Atmospheric precipitable water amount data fusion system based on artificial neural network

Abstract

The invention discloses an atmospheric precipitable water amount data fusion system based on an artificial neural network, and relates to the technical field of data fusion, the atmospheric precipitable water amount data fusion system comprises an arrangement duplicate checking end, a noise detection end and a danger early warning end, the arrangement duplicate checking end is used for collecting real-time data of atmospheric precipitable water amount in real time; by setting an atmospheric precipitable water quantity data parameter model based on an artificial neural network, a noise detection end is used for receiving data in real time, performing hierarchical arrangement on data values after data fusion, detecting noise or abnormal values generated in the data fusion process in real time, and performing data fusion comparison in real time, so that the atmospheric precipitable water quantity is obtained. And the danger early warning end is used for predicting the danger of the current atmospheric precipitable water in advance by detecting the climate in real time and combining historical data. According to the atmospheric precipitable water volume data fusion system based on the artificial neural network, confusion or congestion is avoided during data fusion, the accuracy of data fusion is improved, and early warning reminding of fusion abnormity is sent out in real time.

Description

Technical Field

[0001] The invention relates to the technical field of data fusion, and in particular to an atmospheric precipitable water data fusion system based on artificial neural network. Background Art

[0002] The atmospheric precipitable water data fusion system based on artificial neural network is an advanced technology that uses a generalized regression neural network model to calibrate and optimize multi-source PWV data. With the development of artificial intelligence technology, the application of GRNN model in the field of atmospheric science will be more extensive. In the future, it may be combined with advanced technologies such as deep learning to further improve performance. Continuous research will be committed to solving the challenges in multi-source data fusion, such as data heterogeneity and uncertainty, and promoting the further development of meteorological forecasting technology.

[0003] Publication No. CN115392401A discloses a method for fusion of atmospheric precipitation water content data based on artificial neural network. It includes the following steps: step 1: keep PWV data from different sources consistent in time; step 2: keep PWV data from different sources consistent in space; step 3: perform quality control; step 4: build a generalized regression neural network model; step 5: obtain GRNN model and its accuracy evaluation information. The present invention has the functions of calibrating PWV system deviation and improving PWV accuracy, can realize the fusion of multi-source PWV data, and has the advantages of simple method, high accuracy and high data utilization.

[0004] After searching the above patents, it was found that the atmospheric precipitable water data fusion system based on artificial neural networks still has some shortcomings: 1. When data fusion is required, data values ​​may be repeated. When multiple data are repeated, it will cause confusion or congestion in the data fusion, affecting the accuracy of data fusion; 2. During data fusion, if the input data quality is not high, such as containing a lot of noise or outliers, the performance of the model will be greatly reduced, resulting in errors in the atmospheric precipitable water data, further affecting the accuracy of fusion; 3. Global climate change may lead to changes in the state of the atmosphere. These changes may affect the distribution characteristics of PWV data, and in extreme weather conditions such as heavy rain or drought, it is impossible to warn of danger in advance. The dynamic range of PWV data may exceed the training range of the model, resulting in inaccurate prediction results, thereby affecting the accuracy and reliability of the model.

[0005] Therefore, an atmospheric precipitable water data fusion system based on artificial neural network is proposed to solve the above problems. Summary of the invention

[0006] The main purpose of the present invention is to provide an atmospheric precipitable water data fusion system based on artificial neural network to solve the problems raised in the above background.

