A method for optimizing suspended sediment monitoring accuracy

Abstract

The present invention discloses a method for optimizing the monitoring accuracy of suspended sediment, which belongs to the field of digital measurement. The method includes configuring a number of hydrological instruments to form a monitoring line and constructing a coordinate system; obtaining a number of first features respectively to make a trend graph corresponding to the device code; inputting a second feature into a server and associating the trend graph; taking any trend graph as the first data, mapping the second feature according to the error coefficient on the trend graph corresponding to the first data to form a coverage area, and the server determines whether the trend graph and the coverage area are similar. And obtaining the historical data of the monitored water area, the server determines whether the fluctuation feature is valid through the deviation average value, and the server also inputs the fluctuation feature and the first feature in the coverage area into the training model, and the training model outputs the characteristic value of the water level layer after weighted operation, so as to improve the problem that the automatic monitoring equipment is limited by discontinuity and volatility when sampling sediment at high frequency, resulting in large errors in the obtained results.

Description

Technical Field

[0001] The invention relates to river sediment monitoring, and in particular to a method for optimizing suspended sediment monitoring accuracy. Background Art

[0002] In the field of river sediment monitoring, accurate measurement of suspended sediment is of great significance for river management, flood prevention and disaster reduction, and rational use of water resources. Traditional suspended sediment monitoring methods mainly rely on manual sampling and laboratory analysis. Although the data obtained by traditional methods can provide a relatively accurate basis, they are limited by the sampling frequency and timeliness, making it difficult to obtain the changes in sediment flow during the entire monitoring process.

[0003] Based on this, various automated monitoring equipment has been developed in the existing river monitoring process. Many automated monitoring equipment rely on optical sensors (such as photoelectric sand meters) to measure suspended sediment, combined with optical sensors' sensitivity to particle size, shape and refractive index to collect data, so as to better obtain sediment characteristics. By inputting the parameters of sediment characteristics into the cloud system, the cloud system is used to predict risks.

[0004] However, the existing automated monitoring equipment mainly relies on sensors. In long-term use, the sensors may have zero-point drift or environmental interference (interference from substances in the water), which makes the measurement accuracy deviation random, resulting in the parameters under different river conditions being difficult to reflect the real dynamic changes of sediment. This discontinuity and volatility make the measurement results of different devices or the same device in different time periods have large errors, making long-term trend analysis and model prediction difficult to promote, and seriously affecting the use value of hydrological data. Therefore, it is worth studying how to optimize the timeliness and accuracy of existing equipment in the process of river sediment monitoring and reduce the errors of measurement results of different devices or the same device in different time periods. Summary of the invention

[0005] The purpose of the present invention is to provide a method for optimizing the monitoring accuracy of suspended sediment, in order to improve the problem that when automated monitoring equipment performs sediment sampling at high frequencies, the obtained results have large errors due to discontinuity and volatility.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solution: a method for optimizing the suspended sediment monitoring accuracy, the method comprising the following steps:

[0007] S100, deploying a plurality of hydrographic instruments to the monitoring water area, so that the hydrographic instruments form sampling points in the monitoring water area, and a monitoring line is formed in the vertical height by the plurality of sampling points; setting more than two monitoring lines and dividing them into light wave lines and sampling lines according to their functions.

[0008] S200, the hydrometer exchanges signals with the server, constructs a coordinate system through the location information of the hydrometer, associates and marks the hydrometers of the light wave line and the sampling line, and associates the device codes of the hydrometers placed in the same water level layer with each other.

[0009] S300, configure the monitoring period, the hydrologist on the optical wave line sends several first features obtained by the optical wave element in time-sharing during the monitoring period, and the server makes a trend chart of the first features corresponding to the device code according to the timestamp; obtain the second feature through the hydrologist on the sampling line, input the second feature to the server and associate it with the trend chart; wherein the first feature includes suspended sediment concentration data with a timestamp and a device code; the second feature includes suspended sediment concentration data obtained by weighing.

[0010] S400, the server uses any trend graph as the first data, obtains the second feature of the associated trend graph as the second data, and presets the error coefficient, maps the second data according to the error coefficient on the trend graph corresponding to the first data to form a coverage area, and the server determines whether the trend graph and the coverage area are similar. If the trend graph is completely within the coverage area, it is considered to be similar, and the second data is used as the characteristic value of the water level layer for recording and monitoring. If any area of ​​the trend graph is outside the coverage area, it is considered to be not similar, and the server extracts the first feature outside the coverage area on the trend graph as a fluctuation feature; and executes S500.

[0011] S500, obtain historical data of the monitored water area, determine the average value of the suspended sediment concentration of the water level layer in R months in the historical data and use it as a reference value; wherein the value of R is any positive integer between 1 and 12, and R month represents a specific month; calculate the deviation average value through the difference between each fluctuation feature and the reference value, and the server determines whether the fluctuation feature is valid through the deviation average value.

[0012] If the fluctuation feature is judged to be invalid, the server will alarm and reconfigure the monitoring period.

[0013] Among them, if the fluctuation feature is judged to be valid, the server inputs the fluctuation feature and the first feature in the coverage area into the training model, and the training model outputs the characteristic value of the water level layer after weighted calculation.

