and spatial aggregation characteristics is screened out: if the density change in a grid cell is significantly higher than that of the surrounding grid, and the density change rate of its adjacent grid is small, it is considered that the grid is an "active area" and is suitable for splitting operation; if the density change rate of a grid cell exceeds the change threshold, and the local density difference is large, the splitting operation is triggered. The trigger conditions include: if the density change rate of a grid cell is continuously high in multiple time windows, and the density distribution of the grid presents unevenness (i.e. there are significant high and low density areas in the same grid), the grid splitting is triggered. According to the density change of the surrounding grid, the change threshold of the grid is dynamically adjusted. If the change of the surrounding grid is small, a lower density change rate is allowed to trigger splitting, otherwise, the change threshold is increased to avoid over-refinement.

[0062] Further, the splitting method includes: based on the uneven density distribution, splitting in the high-density area, and dividing the area with higher density into new sub-grids; if the density change rate in the original grid is high and the distribution is relatively concentrated, the original grid can be divided into 2 to 4 sub-grids, using equal proportion division or division based on density hot center. The size of the split grid cell depends on the size of the density difference in the original grid. The area with large density difference is divided into smaller sub-grids, and the area with stable density is kept as a larger grid cell to ensure efficiency. The updated density data of the new grid cell after splitting is directly merged into the global density matrix to ensure smooth connection of new data and not to increase additional calculation pressure.

[0063] For example, the optimization goal of grid merging is to optimize the monitoring efficiency in the area where the fish population changes little, reduce the computational burden, and ensure the accuracy of the fish population density distribution. The grid merging mechanism includes: calculating the variance of the density change of adjacent grid cells, if the variance of multiple grids in multiple time windows in the past is lower than the set threshold, it is considered that the density change is stable; comparing the average of the density gradient of the adjacent grid boundary, if the boundary gradient is close to zero, it is considered that the boundary has no significant separation; only the area between adjacent grids with no significant density gradient and stable density change is selected as the candidate merging unit. Adaptive shape (such as polygon) is used for grid merging instead of fixed rectangular merging, which makes the shape of the merged grid more consistent with the fish population distribution pattern and reduces unnecessary calculation area. If the density change pattern of adjacent grid cells is similar and there is no significant change in the future time period, rectangular or approximately rectangular merging is directly performed to reduce the computational complexity. The grid cell weight center reconstruction method is used to adjust the boundary of the new grid after merging to make it as consistent as possible with the fish population density distribution characteristics instead of simply superimposing the original boundary. The density balance before and after merging is compared, and if it remains stable after merging, the merging operation is finally determined.

[0064] It should be noted that in the high-density change area, if the minimum grid size is not set, each split will further refine the grid, and eventually the size of a single grid cell may tend to zero. In the low change area, if the maximum grid size is not set, multiple grids may continue to merge, eventually resulting in a single grid covering a large area and losing the ability to perceive local density changes. Therefore, a minimum grid size threshold is set, and when the edge length of a certain area grid is close to or less than the minimum grid size threshold after subdivision, further splitting is stopped; a maximum grid merging size is set, and when the edge length of a certain merged grid exceeds the maximum grid merging size, it is no longer merged.

[0065] As can be seen from the above, by combining historical fish migration data and sonar sampling data, a method based on trajectory comparison and dynamic grid optimization is proposed, which realizes accurate prediction and dynamic correction of fish population density change trend, and can efficiently and accurately predict fish population distribution changes in complex water environments.

[0066] In summary, the fishery resource dynamic monitoring method based on the underwater detection equipment according to the embodiment of the present application is illustrated, the fish school density data is collected in real time, the fish school density matrix is constructed, and the grid unit division and the geographic hash coding are combined to realize the fine management of the water area monitoring. The gradient values of the adjacent grid units are calculated based on the density matrix to generate the gradient matrix, the fish school density mutation area can be accurately identified, the abnormal area is marked, and then the sonar sampling frequency of the underwater detection equipment is dynamically adjusted to improve the monitoring accuracy of the abnormal area. Compared with the prior art, the present application has been significantly optimized in terms of monitoring accuracy, abnormal detection response capability and resource utilization efficiency, especially in terms of fish school migration trend prediction and grid division dynamic adjustment, the adaptability and prediction capability of the monitoring system are improved. In summary, the present application provides a more intelligent and efficient technical means for dynamic monitoring of fishery resources, and provides important support for accurate fishery management and sustainable fishery development.

[0067] Figure 4 The block diagram of the fishery resource dynamic monitoring system based on the underwater detection equipment according to the embodiment of the present application is shown. As shown in Figure 4 The fishery resource dynamic monitoring system based on the underwater detection equipment includes: a grid coding module 100, which collects fish school density in real time to form a fish school density matrix; the monitoring water area is divided into grid units, and each grid unit is assigned a unique geographic hash code; a gradient analysis module 200, which calculates the density gradient values of adjacent grid units based on the density matrix to generate a gradient matrix, and marks the abnormal area; a sonar control module 300, which adjusts the sonar sampling frequency of the underwater detection equipment in the abnormal area; a migration prediction and adjustment module 400, which determines the migration rule based on historical fish school migration data, combines the data collected after adjusting the sonar sampling frequency to predict the fish school density change trend, and dynamically corrects the fish school density matrix to adjust the division of the abnormal area grid unit.