[0007] To achieve the above object, the technical solution adopted by the present invention is: an atmospheric precipitation water data fusion system based on artificial neural network, including an arrangement and duplication checking end, a noise detection end and a danger warning end, wherein the arrangement and duplication checking end, the noise detection end and the danger warning end are jointly provided with a visual data fusion module;

[0008] The arrangement and duplication checking end is used to collect real-time data of atmospheric precipitation in real time, and to perform data duplication checking on the received data parameters by setting an atmospheric precipitation data parameter model based on an artificial neural network, and to arrange the data in sequence according to the time sequence after duplication checking, and then to perform a hierarchical secondary arrangement on the sequence-arranged data, and to automatically crush the duplicate data after duplication checking, and to track the crushing status in real time;

[0009] The noise detection end is used to receive data in real time, arrange the data values ​​after data fusion in stages, detect the noise or abnormal values ​​generated in the data fusion process in real time, and issue an early warning reminder of fusion abnormality in real time by performing data fusion comparison in real time;

[0010] The danger warning terminal is used to generate a dynamic precipitation range map after color marking according to climate change through real-time detection of the climate, and to make an early prediction of the danger of the current atmospheric precipitation in combination with historical data, and to make a regional judgment between the predicted danger range and the model range to determine whether the current area is within the danger range;

[0011] The visual data fusion module is used to display the real-time status of data fusion in real time through a visualization device.

[0012] The arrangement and duplication checking terminal includes a data collection module, a duplication checking and truncation module and a sorting and tracking module;

[0013] The data acquisition module includes a data acquisition unit, a time-space unit and a fusion model unit;

[0014] The data acquisition unit is used to collect the atmospheric precipitation water data at the current moment in real time through a data acquisition instrument, and arrange the data in sequence according to time;

[0015] The fusion model unit is used to set a data fusion model, which includes data quantity, fusion parameters and data combination.

[0016] The duplicate checking and truncation module includes a data duplicate checking unit and a sequence number arrangement unit;

[0017] The data duplication checking unit is used to calculate the duplication rate between the data corresponding to the multiple sequence numbers, and the calculation method is as follows:

[0018] Substitute the similarity set into the duplicate checking formula for the real-time received data to be fused, and calculate the duplicate checking rate of the relevant data to be fused. The calculation formula is as follows:

[0019]

[0020] Where D is the repetition rate of the sequence number of the re-integrated data to be checked, sm j is the repetition rate of the sequence number of the j-th query data to be fused, L j is the length of the set of data to be fused for the jth query, D sm =1 means data overlap;

[0021] The sequence number arrangement unit is used to arrange the data after duplication checking by sequence number, find out the repeated data, and arrange the data by sequence number.

[0022] The sorting and tracking module includes a hierarchical arrangement unit, a duplicate checking and shredding unit, and a shredding and tracking unit;

[0023] The hierarchical arrangement unit is used to arrange the data after the calculation and duplication check in hierarchical serial numbers, one layer is the serial number arrangement of normal data, and the other layer is the serial number arrangement of duplicate data;

[0024] The duplicate checking and shredding unit is used to directly shred the duplicate data through a data shredding sensor;

[0025] The crushing tracking unit is used to track the crushing status of duplicate data through a data tracker and serial number arrangement, and calculate the difference between the duplicate data serial numbers before and after crushing. If the difference = 0, it means that the crushing is invalid and continues to crush for the second time. If the difference ≠ 0, it means that the crushing is valid until the duplicate data crushing is completed.

[0026] The noise detection end includes a data receiving module, a noise detection module and a motion tracking module;

[0027] The data receiving module includes a data receiving unit and a fusion and grading unit;

[0028] The data receiving unit is used to receive the data to be fused at the current moment in real time through a data receiver;

[0029] The fusion order unit is used to arrange the data values ​​after data fusion in order in real time through order sorting.

[0030] The noise detection module includes a noise detection unit and a noise threshold unit;

[0031] The noise detection unit is used to calculate the noise generated in the data fusion process at the current moment in real time, and uses the least squares principle and the forgotten information of the variance to calculate the weight of each sensor through mean square error comparison and then perform weighted fusion. The calculation formula is as follows:

[0032]

[0033] Among them, X is the value of the position to be estimated, X1 is the value of the known sampling point, and d i,o is the distance between the position to be estimated and the i-th sampling point, P is the weight index, which is usually 0.5;

[0034] The noise threshold unit is used to set a standard noise threshold in the data fusion process.