[0014] Preferably, in the above S500, the specific steps for the server to determine whether the fluctuation feature is valid through the deviation average value are: S501, preset a blinding parameter and a safety threshold, the above server identifies whether the fluctuation feature reaches the blinding parameter, and marks the fluctuation feature greater than or equal to the blinding parameter as abnormal; and marks the fluctuation feature lower than the blinding parameter as normal. S502, the server calculates the difference between the fluctuation feature marked as normal and the reference value to obtain a difference set. S503, obtain the mean of the difference set to obtain the deviation average value, and compare the deviation average value with the safety threshold; if the deviation average value exceeds the safety threshold, the fluctuation feature is judged to be invalid; if the deviation average value is within the safety threshold range, the fluctuation feature is judged to be valid.

[0015] As a preference, the above-mentioned training model is used to verify the characteristic value of the suspended sediment concentration of any water level layer, and the characteristic value of the suspended sediment concentration of the water level layer is The calculation formula is:

[0016] ;

[0017] In the formula, is the characteristic value of the suspended sediment concentration in the water level layer, is the weight coefficient of the fluctuation characteristics in the water level layer, is the weight coefficient of the first feature in the coverage area in the water level layer, is the mean of the effective fluctuation characteristics in the water level layer, The mean value of the first feature in the coverage area at the water level layer; To adjust the bias parameter of the model's output matching degree, is the error correction parameter of the reference value in the water level layer.

[0018] A further technical solution is that calculating the mean of the first feature in the water level layer includes the following steps: determining the water level layer to be calculated and obtaining the corresponding trend graph and coverage area; extracting all the first features of the corresponding coverage area on the trend graph, obtaining the suspended sediment concentration data corresponding to each first feature, and summarizing them to obtain a first data set; calculating the mean of the first feature in the water level layer in its coverage area through the first data set , the formula is as follows:

[0019] ;

[0020] In the formula, is the total value of all suspended sediment concentration data in the first data set; is the number of the first feature in the first data set; is the qth suspended sediment concentration data in the first data set.

[0021] A further technical solution is that calculating the mean of the effective fluctuation characteristics in the water level layer includes the following steps: obtaining the fluctuation characteristics placed outside the coverage area on the trend graph, determining the number of effective fluctuation characteristics, extracting the suspended sediment concentration data corresponding to the effective fluctuation characteristics, and summarizing them to obtain a second data set; calculating the mean of the effective fluctuation characteristics in the water level layer through the second data set , the formula is as follows:

[0022] ;

[0023] In the formula, is the total value of the suspended sediment concentration data of the fluctuation characteristics in the second data set, N is the number of valid fluctuation characteristics, is the point value corresponding to the i-th valid fluctuation feature in the second data set.

[0024] Preferably, the number of the monitoring lines is greater than three and more than two optical wave lines are set, and the hydrometers on each optical wave line are distributed the same in the longitudinal direction, so that there are at least two or more hydrometers on the optical wave line in the same water level layer. The server marks the hydrometers placed on the optical wave line as the master device and the hydrometers placed on the sampling line as the slave device. When the server obtains the location information of the hydrometer, it marks the master device placed in the same water level layer.

[0025] A further technical solution is that the above-mentioned server sets the same water level layer as a working area, and when executing S300, the server averages and fuses the first features sent by the main devices in the working area at the same time; forms third data, and regards the third data as the first data to execute S400.

[0026] A further technical solution is that the server sets a number of boundary areas in the same water level layer, and regards each boundary area as a working area. When the server obtains the location information of the hydrological instrument, it obtains the number of master devices in the working area, and the server averages and fuses the first features sent by the master devices in the working area at the same time to form fourth data, and regards the fourth data as the first data to execute S400.

[0027] Preferably, the method for optimizing the suspended sediment monitoring accuracy further includes S600, launching the sampling line to the monitored water area and corresponding to the light wave line, so as to re-associate the light wave line with the hydrological instrument of the sampling line; and repeating steps S300 to S500.

[0028] A further technical solution is that when the above-mentioned fluctuation characteristics are judged to be valid, the server records the timestamp and device code of the fluctuation characteristics marked as abnormal, and the server configures the number of verification times. When the number of times the server records the same device code is greater than the number of verification times, the server re-acquires the historical data of the monitored waters, adjusts the parameters of R and adjusts the reference value.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The hydrological instrument layout sampling point method adopted in the present invention allows the sampling points to cover multiple water level layers. Through the linkage between the light wave line and the sampling line, when the light wave line is disturbed by the environment and the equipment drift causes measurement errors, the sampling points are used for identification. The first feature is used to form a trend graph to display the sediment changes, and the second feature is used to determine the upper and lower limits of the deviation, so that the light wave line can observe the sediment data changes and obtain data within a reasonable range. The server determines whether the trend graph and the coverage area are similar, and uses the fluctuation characteristics and historical data to identify abnormal situations. When there is an obvious abnormal environmental impact, the abnormal data can be excluded, so as to obtain more accurate data to adjust the model and finally output parameters with a smaller error risk.

[0031] The training model of the present invention can combine historical data and current monitoring data of each monitoring point according to the weight adjustment result of the water level layer, and recalculate the weighted average value of sediment concentration through the weight coefficient of each monitoring point. It can ensure that the accuracy and reliability of monitoring data are improved under different water levels and different environmental conditions. The training model can also dynamically adjust the characteristic value according to the real-time monitoring data to adapt to environmental changes, thereby realizing real-time and accurate monitoring of sediment concentration changes. In order to improve the accuracy of suspended sediment monitoring, it provides more scientific data support for research and decision-making in related fields.