[0068] Here, those skilled in the art can understand that the specific operation of each module in the above fishery resource dynamic monitoring system based on the underwater detection equipment has been described in detail above with reference to the description of the fishery resource dynamic monitoring method based on the underwater detection equipment Figures 1 to 3 Therefore, the repeated description will be omitted.

[0069] To sum up, the fishery resource dynamic monitoring system based on the underwater detection equipment is illustrated based on the embodiment of the application, the fish school density data is collected in real time, the fish school density matrix is constructed, and the fine management of water area monitoring is realized by combining the grid element division and the geographic hash coding. The gradient value of the adjacent grid element is calculated based on the density matrix, the gradient matrix is generated, the fish school density mutation area can be accurately identified, the abnormal area is marked, and then the sonar sampling frequency of the underwater detection equipment is dynamically adjusted, and the monitoring accuracy of the abnormal area is improved. Compared with the prior art, the monitoring accuracy, the abnormal detection response capability and the resource utilization efficiency are all significantly optimized, especially in the fish school migration trend prediction and the grid division dynamic adjustment, the adaptability and the prediction capability of the monitoring system are improved. In summary, the application provides a more intelligent and efficient technical means for the dynamic monitoring of fishery resources, and provides important support for accurate fishery management and sustainable fishery development.

Claims

1. A method for dynamic monitoring of fishery resources based on underwater detection equipment, characterized in that, include: Real-time fish density is collected to form a fish density matrix; the monitored water area is divided into grid units, and each grid unit is assigned a unique geohash code; Based on the density matrix, the density gradient values ​​of adjacent grid cells are calculated, a gradient matrix is ​​generated, and abnormal regions are marked. Adjust the sonar sampling frequency of the underwater detection equipment in the abnormal area; Based on historical fish migration data, migration patterns are determined. Combined with data collected after adjusting the sonar sampling frequency, the trend of fish density change is predicted, and the fish density matrix is ​​dynamically corrected to adjust the division of grid cells in abnormal areas.

2. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 1, characterized in that, The division of the grid cells includes: determining the size threshold of the grid cells based on the geographical range, fish density and underwater environmental characteristics of the monitored water area, and dividing the monitored water area into grids according to the size threshold.

3. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 2, characterized in that, The geohash code includes the grid center coordinates, depth level, and preset resource density threshold.

4. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 1, characterized in that, The calculation of density gradient values ​​between adjacent grid cells and the generation of a gradient matrix includes: based on the fish density matrix, sequentially calculating the fish density difference between adjacent grid cells, and calculating the density gradient value of each grid cell to form a gradient dataset; according to the density gradient value, assigning gradient weights to each grid cell, and generating a gradient matrix according to the spatial distribution relationship of the grid cells, so that it can characterize the spatial trend of fish density changes in the water area.

5. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 4, characterized in that, When calculating the density gradient value of each grid cell, geohashing is used as an index to ensure that the gradient information of each grid cell can accurately correspond to its spatial location.

6. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 5, characterized in that, The determination of the abnormal region includes: comparing the data in the gradient matrix with the density gradient threshold, marking the grid cells that exceed the density gradient threshold, and marking their geohash codes.

7. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 1, characterized in that, The prediction of fish density change trends includes: extracting historical fish migration data matching the current abnormal area from the historical database based on the marked geohash code; matching the sonar data of the abnormal area with the data of the corresponding historical period using a time series data alignment method; and calculating the movement direction and density increase / decrease trend of the fish population using a trajectory comparison analysis method, and comparing it with the historical fish migration path to infer the future spread or aggregation of the fish population.

8. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 7, characterized in that, The dynamic correction of the fish density matrix includes updating the time dimension data of the fish density matrix based on the prediction results of the fish density change trend.

9. The method for dynamic monitoring of fishery resources based on underwater detection equipment according to claim 1, characterized in that, The adjustment of the grid cell division in the abnormal region includes: calculating the density change rate of the grid cells based on the corrected fish density matrix; using dynamic time warping or spatial weighted correlation analysis to assess whether the density change patterns between adjacent grid cells are similar; if the density change patterns of multiple adjacent grid cells are highly similar, they are marked as density change related regions; if the density change rate of a certain grid cell is consistently higher than the change threshold within multiple time windows and the density distribution is uneven, grid splitting is triggered; if the density change rate of a certain grid cell is consistently lower than the change threshold within multiple time windows and the long-term change is stable, grid merging is triggered; merging or splitting operations are performed in the density change related regions.

10. A dynamic monitoring system for fishery resources based on underwater detection equipment, and a dynamic monitoring method for fishery resources based on underwater detection equipment as described in any one of claims 1 to 9, characterized in that: Also includes: The grid coding module collects fish density data in real time and forms a fish density matrix; it divides the monitored water area into grid units, and assigns a unique geohash code to each grid unit. The gradient analysis module calculates the density gradient values ​​of adjacent grid cells based on the density matrix, generates a gradient matrix, and marks abnormal regions. The sonar control module adjusts the sonar sampling frequency of the underwater detection equipment in the abnormal area; The migration prediction and adjustment module determines migration patterns based on historical fish migration data, combines data collected after adjusting the sonar sampling frequency, predicts fish density change trends, dynamically corrects the fish density matrix, and adjusts the division of grid cells in abnormal areas.

Images