[0035] The motion tracking module includes a fusion comparison unit and a fusion warning unit;

[0036] The fusion comparison unit is used to compare the data to be fused in real time to determine whether there is any abnormality in the data fusion. The specific steps are as follows:

[0037] Step 1: Implement exception fusion through the function formula. The function formula is as follows

[0038]

[0039] Among them, A n Characterize the data anomaly fusion classification results of each period. X, Y, Z and C are the first-order anomaly perception coefficient, the second-order anomaly perception coefficient, the third-order anomaly perception coefficient and the fourth-order anomaly perception coefficient, respectively. X O , Y O , Z O and C O are the first-order fusion value, the second-order fusion value, the third-order fusion value and the fourth-order fusion value, respectively; T represents the time period of data anomaly tracking; α, β, γ and σ are the first-order fusion anomaly weight, the second-order fusion anomaly weight, the third-order fusion anomaly weight and the fourth-order fusion anomaly weight, respectively; and α+β+γ+σ=1;

[0040] Step 2: Detect the fusion anomaly values ​​of each time period. If α+β+γ+σ≠1, it is judged that data fusion is abnormal.

[0041] The fusion warning unit is used to report abnormal fusion to the system, issue a voice alarm, and provide feedback for manual processing.

[0042] The danger warning terminal includes a climate detection module, a data dynamic module and a warning tracking module;

[0043] The climate detection module includes a climate detection unit and a historical data unit;

[0044] The climate detection unit is used to detect the climate parameters at the current moment in real time through the environmental monitoring equipment;

[0045] The historical data unit is used to record the atmospheric precipitable water and climate parameters at the current moment in real time through a data recorder.

[0046] The data dynamic module includes a dynamic range unit;

[0047] The dynamic range unit is used for dynamic demonstration through GIS technology, and produces a dynamic precipitation range map after color marking according to climate change at different times.

[0048] The warning tracking module includes a danger warning unit, a model determination unit and a fusion adjustment unit;

[0049] The danger warning unit is used to predict whether the current atmospheric precipitation is within the danger range. The prediction method is as follows:

[0050] The precipitation range parameter of the corresponding atmospheric position is calculated with the precipitation standard parameter. If the difference is equal to ±0.05, it is judged that the precipitation range is close to the safe value. If the difference is greater than 0.05 or less than -0.05, it is judged that the precipitation range is in the dangerous range, and a disaster warning is issued. One week is counted as one cycle. The precipitation range of the corresponding atmospheric position in the next cycle is predicted based on the precipitation range parameters of three cycles. If the average values ​​of the environmental parameters of the three cycles all exceed the precipitation standard parameters, it is judged that the precipitation range of the corresponding atmospheric position is in the dangerous range, and a danger warning message is issued;

[0051] The model determination unit is used to perform regional determination between the danger range for issuing the danger warning information and the model range, and obtain the specific location of the danger range through the navigation device;

[0052] The fusion adjustment unit is used to adjust abnormal data in real time through data fusion abnormality until the data fusion is normal.

[0053] The present invention has the following beneficial effects:

[0054] 1. In the present invention, by setting an arrangement and duplicate checking end, when fusing the atmospheric precipitation data based on the artificial neural network, by setting the atmospheric precipitation data parameter model based on the artificial neural network, the received data parameters are used to implement data duplicate checking, and then the data arranged in sequence are arranged in layers for a second time, and the crushing status is tracked in real time to avoid the situation where data values ​​are repeated in data fusion, automatically crush duplicate data, and arrange the data in layers, so as to avoid confusion or congestion during data fusion and increase the accuracy of data fusion.