[0032] The present invention realizes the construction of river sediment concentration monitoring areas and forms multiple working areas through multiple monitoring lines. By collecting and monitoring the dynamic changes of river sediment in the working area. The weighted algorithm can also be adjusted using the training model to calculate the working area. By monitoring different working areas separately, it is beneficial to provide research data for the movement law of river sediment in the later stage. The data obtained objectively helps to understand the transport, deposition and erosion process of river sediment in the later stage more deeply, and provides more comprehensive data support for scientific decision-making in river management and water resources management. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the process of the present invention.

[0034] Figure 2 Schematic diagram of the coverage area of ​​the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0036] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. in a certain working state. If the specific posture changes, the directional indication will also change accordingly. In the present invention, unless otherwise clearly specified and limited, the term "connection" and the like should be understood in a broad sense. For example, "connection" can be an electrical signal connection or a signal connection; it can also be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0037] If there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "and / or" appearing in the full text includes three parallel schemes. Taking "A and / or B" as an example, it includes scheme A, or scheme B, or a scheme that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0038] Considering the modern hydrological monitoring field, it is crucial to collect and analyze data accurately and efficiently. At present, hydrological instruments generally integrate advanced sensors and remote communication equipment. For example, by putting the existing integrated lead fish into the water area as a hydrological instrument, the concentration of suspended sediments in the water area can be monitored and analyzed in real time. Its hydrological instrument integrates multiple functional components. The hydrological instrument is usually equipped with flow measurement equipment (acoustic Doppler current meter ADCP), photoelectric sand meter, infrared imaging equipment, sampler and underwater camera.

[0039] Among them, the photoelectric sand meter mainly uses optical back projection imaging technology to determine the position, distribution, particle size, etc. of suspended matter in the water body. Through the algorithm, a plane projection simulation pixel is formed to obtain the three-dimensional distribution and volume proportion of suspended matter in the irradiated water body, thereby obtaining the weight of sediment in the measured water body.

[0040] Among them, the acoustic Doppler current meter ADCP uses the Doppler effect produced by emitting sound waves and receiving reflected waves, and uses the frequency changes of the reflected light waves to analyze and estimate the concentration and distribution of suspended sediment.

[0041] Among them, infrared imaging equipment mainly uses infrared imaging technology to observe the optical properties of sediment, such as color and distribution density. Its underwater camera directly observes the behavior and distribution of sediment underwater, and its sampler is mainly used to obtain sediment samples to facilitate direct collection of sediment samples. Limited by the sampling of hydrographic instruments, its data deviation is large. The reason is that hydrographic instruments are affected by many factors in actual operation, such as turbidity of water bodies, flow rate changes, underwater obstacles, and the accuracy limitations of the equipment itself. These factors may cause deviations in the measurement results of sensors such as acoustic Doppler current meters ADCP and photoelectric sand meters.

[0042] Based on this, in order to improve the monitoring accuracy, refer to Figure 1 As shown, one embodiment of the present invention is a method for optimizing the monitoring accuracy of suspended sediment, the method comprising the following steps: S100, configuring a plurality of hydrographic instruments to the monitoring waters, so that the hydrographic instruments form sampling points in the monitoring waters, and a monitoring line is formed by the plurality of sampling points in the longitudinal height; more than two monitoring lines are set and divided into light wave lines and sampling lines according to their functions. By configuring a plurality of hydrographic instruments that interact with server signals, the hydrographic instruments are dropped into the monitoring waters through cables, the above-mentioned hydrographic instruments form sampling points in the monitoring waters, and a monitoring line is formed by the plurality of sampling points; wherein more than two monitoring lines are set and divided into light wave lines and sampling lines according to their functions.

[0043] For reference, by reasonably deploying a certain number of hydrological instruments as hydrological monitoring equipment in the monitored waters, multiple hydrological instruments are placed at predetermined positions through cables to form several sampling points. Limited by the influence of the cables, the sampling points are arranged in the longitudinal height to form a series of monitoring lines. In order to make more effective use of these monitoring lines, the server divides the hydrological instruments on the cables into optical wave lines and sampling lines. The optical wave lines are mainly responsible for using optical wave technology to quickly obtain basic information in the waters; while the sampling lines are responsible for collecting samples of suspended sediments for subsequent collaborative analysis and correction of the parameters of the optical wave lines.

[0044] For reference, existing hydrological instruments can collect physical properties such as particle size distribution, shape, density, etc. of sediment according to actual conditions, so as to facilitate the subsequent analysis of relevant data and determine the performance of sediment in light waves and infrared imaging. When abnormal data is found, the performance rating of the particle distribution density and color characteristics under infrared imaging can be used to reflect the optical properties of the sediment. The distribution and density of sediment in the water can also be reflected by the particle size, shape, light wave reflection intensity and sound speed changes. If necessary, the impact of other suspended matter can also be judged through the camera.

[0045] S200, the hydrometer exchanges signals with the server, constructs a coordinate system through the location information of the hydrometer, associates and marks the hydrometers of the light wave line and the sampling line, and associates the device codes of the hydrometers placed in the same water level layer with each other.

