Marine environment information measurement system and method based on multi-source data fusion

Abstract

The invention relates to a marine environment information measuring system and method based on multi-source data fusion. The system is characterized in that a seawater monitoring system is used for monitoring and collecting seawater pressure, temperature, salinity, sound velocity and seawater transparency data; the hydrological monitoring subsystem is used for detecting and collecting flow velocity and flow direction data of seawater; the atmospheric monitoring system is used for monitoring atmospheric visibility data in real time, measuring and recording air temperature, humidity, air pressure, wind speed, wind direction and rainfall data of monitoring points of the monitoring platform in real time, acquiring influences of the monitoring points on sea surface atmospheric meteorological data and acquiring position data of observation points; the data transmission and processing system performs marine surveying and mapping data inspection on the monitoring data through a robust M estimation inspection method, and accurately locates and eliminates abnormal values in the data; and data parameters are corrected to obtain hydrological monitoring data correctly reflecting the measuring points. According to the invention, marine environment monitoring information can be comprehensively presented from different dimensions.

Description

Technical Field

[0001] The present invention belongs to the technical field of ocean surveying and mapping, and in particular relates to an ocean environment information measurement system and method for multi-source data fusion. Background Art

[0002] With the development of the economy and the requirements of national development, marine environmental information services are of great significance for the safe navigation of ships and the protection of the maritime battlefield environment. The acquisition of marine environmental information relies on marine environmental measurement operations. It is an inevitable trend for the development of marine environment to establish a marine environmental measurement system with a wide monitoring coverage area and rich monitoring methods. Traditional marine environmental information is measured based on independent sensors, without fully exploring data relationships and correcting data. The degree of systematization is not enough and cannot fully reflect the trend of marine environmental changes and the overall situation. Therefore, there is an urgent need for a marine environmental information measurement system with multi-source data fusion and a method for realizing measurement, which expands the technical ideas for subsequent marine environmental mapping. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention proposes a marine environment information measurement system and method that can comprehensively present marine environment monitoring information from different dimensions and provide users with high-quality marine environment measurement data through multi-source data fusion.

[0004] One of the above objects of the present invention is achieved by the following technical solution:

[0005] A multi-source data fusion marine environment information measurement system, characterized in that it includes a seawater monitoring subsystem, a hydrological monitoring subsystem, a meteorological monitoring subsystem and a data transmission and processing system; the seawater monitoring system includes a temperature, salinity, depth and sound velocity measuring device and a seawater visibility measuring device, which are used to monitor and collect seawater pressure, temperature, salinity, sound velocity and seawater transparency data; the hydrological monitoring subsystem adopts a current profiler, which is used to detect and collect seawater flow velocity and flow direction data; the atmospheric monitoring system includes an atmospheric visibility meter, a buoy, a meteorological sensor and a GPS unit; the atmospheric visibility meter is installed on a monitoring platform and is used to monitor atmospheric visibility data in real time; a plurality of buoys are arranged around the monitoring platform, and each buoy is provided with a meteorological sensor and a GPS unit, which are used to measure and record the temperature, humidity, air pressure, wind speed, wind direction and rainfall data of the corresponding monitoring point of the monitoring platform in real time, and obtain the influence of the monitoring point on the sea surface atmospheric meteorological data; GPS is used to obtain the location data of the observation point, mainly including coordinate data and time data;

[0006] The data transmission and processing system includes a data transmission unit and a data processing unit. The data transmission unit transmits the monitoring data to the data processing unit quickly and effectively by wireless or wired means. The data processing unit performs marine surveying and mapping data inspection on the monitoring data by using the robust M estimation inspection method to accurately locate and remove abnormal values ​​in the data. The monitoring data with abnormal values ​​removed is further corrected for data parameters to obtain hydrological monitoring data that correctly reflects the measurement point.

[0007] The second objective of the present invention is achieved by the following technical solutions:

[0008] A measurement method of the marine environment information measurement system based on the above multi-source data fusion is characterized by comprising the following steps:

[0009] Step 1: Obtain monitoring data of the sea area through the hydrological monitoring subsystem, the seawater monitoring subsystem and the meteorological monitoring subsystem;

[0010] Step 2: The monitoring data is quickly and effectively transmitted to the data processing unit by wireless or wired means through the data transmission unit. The data processing unit performs marine surveying and mapping data inspection on the monitoring data through the robust M estimation inspection method, accurately locates the outliers in the data, and marks and removes the outliers;

[0011] Step 3: further correct the data parameters by using the correction formula after removing the outliers to obtain the hydrological monitoring data that correctly reflects the measurement point.