[0055] 2. In the present invention, by setting a noise detection end, when fusing the atmospheric precipitation water content data based on the artificial neural network, the data values ​​after data fusion are arranged in stages, and the noise or abnormal values ​​generated in the data fusion process are detected in real time, so that when the data is fused, the quality of the input data is increased in real time, and the noise or abnormal values ​​generated during the data fusion are reduced. In addition, by performing real-time data fusion comparison, an early warning reminder of fusion abnormality is issued in real time, thereby improving the accuracy of atmospheric precipitation water content data fusion and further reducing the data deviation existing in data fusion.

[0056] 3. In the present invention, by setting a danger warning terminal, when fusing the atmospheric precipitable water data based on the artificial neural network, the current atmospheric precipitable water is predicted in advance by combining historical data, and the predicted danger range and the model range are regionally determined to determine whether the current area is within the danger range, so that the system can predict extreme weather conditions such as heavy rain or drought in advance, warn of danger in advance, avoid the dynamic range of PWV data exceeding the training range of the model, increase the accuracy of the atmospheric precipitable water prediction results, and thus increase the reliability of the model preview. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the overall system architecture of an atmospheric precipitation data fusion system based on an artificial neural network of the present invention;

[0058] Figure 2 It is a schematic diagram of the structure of the arrangement and duplicate checking end of the atmospheric precipitation water data fusion system based on artificial neural network of the present invention;

[0059] Figure 3 It is a schematic diagram of the architecture of a noise detection end of an atmospheric precipitation water data fusion system based on an artificial neural network according to the present invention;

[0060] Figure 4 The present invention is a schematic diagram of the architecture of a danger warning terminal of an atmospheric precipitable water content data fusion system based on an artificial neural network. DETAILED DESCRIPTION

[0061] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0062] Embodiment 1

[0063] Please refer to Figure 1 to Figure 2 As shown: an atmospheric precipitation water data fusion system based on artificial neural network, including an arrangement and duplication checking end, a noise detection end and a danger warning end, wherein the arrangement and duplication checking end, the noise detection end and the danger warning end are jointly provided with a visual data fusion module;

[0064] The arrangement and duplicate checking end is used to collect real-time data of atmospheric precipitation in real time, and to check the received data parameters for duplicates by setting the atmospheric precipitation data parameter model based on artificial neural network. After the data is checked for duplicates, the data is arranged in sequence according to the time sequence, and then the data arranged in sequence is arranged in layers for secondary arrangement. After the duplicate checking, the duplicate data is automatically crushed, and the crushing status is tracked in real time.

[0065] The noise detection end is used to receive data in real time, arrange the data values ​​after data fusion in stages, detect the noise or abnormal values ​​generated in the data fusion process in real time, and issue early warning reminders of fusion abnormalities in real time by performing real-time data fusion comparison;

[0066] The danger warning terminal is used to generate a dynamic precipitation range map through real-time climate detection and color coding according to climate change, and to make an early prediction of the current atmospheric precipitation in combination with historical data. The predicted danger range and the model range are used for regional judgment to determine whether the current area is within the danger range.

[0067] The visual data fusion module is used to display the real-time status of data fusion through visualization equipment.

[0068] The arrangement and duplication checking end includes a data collection module, a duplication checking and truncation module, and a sorting and tracking module;

[0069] The data acquisition module includes a data acquisition unit, a time-space unit and a fusion model unit;

[0070] The data acquisition unit is used to collect the atmospheric precipitation water data at the current moment in real time through the data acquisition instrument, and arrange the data in sequence according to time;

[0071] The fusion model unit is used to set the data fusion model, which includes the number of data, fusion parameters and data combination.

[0072] The duplicate checking and truncation module includes a data duplicate checking unit and a sequence number arrangement unit;

[0073] The data duplication checking unit is used to calculate the duplication rate between the data corresponding to multiple serial numbers. The calculation method is as follows:

[0074] Substitute the similarity set into the duplicate checking formula for the real-time received data to be fused, and calculate the duplicate checking rate of the relevant data to be fused. The calculation formula is as follows:

[0075]

[0076] Where D is the repetition rate of the sequence number of the re-integrated data to be checked, sm j is the repetition rate of the sequence number of the j-th query data to be fused, L j is the length of the set of data to be fused for the jth query, Dsm =1 means data overlap;

[0077] The serial number arrangement unit is used to arrange the data after duplication checking by serial number, find out the duplicate data, and arrange them by serial number.