[0046] By establishing a stable signal interaction between the hydrometer and the server, the server can use the location information of the hydrometer to build a coordinate system. The coordinate system helps the server track the exact location of each hydrometer on the one hand, and on the other hand, it can associate and mark the hydrometers on the light wave line and the sampling line according to their locations, especially for hydrometers located in the same water level layer, and the device codes will be associated with each other.

[0047] On the one hand, the data collected by various hydrological instruments in the same hydrological layer usually have similar characteristics. By using association tags to associate the corresponding hydrological instruments with each other, the data collected by the hydrological instruments can be integrated to form a complete data set. This helps to improve the efficiency of data processing and reduce errors and omissions in the data integration process.

[0048] On the other hand, by associating the equipment codes of the same hydrological layer, the data correspondence between different hydrological instruments between devices can be established. This helps to accurately correspond the data to the corresponding equipment in the subsequent data analysis, thereby ensuring the accuracy and reliability of the data. And the use of equipment codes to associate the same water level layer can also facilitate the traceability and verification of data, and timely discover and correct data errors.

[0049] S300, configure the monitoring period, the hydrological instrument of the light wave line sends several first features obtained by the light wave element in time-sharing during the monitoring period, and the server makes a trend chart of the corresponding device code according to the timestamp of the first feature; obtain the second feature through the hydrological instrument on the sampling line, input the second feature to the server and associate it with the trend chart. Among them, the first feature includes the suspended sediment concentration data with the timestamp and the device code; the second feature includes the suspended sediment concentration data obtained by weighing.

[0050] Among them, the monitoring period is usually a time period, and the time unit is usually calculated in hours, such as 24 hours, 48 ​​hours, etc. During the monitoring period, multiple hydrographic instruments on the optical wave line will send several first features obtained by the optical wave element in a time-sharing manner. Since the first feature includes suspended sediment concentration data with a timestamp and a device code, the first feature is received by the server in real time and organized into a trend chart corresponding to the device code. At the same time, the hydrographic instruments on the sampling line will also regularly collect the second feature. Usually, the second feature is collected at the end of the monitoring period. The second feature can be the suspended sediment concentration data obtained by weighing. After the second feature is entered into the server, the server will find the associated hydrographic instrument through the device code corresponding to the second feature, and obtain the trend chart of the associated hydrographic instrument.

[0051] It should be noted that hydrological equipment is usually in self-operation, and it is difficult to monitor abnormal data. However, in actual work, humans will also regularly observe the water conditions through cameras. If any abnormal or erroneous data is observed, it is necessary to immediately mark it with the equipment code and notify relevant personnel for processing.

[0052] S400, the server uses any trend graph as the first data, obtains the second feature of the associated trend graph as the second data, and presets the error coefficient, maps the second data according to the error coefficient on the trend graph corresponding to the first data to form a coverage area, and the server determines whether the trend graph and the coverage area are similar. The server uses any trend graph as the first data set, and obtains the second feature data associated with it as the second data set. Then, according to the preset error coefficient, the server maps the second data set according to the error coefficient to the trend graph corresponding to the first data set and forms a coverage area.

[0053] The value of the error coefficient can be determined by referring to historical data and expert experience. The second data is usually a specific point value. By combining the second data with the error coefficient, an upper and lower limit range value can be formed, and the range value is used as the coverage area in the trend chart. For details, see Figure 2 As shown, the first data mainly includes multiple first features in the trend graph, and the second feature of the associated trend graph is usually a line connecting the point value data collected by the hydrological instrument on the optical wave line according to the timestamp. When the point value corresponding to the second feature is mapped to the trend graph, it is usually a straight line. The error coefficient is a parameter used to adjust the coverage of the second data set on the first data set. When the second feature is mapped to the trend graph according to the error coefficient, straight lines representing the upper and lower limits are formed above and below the straight line position of the second feature, respectively. The area between the upper and lower limit straight lines is the coverage area, and the coverage area corresponds to the monitoring period of the trend graph in the time dimension.

[0054] Among them, the server performs similarity judgment through trend graph and coverage area.

[0055] Among them, if the trend graph is completely within the coverage area, it is regarded as approximate, and the second data is used as the characteristic value of the water level layer for recording and monitoring; the two are approximate, indicating that the second data set is consistent with the first data set in trend, which helps to verify the accuracy and reliability of the data. Since the fluctuation range is within a reasonable stage, the second data can be directly used as the characteristic value of the water level layer.

[0056] Among them, if any area of ​​the trend graph is outside the coverage area, it is considered to be not approximate, and the server extracts the first feature outside the coverage area on the trend graph as the fluctuation feature; and executes S500; if any area is outside the coverage area, it indicates that there is a significant difference between the trend graph and the coverage area, and it is necessary to obtain the first feature of the difference, and then execute the analysis and disposal measures of S500.

[0057] S500, obtain historical data of the monitored water area, determine the mean value of the suspended sediment concentration of the water level layer in R months in the historical data and use it as a reference value; wherein R is any positive integer between 1 and 12, and R months represent specific months; calculate the deviation average value by the difference between each fluctuation feature and the reference value, and the server determines whether the fluctuation feature is valid through the deviation average value. If the fluctuation feature is judged to be invalid, the server will alarm and reconfigure the monitoring cycle. If the fluctuation feature is judged to be valid, the server will input the fluctuation feature and the first feature in the coverage area into the training model, and the training model will output the characteristic value of the water level layer after weighted calculation.