[0012] Moreover, in step 2, the method used to remove outliers is to "smoothly" resist the interference of abnormal data through continuous weight reduction and iterative calculation. The process is as follows:

[0013] The general expression of the fitted model is:

[0014] V=AX-L (1)

[0015] The robust solution of the robust M estimate is:

[0016]

[0017] Its iterative solution is:

[0018]

[0019] In the formula, is the equivalent full matrix; A is the known coefficient matrix; X is the model parameter vector to be determined; L is the free term; the IGGIII weight function is used

[0020]

[0021] Among them, k0 is generally 1.5, k1 is generally 2.5,

[0022] Through the iterative solution, the calculation formula of the iterative weighted average based on the robust M estimation can be written as follows:

[0023]

[0024] In the formula, z i Represents the observed value in the neighborhood around the detected point, and the equivalent weight Determined by formula (4), formula (5) is exactly the same as the traditional weighted average calculation formula in form. The difference is that the weight of this robust weighted average changes during the iteration process, and the value of each parameter in formula (5) is determined by the residual obtained in the previous iteration;

[0025] The original weight factor p of the observations used to calculate the mean i , according to the distance from the observation point to the detected point, it is calculated by the following formula:

[0026] p i =1 / (d i +ε) 2 (6)

[0027] Among them, d i is the horizontal distance from the observation point i to the detected point. In order to prevent the denominator of the weight function from being 0, the distance d needs to be given i Add a small positive number ε;

[0028] Suppose after n iterations, the estimated value of the detected point p is The actual data value of point p is z p The prediction residual can be expressed as:

[0029]

[0030] The abnormality test standard of point P is:

[0031]

[0032] In the formula, is the mean square error of the inferred value, m p The mean square error is the observed error, and the measurement accuracy index is determined. k is usually 2 or 3;

[0033] According to the weighted iterative weighted average model, the value of the prediction is basically stable after n iterations. And find the mean square error of the estimated value and the observation error m p If the value of If it holds, then point P is an outlier and should be marked or removed.

[0034] Moreover, in step 3, the correction formula is: parameter true value = parameter reading * slope, where

[0035]

[0036] Where: a i represents the parameter measured by the sensor for the i-th water chemical analysis sample, b i represents the parameter of the i-th water chemical analysis sample obtained from the water chemical analysis experiment, and slope represents the corrected value of the slope parameter after correction.

[0037] The advantages and positive effects of the present invention are:

[0038] The present invention realizes three-dimensional monitoring of the marine environment and detection of the atmospheric environment around the monitoring platform through the seawater monitoring subsystem, the hydrological monitoring subsystem and the meteorological monitoring subsystem. All monitoring data are transmitted and processed through the data transmission and processing system to locate and delete abnormal values ​​in the monitoring data, and further correct them, so as to finally obtain high-quality marine environment measurement data. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a marine environment information measurement system integrating multi-source data;

[0040] Figure 2 The flow chart of the marine environment information measurement method based on multi-source data fusion;

[0041] Figure 3 Software architecture design diagram for the marine environment information application system with multi-source data fusion. DETAILED DESCRIPTION

[0042] The structure of the present invention is further described below with reference to the accompanying drawings and by way of examples. It should be noted that the present examples are descriptive rather than restrictive.

[0043] A multi-source data fusion marine environment information measurement system, the invention point is: including a seawater monitoring subsystem, a hydrological monitoring subsystem, a meteorological monitoring subsystem and a data detection and processing system, such as Figure 1 shown.

[0044] 1. Seawater monitoring subsystem

[0045] The seawater monitoring system mainly uses temperature, salinity, depth and sound velocity measuring equipment and seawater visibility measuring equipment to collect seawater data. The collected data mainly includes seawater pressure, temperature, salinity, sound velocity, seawater transparency, etc., to conduct three-dimensional monitoring of the entire marine environment.

[0046] 2. Hydrological monitoring subsystem

[0047] The hydrological monitoring subsystem mainly uses current measuring instruments to measure flow velocity and direction based on acoustic Doppler technology. It has the characteristics of wide measurement range, high precision and strong accuracy.

[0048] 3. Atmospheric monitoring subsystem

[0049] The atmospheric monitoring system mainly includes atmospheric visibility meters, buoys, meteorological sensors, GPS units, etc.

[0050] The integrated atmospheric visibility meter is installed on the monitoring platform to monitor the atmospheric visibility data in real time. There are multiple buoys around the monitoring platform, and each buoy is equipped with a meteorological sensor and GPS system to measure and record key data such as temperature, humidity, air pressure, wind speed, wind direction, rainfall, etc. around the monitoring platform in real time, and obtain the impact of the monitoring point on the sea surface atmospheric meteorological data; GPS is used to obtain the location data of the observation point, mainly including coordinate data and time data, and to conduct qualitative analysis on the collected data.