[0078] The sorting and tracking module includes a hierarchical arrangement unit, a duplicate checking and shredding unit, and a shredding and tracking unit;

[0079] The hierarchical arrangement unit is used to arrange the data after the calculation and duplication check in hierarchical serial numbers, one layer is the serial number arrangement of normal data, and the other layer is the serial number arrangement of duplicate data;

[0080] The duplicate checking and shredding unit is used to directly shred the duplicate data through the data shredding sensor;

[0081] The crushing tracking unit is used to track the crushing status of duplicate data through a data tracker and serial number arrangement, and calculate the difference between the duplicate data serial numbers before and after crushing. If the difference = 0, it means that the crushing is invalid and a second crushing is continued. If the difference ≠ 0, it means that the crushing is valid until the duplicate data crushing is completed.

[0082] By collecting real-time data of atmospheric precipitation and setting an atmospheric precipitable water data parameter model based on an artificial neural network, the received data parameters are checked for duplicate data. After the data is checked for duplicate data, the data are arranged in sequence in chronological order, and then the data arranged in sequence are arranged in layers for a second time. After the duplicate data is checked for duplicate data, the duplicate data are automatically crushed, and the crushing status is tracked in real time to avoid duplicate data values ​​in data fusion. Duplicate data are automatically crushed and the data are arranged in layers to avoid confusion or congestion during data fusion and increase the accuracy of data fusion.

[0083] Embodiment 2

[0084] Please refer to Figure 3 As shown: Based on the first embodiment, the noise detection end includes a data receiving module, a noise detection module and a motion tracking module;

[0085] The data receiving module includes a data receiving unit and a fusion and grading unit;

[0086] The data receiving unit is used to receive the data to be fused at the current moment in real time through a data receiver;

[0087] The fusion hierarchical unit is used to arrange the data values ​​after data fusion in real time through hierarchical sorting.

[0088] The noise detection module includes a noise detection unit and a noise threshold unit;

[0089] The noise detection unit is used to calculate the noise generated in the data fusion process at the current moment in real time. By using the least squares principle and the forgotten information of the variance, the weights of each sensor are calculated through mean square error comparison and then weighted fusion is performed. The calculation formula is as follows:

[0090]

[0091] Among them, X is the value of the position to be estimated, X1 is the value of the known sampling point, and d i,o is the distance between the position to be estimated and the i-th sampling point, P is the weight index, which is usually 0.5;

[0092] The noise threshold unit is used to set the standard noise threshold in the data fusion process.

[0093] The motion tracking module includes a fusion comparison unit and a fusion warning unit;

[0094] The fusion comparison unit is used to compare the data to be fused in real time to determine whether there is any abnormality in the data fusion. The specific steps are as follows:

[0095] Step 1: Implement exception fusion through the function formula. The function formula is as follows

[0096]

[0097] Among them, A n Characterize the data anomaly fusion classification results of each period. X, Y, Z and C are the first-order anomaly perception coefficient, the second-order anomaly perception coefficient, the third-order anomaly perception coefficient and the fourth-order anomaly perception coefficient, respectively. X O , Y O , Z O and C O are the first-order fusion value, the second-order fusion value, the third-order fusion value and the fourth-order fusion value, respectively; T represents the time period of data anomaly tracking; α, β, γ and σ are the first-order fusion anomaly weight, the second-order fusion anomaly weight, the third-order fusion anomaly weight and the fourth-order fusion anomaly weight, respectively; and α+β+γ+σ=1;

[0098] Step 2: Detect the fusion anomaly values ​​of each time period. If α+β+γ+σ≠1, it is judged that data fusion is abnormal.