[0058] For reference, historical data provides a baseline level of sediment concentration in waters under normal conditions. This baseline can be used as a basis for comparison; the mean value of suspended sediment concentration in month R is selected mainly because the long-term data is expected to reveal seasonal changes, interannual changes, and possible long-term trends in sediment concentration; the deviation average is calculated by the difference between each fluctuation feature and the reference value. The deviation average can measure the difference between the current monitoring data and the historical reference value, and then the server uses the deviation average to determine whether the fluctuation feature is valid.

[0059] Among them, the suspended sediment concentration in month R refers to the data of month R of the corresponding season in recent years, where R is a positive integer from 1 to 12. The value of R usually corresponds to the current season. In principle, R can be selected from the current month or the nearby month to ensure that the referenced historical data can correspond to the seasonal environment currently monitored.

[0060] On the one hand, the deviation average exceeds the preset safety threshold, which indicates that most of the fluctuation characteristics have exceeded the safety limit, indicating that the difference between the fluctuation characteristics and the coverage area exceeds the normal change. Therefore, all fluctuation characteristics should be regarded as abnormal.

[0061] On the other hand, if the deviation average value is within the preset safety threshold, it means that the fluctuation characteristics have not exceeded the safety limit and most of the fluctuation characteristics are valid. Therefore, the server needs to input the fluctuation characteristics and the first feature in the coverage area into the training model, and the training model will output the characteristic value of the water level layer after weighted calculation.

[0062] Based on the above embodiment, another embodiment of the present invention is that in order to better judge the validity of the fluctuation feature by the deviation average value, in the above S500, the specific steps of the server judging whether the fluctuation feature is valid by the deviation average value are:

[0063] S501, preset blinding parameters and safety thresholds. The above server identifies whether the fluctuation characteristics have reached the blinding parameters, and marks the fluctuation characteristics that are greater than or equal to the blinding parameters as abnormal; and marks the fluctuation characteristics that are lower than the blinding parameters as normal. Among them, the blinding parameter is a parameter for a single fluctuation characteristic that obviously exceeds the normal data, and the safety parameter is a floating reference value during the sediment change process. Usually, the blinding parameter and the safety parameter are set based on historical monitoring data and expert experience. The blinding parameter is mainly used to identify fluctuation characteristics that obviously exceed the constant, and the blinding parameter is set to filter out obvious abnormal data caused by measurement errors or temporary environmental factors. The safety parameter is mainly used to indicate the normal floating reference range of sediment in the same water level layer.

[0064] S502, the server calculates the difference between the fluctuation feature marked as normal and the reference value to obtain a difference set; the server excludes the fluctuation feature marked as abnormal and only uses the fluctuation feature marked as normal for calculation, thereby reducing the impact of abnormal fluctuation features on accuracy. The change trend of the fluctuation feature is obtained by calculating the difference between each fluctuation feature and the reference value; and all differences are aggregated into a difference set.

[0065] S503, obtaining the average value of the difference set to obtain the deviation average value, and comparing the deviation average value with the safety threshold; if the deviation average value exceeds the safety threshold, the fluctuation feature is judged to be invalid; if the deviation average value is within the safety threshold range, the fluctuation feature is judged to be valid.

[0066] By calculating the average value of the difference set to obtain the average deviation, the average difference trend of the normal fluctuation characteristics can be determined. By comparing the average deviation of the difference trend with the safety threshold, if the average deviation exceeds the safety threshold, it means that the difference trend has exceeded the normal fluctuation and there may be abnormal conditions, so it is necessary to pause and check the risks. Therefore, the fluctuation characteristics are judged to be invalid and are not used as relevant data for subsequent processing. On the contrary, if the average deviation is within the safety threshold range, it means that the difference trend is within the normal fluctuation range, and the fluctuation characteristics are judged to be valid and can be used for subsequent processing.

[0067] Based on the above embodiment, another embodiment of the present invention is that the above training model is used to verify the characteristic value of the suspended sediment concentration of any water level layer, and the characteristic value of the suspended sediment concentration of the water level layer is The calculation formula is:

[0068] ;

[0069] In the formula, is the characteristic value of the suspended sediment concentration in the water level layer, is the weight coefficient of the fluctuation characteristics in the water level layer, is the weight coefficient of the first feature in the coverage area in the water level layer, is the mean of the effective fluctuation characteristics in the water level layer, The mean value of the first feature in the coverage area at the water level layer; To adjust the bias parameter of the model's output matching degree, is the error correction parameter of the reference value in the water level layer.

[0070] Among them, the tuning model introduces weight coefficients to reflect the degree of influence of different fluctuation characteristics on the concentration of suspended sediments. The weight coefficient takes into account the mean data, which objectively helps to reduce the accuracy and influence of abnormal fluctuations or deviation data. At the same time, the model also sets a bias parameter for the degree of output matching, so that the specific condition influencing parameters of different water level layers can be incorporated into the tuning model when necessary. Finally, the introduction of error correction parameters is mainly used to compensate for the differences in the characteristic values ​​of suspended sediment concentration in the tuning model, so that the tuning model can be adjusted for possible accuracy deviations of the sensors of the hydrographic instrument.