[0051] 4. The data transmission and processing system includes a data transmission unit and a data processing unit. The data transmission unit transmits the monitoring data to the data processing unit quickly and effectively by wireless or wired means; the data processing unit adopts marine environment data information processing software to comprehensively present the marine environment monitoring information from different dimensions and provide users with high-quality marine environment measurement data.

[0052] The present invention also proposes a method for measuring marine environment information by fusion of multi-source data. Since the amount of marine environment measurement data is very large, the data processing and fusion are divided into three steps:

[0053] Step 1: Obtain monitoring data of the sea area through the hydrological monitoring subsystem, the seawater monitoring subsystem and the meteorological monitoring subsystem;

[0054] Step 2: The monitoring data is quickly and effectively transmitted to the data processing unit by wireless or wired means through the data transmission unit. The data processing unit performs marine surveying and mapping data inspection on the monitoring data through the robust M estimation inspection method, accurately locates the outliers in the data, and marks or removes the outliers;

[0055] Step 3: The data parameters are further corrected by the correction formula after the outliers are removed. The fused data can correctly reflect the seawater environment data of the measurement point, comprehensively present the marine environment monitoring information from different dimensions, and provide users with high-quality marine environment measurement data. Figure 2 shown.

[0056] In step 2, the data processing unit uses the robust M estimation test method to test the monitoring data for marine surveying and mapping. The process of accurately locating outliers in the data is as follows:

[0057] The acquired seawater temperature, salinity, depth, sound speed, seawater transparency, meteorological data and other factors are used to eliminate abnormal data using the robust M estimation test method; through the method of continuous weight reduction and iterative calculation, the interference of abnormal data is "smoothly" resisted.

[0058] The general expression of the fitted model is:

[0059] V=AX-L (1)

[0060] The robust solution of the robust M estimate is:

[0061]

[0062] Its iterative solution is:

[0063]

[0064] In the formula is the equivalent full matrix; A is the known coefficient matrix; X is the model parameter vector to be determined; L is the free term. The IGGIII weight function is used

[0065]

[0066] Among them, k0 is generally 1.5, k1 is generally 2.5,

[0067] Through the iterative solution, the calculation formula of the iterative weighted average based on the robust M estimation can be written as follows:

[0068]

[0069] In the formula, z i Represents the observed value in the neighborhood around the detected point, and the equivalent weight Determined by formula (4), formula (5) is exactly the same as the traditional weighted average calculation formula in form. The difference is that the weight of this robust weighted average changes during the iteration process, and the value of the parameter in formula (5) is determined by the residual obtained from the previous iteration.

[0070] The original weight factor p of the observations used to calculate the mean i , according to the distance from the observation point to the detected point, it is calculated by the following formula:

[0071] p i =1 / (d i +ε) 2 (6)

[0072] Among them, d i is the horizontal distance from the observation point i to the detected point. In order to prevent the denominator of the weight function from being 0, the distance d needs to be given iAdd a small positive number ε. Combining (4) and (6), it can be seen that when the observation point is far away from the detected point, the initial weight p determined by (6) is i When the observed value is abnormal, the absolute value of the corresponding residual V will increase. At this time, the equivalent weight determined by formula (4) will decrease, indicating that the influence of such abnormal data on the detected point will also be weakened. but That is, when an observation value is significantly abnormal, its equivalent weight is zero, and such abnormal values ​​will not affect the prediction results.

[0073] Suppose after n iterations, the estimated value of the detected point p is The actual data value of point p is z p (where z p Not participating The prediction residual can be expressed as:

[0074]

[0075] The abnormality test standard of point P is:

[0076]

[0077] In the formula, is the mean square error of the inferred value, m p To observe the mean square error, the measurement accuracy index is determined, and k is generally taken as 2 or 3.

[0078] According to the weighted iterative weighted average model, the value of the prediction is basically stable after n iterations. And find the mean square error of the estimated value and the observation error m p If the value of If it holds, then point P is an outlier and should be marked or removed.

[0079] Step 3 is as follows:

[0080] After the abnormal data is eliminated by the above robust M estimation test method, the data parameters are further corrected to correctly reflect the hydrological monitoring data of the measurement point. The correction formula is: parameter true value = parameter reading * slope, where

[0081]

[0082] Where: a i represents the parameter measured by the sensor for the i-th water chemical analysis sample, b i represents the parameter of the i-th water chemical analysis sample obtained from the water chemical analysis experiment, and slope represents the corrected value of the slope parameter after correction.