[0099] The fusion warning unit is used to report abnormal fusion and send a voice alarm to the reporting system for feedback for manual processing.

[0100] By collecting real-time data of atmospheric precipitation and setting an atmospheric precipitable water data parameter model based on an artificial neural network, the received data parameters are checked for duplicate data. After the data is checked for duplicate data, the data are arranged in sequence in chronological order, and then the data arranged in sequence are arranged in layers for a second time. After the duplicate data is checked for duplicate data, the duplicate data are automatically crushed, and the crushing status is tracked in real time to avoid duplicate data values ​​in data fusion. Duplicate data are automatically crushed and the data are arranged in layers to avoid confusion or congestion during data fusion and increase the accuracy of data fusion.

[0101] Embodiment 3

[0102] Please refer to Figure 4 As shown: Based on the first embodiment, the danger warning terminal includes a climate detection module, a data dynamic module and a warning tracking module;

[0103] The climate detection module includes a climate detection unit and a historical data unit;

[0104] The climate detection unit is used to detect the climate parameters at the current moment in real time through the environmental monitoring equipment;

[0105] The historical data unit is used to record the atmospheric precipitable water and climate parameters at the current moment in real time through a data recorder.

[0106] The data dynamic module includes a dynamic range unit;

[0107] The dynamic range unit is used for dynamic demonstration through GIS technology. According to the climate change at different times, dynamic precipitation range map is produced through color identification.

[0108] The early warning tracking module includes a danger early warning unit, a model determination unit, and a fusion adjustment unit;

[0109] The danger warning unit is used to predict whether the current atmospheric precipitation is within the danger range. The prediction method is as follows:

[0110] The precipitation range parameter of the corresponding atmospheric position is calculated with the precipitation standard parameter. If the difference is equal to ±0.05, it is judged that the precipitation range is close to the safe value. If the difference is greater than 0.05 or less than -0.05, it is judged that the precipitation range is in the dangerous range, and a disaster warning is issued. One week is counted as one cycle. The precipitation range of the corresponding atmospheric position in the next cycle is predicted based on the precipitation range parameters of three cycles. If the average values ​​of the environmental parameters of the three cycles all exceed the precipitation standard parameters, it is judged that the precipitation range of the corresponding atmospheric position is in the dangerous range, and a danger warning message is issued;

[0111] The model determination unit is used to perform regional determination between the danger range for which the danger warning information is issued and the model range, and obtain the specific location of the danger range through the navigation device;

[0112] The fusion adjustment unit is used to adjust abnormal data in real time through data fusion abnormality until the data fusion is normal.

[0113] Through real-time detection of the climate, a dynamic precipitation range map is generated after color coding according to climate change, and the current atmospheric precipitable water is predicted in advance based on historical data. The predicted danger range and the model range are regionally determined to determine whether the current area is within the danger range. If it is within the danger range, an early warning is issued and warning information is issued. If it is not within the danger range, the dynamic precipitation range map is tracked in real time, so that the system can predict extreme weather conditions such as heavy rain or drought in advance, issue early warnings of danger, avoid the dynamic range of PWV data exceeding the training range of the model, increase the accuracy of the atmospheric precipitable water prediction results, and thus increase the reliability of model rehearsals.