[0071] Among them, the training model uses The mean value of the effective fluctuation characteristics in the water level layer is mathematically transformed. Considering that when the concentration of suspended sediment increases, the increase in concentration may cause the light wave characteristics to change in a nonlinear manner, which easily leads to a non-linear relationship between the variable and the output result. Therefore, by using a logarithmic function, the nonlinear relationship can be effectively converted into an approximately linear form, thereby The description can enhance the robustness of the model, enabling the model to better handle extreme values ​​and nonlinear relationships in the data.

[0072] Among them, among them, is an exponential transformation of the mean of the first feature in the coverage area in the water level layer. When the value is large, it can be significantly enlarged The influence of this data is so that the first feature within the coverage area produces a more significant output impact through the exponential function.

[0073] Furthermore, calculating the mean of the first feature in the water level layer includes the following steps: determining the water level layer to be calculated and obtaining the corresponding trend graph and coverage area; extracting all the first features of the corresponding coverage area on the trend graph, obtaining the suspended sediment concentration data corresponding to each first feature, and summarizing them to obtain a first data set; calculating the mean of the first feature in the water level layer in its coverage area through the first data set , the formula is as follows:

[0074] ;

[0075] In the formula, is the total value of all suspended sediment concentration data in the first data set; is the number of the first feature in the first data set; is the qth suspended sediment concentration data in the first data set.

[0076] For reference, by calculating the mean of all suspended sediment concentration data in the first data set of the coverage area, representative data of all first data in the water level layer coverage area can be obtained. And the mean value, as an indicator of central tendency, can reflect the point value of the overall sediment concentration in the coverage area, while reducing the impact of a single first data in the coverage area on the overall judgment.

[0077] Furthermore, calculating the mean of the effective fluctuation characteristics in the water level layer includes the following steps: obtaining the fluctuation characteristics placed outside the coverage area on the trend graph, determining the number of effective fluctuation characteristics, extracting the suspended sediment concentration data corresponding to the effective fluctuation characteristics, and summarizing them to obtain a second data set; calculating the mean of the effective fluctuation characteristics in the water level layer through the second data set , the formula is as follows:

[0078] ;

[0079] In the formula, is the total value of the suspended sediment concentration data of the fluctuation characteristics in the second data set, N is the number of valid fluctuation characteristics, is the point value corresponding to the i-th valid fluctuation feature in the second data set.

[0080] The principle is the same as the above calculation of the mean value of the first feature in the water level layer. During the monitoring process, the fluctuation characteristics outside the coverage area are obtained. By calculating the mean value of these fluctuation characteristics, the overall fluctuation trend of the effective fluctuation characteristics outside the entire coverage area in the water level layer can be evaluated, and a more objective point value can be obtained to reduce the impact of individual anomalies.

[0081] Based on the above embodiment, another embodiment of the present invention is that the number of the above monitoring lines is greater than three and more than two optical wave lines are set, and the hydrographic instruments on each optical wave line are distributed the same in the longitudinal direction, so that there are at least two or more hydrographic instruments on the same water level layer. By setting more monitoring lines and corresponding optical wave lines, the collection of hydrographic instruments in the monitored water area is more intensive. Among them, the above server marks the hydrographic instruments placed on the optical wave line as the master device, and the hydrographic instruments placed on the sampling line as the slave device. Through the linkage between the master device and the slave device, it is ensured that the data collected by each master device can be corrected by the slave device. Among them, when the server obtains the location information of the hydrographic instrument, the master device placed in the same water level layer is marked. By marking, the second data obtained by the slave device can correspond to the first data obtained by the master device, in the hope of correcting the first data by the second data.

[0082] One setting method of this embodiment is that the above-mentioned server sets the same water level layer as the working area. When executing S300, the server averages and fuses the first features sent by the main device of the working area at the same time; third data is formed, and the third data is regarded as the first data to execute S400. By setting the same water level layer as the working area, the first feature of the water level layer is obtained, and the first feature is averaged and fused. By numerically averaging the first features collected by the hydrological instrument placed on the light wave line at the same time point and the same water level layer, a comprehensive and representative third data is generated. By averaging the data of multiple devices, the measurement error caused by environmental interference or self-failure of individual devices can be reduced; and the averaged and fused data comes from multiple points in the water level layer, which can more comprehensively reflect the comprehensive situation of the suspended sediment concentration in the entire water level layer and reduce the influence of local or single device locations.

[0083] If necessary, the server can dynamically adjust the emission frequency and power of the light wave line and the sampling frequency of the monitoring line according to the real-time monitored sediment concentration changes. When the monitored sediment concentration is high, the emission frequency and power of the light wave line will be increased to improve the sensitivity and accuracy of monitoring; otherwise, it will be appropriately reduced to save energy and reduce equipment wear. Through remote control by the server, the sampling frequency of the hydrometer can be adjusted according to the sampling of changes in sediment concentration to ensure that the change information can be captured in time when the sediment concentration changes drastically, and reduce unnecessary sampling when the sediment concentration is stable, thereby optimizing the overall monitoring efficiency.