[0083] Through the fusion of multi-source data, the marine environment monitoring information can be comprehensively presented from different dimensions, providing users with high-quality marine environment measurement data. Figure 3 This is a software architecture design diagram for the marine environment information application system that integrates multi-source data. It can be used to complete functions such as storage, query, graphical display, and output of all data, and has the ability to expand functions.

[0084] Although the embodiments and drawings of the present invention are disclosed for illustrative purposes, those skilled in the art will appreciate that various substitutions, changes and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims

1. A marine environment information measurement system based on the above multi-source data fusion, characterized in that: Including seawater monitoring subsystem, hydrological monitoring subsystem, meteorological monitoring subsystem and data transmission and processing system; The seawater monitoring system includes a temperature, salinity, depth, and sound velocity measuring device and a seawater visibility measuring device, which are used to monitor and collect seawater pressure, temperature, salinity, sound velocity, and seawater transparency data; the hydrological monitoring subsystem uses a current profiler to detect and collect seawater flow velocity and flow direction data; the atmospheric monitoring system includes an atmospheric visibility meter, a buoy, a meteorological sensor, and a GPS unit; the atmospheric visibility meter is installed on a monitoring platform and is used to monitor atmospheric visibility data in real time; a plurality of buoys are arranged around the monitoring platform, and each buoy is provided with a meteorological sensor and a GPS unit, which are used to measure and record the temperature, humidity, air pressure, wind speed, wind direction, and rainfall data of the monitoring point of the monitoring platform in real time, and obtain the influence of the monitoring point on the sea surface atmospheric meteorological data; GPS is used to obtain the location data of the observation point, mainly including coordinate data and time data; The data transmission and processing system comprises a data transmission unit and a data processing unit. The data transmission unit transmits the monitoring data to the data processing unit quickly and effectively by wireless or wired means. The data processing unit performs marine surveying and mapping data inspection on the monitoring data by using the robust M estimation inspection method to accurately locate and remove abnormal values ​​in the data. The monitoring data with outliers will be eliminated and the data parameters will be further corrected to obtain hydrological monitoring data that correctly reflects the measurement points.

2. The measurement method of the marine environment information measurement system based on the multi-source data fusion according to claim 1 is characterized in that: The steps include: Step 1: Obtain monitoring data of the sea area through the hydrological monitoring subsystem, the seawater monitoring subsystem and the meteorological monitoring subsystem; Step 2: The monitoring data is quickly and effectively transmitted to the data processing unit by wireless or wired means through the data transmission unit. The data processing unit performs marine surveying and mapping data inspection on the monitoring data through the robust M estimation inspection method, accurately locates the outliers in the data, and marks and removes the outliers; Step 3: further correct the data parameters by using the correction formula after removing the outliers to obtain the hydrological monitoring data that correctly reflects the measurement point.

3. The measuring method of the measuring system based on the above-mentioned multi-source data fusion marine environment information measuring system according to claim 2 is characterized in that: The steps include: In step 2, the method used to remove outliers is to "smoothly" resist the interference of abnormal data through continuous weight reduction and iterative calculation. The process is as follows: The general expression of the fitted model is: V=AX-L(1) The robust solution of the robust M estimate is: Its iterative solution is: In the formula, is the equivalent full matrix; A is the known coefficient matrix; X is the model parameter vector to be determined; L is the free term; the IGGIII weight function is used Among them, k0 is generally 1.5, k1 is generally 2.5, Through the iterative solution, the calculation formula of the iterative weighted average based on the robust M estimation can be written as follows: In the formula, z i Represents the observed value in the neighborhood around the detected point, and the equivalent weight Determined by formula (4), formula (5) is exactly the same as the traditional weighted average calculation formula in form. The difference is that the weight of this robust weighted average changes during the iteration process, and the value of each in formula (5) is determined by the residual obtained in the previous iteration; The original weight factor p of the observations used to calculate the mean i , according to the distance from the observation point to the detected point, it is calculated by the following formula: p i =1 / (d i +e) 2 (6) Among them, d i is the horizontal distance from the observation point i to the detected point. In order to prevent the denominator of the weight function from being 0, the distance d needs to be given i Add a small positive number ε; Suppose after n iterations, the estimated value of the detected point p is The actual data value of point p is z p The prediction residual can be expressed as: The abnormality inspection standard of point P is: In the formula, is the mean square error of the inferred value, m p The mean square error is the observed error, and the measurement accuracy index is determined. k is usually 2 or 3; According to the weighted iterative weighted average model, the value of the prediction is basically stable after n iterations. And find the mean square error of the estimated value and the observation error m p If the value of If it holds, then point P is an outlier and should be marked or removed.

Images