[0114] In the present invention, a system for fusion of atmospheric precipitation data based on artificial neural network is provided. When the system is in operation, the system collects real-time data of atmospheric precipitation in real time, and realizes data duplication check on the received data parameters by setting the atmospheric precipitation data parameter model based on artificial neural network. After the data duplication check, the data are arranged in sequence according to time order, and then the data arranged in sequence are arranged in layers for secondary arrangement. After the duplication check, the duplicate data are automatically crushed, and the crushing status is tracked in real time to avoid the situation where data values ​​are repeated in data fusion, the duplicate data are automatically crushed, and the data are arranged in layers, so as to avoid confusion or congestion during data fusion and increase the accuracy of data fusion. The system receives data in real time, arranges the data values ​​after data fusion in stages, and detects the noise or abnormal values ​​generated in the process of data fusion in real time. By performing real-time data fusion comparison, an early warning reminder of fusion abnormality is issued in real time, so that the data fusion is accurate. The quality of input data can be increased, the noise or outliers that appear during data fusion can be reduced, and through real-time data fusion comparison, early warning reminders of fusion anomalies can be issued in real time, the accuracy of atmospheric precipitation data fusion can be improved, and the data bias in data fusion can be further reduced; through real-time detection of the climate, a dynamic precipitation range map is generated after color marking according to climate change, and the current atmospheric precipitation can be predicted in advance based on historical data. The predicted danger range and the model range are regionally determined to determine whether the current area is within the danger range. If it is within the danger range, an early warning is issued and a warning message is issued. If it is not within the danger range, the dynamic precipitation range map is tracked in real time, so that the system can predict extreme weather conditions such as heavy rain or drought in advance, warn of danger in advance, avoid the dynamic range of PWV data exceeding the training range of the model, increase the accuracy of the atmospheric precipitation prediction results, and thus increase the reliability of the model rehearsal.

[0115] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An atmospheric precipitable water data fusion system based on artificial neural network, characterized in that: The system includes an arrangement and duplication checking end, a noise detection end and a danger warning end, wherein the arrangement and duplication checking end, the noise detection end and the danger warning end are jointly provided with a visual data fusion module; The arrangement and duplication checking end is used to collect real-time data of atmospheric precipitation in real time, and to perform data duplication checking on the received data parameters by setting an atmospheric precipitation data parameter model based on an artificial neural network, and to arrange the data in sequence according to the time sequence after duplication checking, and then to perform a hierarchical secondary arrangement on the sequence-arranged data, and to automatically crush the duplicate data after duplication checking, and to track the crushing status in real time; The noise detection end is used to receive data in real time, arrange the data values ​​after data fusion in stages, detect the noise or abnormal values ​​generated in the data fusion process in real time, and issue an early warning reminder of fusion abnormality in real time by performing data fusion comparison in real time; The danger warning terminal is used to generate a dynamic precipitation range map after color marking according to climate change through real-time detection of the climate, and to make an early prediction of the danger of the current atmospheric precipitation in combination with historical data, and to make a regional judgment between the predicted danger range and the model range to determine whether the current area is within the danger range; The visual data fusion module is used to display the real-time status of data fusion in real time through a visualization device.

2. The system according to claim 1, characterized in that The arrangement and duplication checking terminal includes a data collection module, a duplication checking and truncation module and a sorting and tracking module; The data acquisition module includes a data acquisition unit, a time-space unit and a fusion model unit; The data acquisition unit is used to collect the atmospheric precipitation water data at the current moment in real time through a data acquisition instrument, and arrange the data in sequence according to time; The fusion model unit is used to set a data fusion model, which includes data quantity, fusion parameters and data combination.

3. The system according to claim 2, characterized in that The duplicate checking and truncation module includes a data duplicate checking unit and a sequence number arrangement unit; The data duplication checking unit is used to calculate the duplication rate between the data corresponding to the multiple sequence numbers, and the calculation method is as follows: Substitute the similarity set into the duplicate checking formula for the real-time received data to be fused, and calculate the duplicate checking rate of the relevant data to be fused. The calculation formula is as follows: Where D is the repetition rate of the sequence number of the re-integrated data to be checked, sm j is the repetition rate of the sequence number of the j-th query data to be fused, L j is the length of the set of data to be fused for the jth query, D sm =1 means data overlap; The sequence number arrangement unit is used to arrange the data after duplication checking by sequence number, find out the repeated data, and arrange the data by sequence number.