[0084] One setting method of this embodiment is that the server sets several boundary areas in the same water level layer, and regards each boundary area as a working area. When the server obtains the location information of the hydrometer, it obtains the number of master devices in the working area, and the server averages and fuses the first features sent by the master devices in the working area at the same time to form fourth data, and regards the fourth data as the first data to execute S400

[0085] By setting boundary areas, each boundary area is regarded as a work area. This method is usually suitable for different geographical influencing factors in a water level layer. By dividing different work areas, each work area is used as an independent collection area. Independent work areas can isolate the impact of different geographical factors on the data and avoid abnormal conditions in one area, such as local surges or obstacles, which interfere with data collection and analysis in other areas. In addition, independent work area management helps to provide real-time feedback on the monitoring status of each area and divide large-scale monitoring data into multiple small-scale data sets, which facilitates data storage, management and retrieval and reduces the complexity of data processing.

[0086] Based on the above embodiment, another embodiment of the present invention further includes S600, where a sampling line is sent to the monitored water area and corresponds to the optical wave line, so that the optical wave line and the hydrological instrument of the sampling line are re-associated; and steps S300 to S500 are repeated. By sending the sampling line again, the sampling line can be associated with the optical wave line each time. By repeating steps S300 to S500, the sampling line can be cyclically used to correct the first feature collected by the optical wave line.

[0087] Furthermore, in order to better understand the impact of fluctuations in the water environment, when the above-mentioned fluctuation characteristics are judged to be valid, the server records the timestamp and device code of the fluctuation characteristics marked as abnormal, and the server configures the number of verifications. By recording the timestamp and device code of the abnormal fluctuation characteristics, when a certain device frequently records abnormal data, it can determine the data abnormality and increase attention. By recording the number of abnormalities, on the one hand, it is possible to filter out the device data that has always had problems, and on the other hand, it may be that after the sampling environment changes, the correlation of historical data is reduced. Therefore, when the number of times the server records the same device code is greater than the number of verifications, the server re-acquires the historical data of the monitored water area, adjusts the parameters of R and adjusts the reference value. The historical data acquisition time is modified by adjusting the parameters of R to ensure that the baseline value reflects the actual situation and the latest changes in the current water area.

[0088] For reference, the relevance of historical data may decrease over time. For example, the water environment is significantly affected by seasonal changes. For example, the sediment concentration in the rainy and dry seasons will fluctuate greatly. The R month selected for its historical data may not reflect these seasonal changes in a timely manner, so the selection of months needs to be adaptively adjusted. By dynamically adjusting the parameters of R, it can ensure that the baseline value is based on more relevant and recent data, improving the accuracy of data analysis and anomaly detection.

[0089] The "one embodiment", "another embodiment", "embodiment", "preferred embodiment" and the like mentioned in this specification refer to the specific features, structures or characteristics described in conjunction with the embodiment as included in at least one embodiment generally described in this application. The same expression appearing in multiple places in the specification does not necessarily refer to the same embodiment. Further, when a specific feature, structure or characteristic is described in conjunction with any embodiment, it is claimed that the realization of such feature, structure or characteristic in conjunction with other embodiments also falls within the scope of the present invention.

[0090] Although the present invention is described herein with reference to a number of illustrative embodiments of the present invention, it will be appreciated that those skilled in the art may devise many other modifications and implementations that fall within the scope and spirit of the principles disclosed herein. More specifically, within the scope of the present disclosure, drawings, and claims, a variety of variations and modifications may be made to the components and / or layout of the subject combination layout. In addition to the variations and modifications made to the components and / or layout, other uses will also be apparent to those skilled in the art.

Claims

1. A method for optimizing suspended sediment monitoring accuracy, characterized in that: The method comprises the following steps: S100, deploying a plurality of hydrographic instruments to the monitoring water area, so that the hydrographic instruments form sampling points in the monitoring water area, and a monitoring line is formed in the vertical height by the plurality of sampling points; setting more than two monitoring lines and dividing them into light wave lines and sampling lines according to their functions; S200, the hydrometer exchanges signals with the server, constructs a coordinate system through the location information of the hydrometer, associates and marks the hydrometers of the light wave line and the sampling line, and associates the device codes of the hydrometers placed in the same water level layer with each other; S300, configuring a monitoring period, the hydrological instrument of the optical wave line sends a plurality of first features acquired by the optical wave element in a time-sharing manner during the monitoring period, and the server generates a trend graph of the corresponding device code according to the timestamp of the first features; the hydrological instrument on the sampling line acquires a second feature, inputs the second feature to the server and associates it with the trend graph; The first feature includes suspended sediment concentration data with a timestamp and a device code; the second feature includes suspended sediment concentration data obtained by weighing; S400, the server uses any trend graph as first data, obtains a second feature associated with the trend graph as second data, and presets an error coefficient, maps the second data on the trend graph corresponding to the first data according to the error coefficient to form a coverage area, and the server determines whether the trend graph and the coverage area are similar; If the trend graph is completely within the coverage area, it is considered to be approximate, and the second data is used as the characteristic value of the water level layer for recording and monitoring; If any area of ​​the trend graph is outside the coverage area, it is considered not approximate, and the server extracts the first feature outside the coverage area on the trend graph as a fluctuation feature; and executes S500; S500, obtaining historical data of the monitored water area, determining the average value of the suspended sediment concentration of the water level layer in R months in the historical data and using it as a reference value; wherein R is any positive integer between 1 and 12, and R months represent specific months; obtaining the difference between each fluctuation feature and the reference value to calculate the deviation average value, and the server determines whether the fluctuation feature is valid through the deviation average value; wherein, if the fluctuation feature is determined to be invalid, the server issues an alarm and reconfigures the monitoring period; If the fluctuation feature is judged to be valid, the server inputs the fluctuation feature and the first feature in the coverage area into the training model, and the training model performs weighted calculation and outputs the feature value of the water level layer; The training model calculates the characteristic value of the suspended sediment concentration in the water level layer The calculation formula is: ; In the formula, is the characteristic value of the suspended sediment concentration in the water level layer, is the weight coefficient of the fluctuation characteristics in the water level layer, is the weight coefficient of the first feature in the coverage area in the water level layer, is the mean value of the effective fluctuation characteristics in the water level layer, The mean value of the first feature in the coverage area at the water level layer; To adjust the bias parameter of the model's output matching degree, is the error correction parameter of the reference value in the water level layer; Calculating the mean value of the first feature in the water level layer includes the following steps: determining the water level layer to be calculated and obtaining the corresponding trend graph and coverage area; extracting all the first features of the corresponding coverage area on the trend graph, obtaining the suspended sediment concentration data corresponding to each first feature, and summarizing them to obtain a first data set; calculating the mean value of the first feature in the water level layer in its coverage area through the first data set. , Calculating the mean of the effective fluctuation characteristics in the water level layer includes the following steps: obtaining the fluctuation characteristics placed outside the coverage area on the trend graph, determining the number of effective fluctuation characteristics, extracting the suspended sediment concentration data corresponding to the effective fluctuation characteristics, and summarizing them to obtain a second data set; calculating the mean of the effective fluctuation characteristics in the water level layer through the second data set .

2. The method for optimizing suspended sediment monitoring accuracy according to claim 1, characterized in that: In S500, the specific steps of the server judging whether the fluctuation feature is valid by the deviation average value are: S501, presetting a blinding parameter and a safety threshold, the server identifying whether the fluctuation feature reaches the blinding parameter, and marking the fluctuation feature greater than or equal to the blinding parameter as abnormal; Fluctuating features below the blinding parameter were labeled as normal; S502, the server calculates the difference between the fluctuation feature marked as normal and the reference value to obtain a difference set; S503, obtaining the average value of the difference set to obtain the average deviation value, and comparing the average deviation value with the safety threshold; If the average deviation exceeds the safety threshold, the fluctuation feature is judged to be invalid; if the average deviation is within the safety threshold, the fluctuation feature is judged to be valid.

3. The method for optimizing suspended sediment monitoring accuracy according to claim 1, characterized in that: Calculate the mean value of the first feature in the water level layer in its coverage area through the first data set , the formula is as follows: ； In the formula, is the total value of all suspended sediment concentration data in the first data set; is the number of the first feature in the first data set; is the qth suspended sediment concentration data in the first data set.

4. The method for optimizing suspended sediment monitoring accuracy according to claim 3 is characterized in that: Calculate the mean of the effective fluctuation characteristics in the water level layer using the second data set , the formula is as follows: ； In the formula, is the total value of the suspended sediment concentration data of the fluctuation characteristics in the second data set, N is the number of valid fluctuation characteristics, is the point value corresponding to the i-th valid fluctuation feature in the second data set.

5. The method for optimizing suspended sediment monitoring accuracy according to claim 1, characterized in that: The number of monitoring lines is greater than three and more than two optical wave lines are provided, and the hydrological instruments on each optical wave line are distributed in the same manner in the longitudinal direction, so that at least two or more hydrological instruments on the optical wave line exist in the same water level layer; The server marks the hydrographic instrument placed on the optical wave line as the master device and marks the hydrographic instrument placed on the sampling line as the slave device; When the server obtains the location information of the hydrometer, it marks the main device placed in the same water level layer.

6. The method for optimizing suspended sediment monitoring accuracy according to claim 5, characterized in that: The server sets the same water level layer as a working area, and when executing S300, the server averages and fuses the first features sent by the main devices in the working area at the same time; The third data is formed, and S400 is executed by treating the third data as the first data.

7. The method for optimizing suspended sediment monitoring accuracy according to claim 5, characterized in that: The server sets a number of boundary areas at the same water level layer, and regards each boundary area as a working area; When the server obtains the location information of the hydrological instrument, it obtains the number of main devices in the working area, and the server averages and fuses the first features sent by the main devices in the working area at the same time to form fourth data, and regards the fourth data as the first data to execute S400.

8. The method for optimizing suspended sediment monitoring accuracy according to claim 1, characterized in that: The method further includes S600, wherein the sampling line is sent to the monitored water area and corresponds to the light wave line, so that the light wave line and the hydrological instrument of the sampling line are re-associated; and steps S300 to S500 are repeated.

9. The method for optimizing suspended sediment monitoring accuracy according to claim 8, characterized in that: When the fluctuation feature is judged to be valid, the server records the timestamp and device code of the fluctuation feature marked as abnormal, and the server configures the number of verifications. When the number of times the server records the same device code is greater than the number of verifications, the server re-acquires the historical data of the monitored waters, adjusts the parameters of R and adjusts the reference value.

Images