4. The system according to claim 3, characterized in that The sorting and tracking module includes a hierarchical arrangement unit, a duplicate checking and shredding unit, and a shredding and tracking unit; The hierarchical arrangement unit is used to arrange the data after the calculation and duplication check in hierarchical serial numbers, one layer is the serial number arrangement of normal data, and the other layer is the serial number arrangement of duplicate data; The duplicate checking and shredding unit is used to directly shred the duplicate data through a data shredding sensor; The crushing tracking unit is used to track the crushing status of duplicate data through a data tracker and serial number arrangement, and calculate the difference between the duplicate data serial numbers before and after crushing. If the difference = 0, it means that the crushing is invalid and continues to crush for the second time. If the difference ≠ 0, it means that the crushing is valid until the duplicate data crushing is completed.

5. The system according to claim 1, characterized in that The noise detection end includes a data receiving module, a noise detection module and a motion tracking module; The data receiving module includes a data receiving unit and a fusion and grading unit; The data receiving unit is used to receive the data to be fused at the current moment in real time through a data receiver; The fusion order unit is used to arrange the data values ​​after data fusion in order in real time through order sorting.

6. The system according to claim 5, characterized in that The noise detection module includes a noise detection unit and a noise threshold unit; The noise detection unit is used to calculate the noise generated in the data fusion process at the current moment in real time, and uses the least squares principle and the forgotten information of the variance to calculate the weight of each sensor through mean square error comparison and then perform weighted fusion. The calculation formula is as follows: Among them, X is the value of the position to be estimated, X1 is the value of the known sampling point, and d i,o is the distance between the position to be estimated and the i-th sampling point, P is the weight index, which is usually 0.5; The noise threshold unit is used to set a standard noise threshold in the data fusion process.

7. The system according to claim 6, characterized in that The motion tracking module includes a fusion comparison unit and a fusion warning unit; The fusion comparison unit is used to compare the data to be fused in real time to determine whether there is any abnormality in the data fusion. The specific steps are as follows: Step 1: Implement exception fusion through the function formula. The function formula is as follows Among them, A n Characterize the data anomaly fusion classification results of each period. X, Y, Z and C are the first-order anomaly perception coefficient, the second-order anomaly perception coefficient, the third-order anomaly perception coefficient and the fourth-order anomaly perception coefficient, respectively. X O , Y O , Z O and C O are the first-order fusion value, the second-order fusion value, the third-order fusion value and the fourth-order fusion value, respectively; T represents the time period of data anomaly tracking; α, β, γ and σ are the first-order fusion anomaly weight, the second-order fusion anomaly weight, the third-order fusion anomaly weight and the fourth-order fusion anomaly weight, respectively; and α+β+γ+σ=1; Step 2: Detect the fusion anomaly values ​​of each time period. If α+β+γ+σ≠1, it is judged that data fusion is abnormal. The fusion warning unit is used to report abnormal fusion to the system, issue a voice alarm, and provide feedback for manual processing.

8. The system according to claim 1, characterized in that The danger warning terminal includes a climate detection module, a data dynamic module and a warning tracking module; The climate detection module includes a climate detection unit and a historical data unit; The climate detection unit is used to detect the climate parameters at the current moment in real time through the environmental monitoring equipment; The historical data unit is used to record the atmospheric precipitable water and climate parameters at the current moment in real time through a data recorder.

9. The system according to claim 8, characterized in that The data dynamic module includes a dynamic range unit; The dynamic range unit is used for dynamic demonstration through GIS technology, and produces a dynamic precipitation range map after color marking according to climate change at different times.

10. The system according to claim 9, characterized in that The warning tracking module includes a danger warning unit, a model determination unit and a fusion adjustment unit; The danger warning unit is used to predict whether the current atmospheric precipitation is within the danger range. The prediction method is as follows: The precipitation range parameter of the corresponding atmospheric position is calculated with the precipitation standard parameter. If the difference is equal to ±0.05, it is judged that the precipitation range is close to the safe value. If the difference is greater than 0.05 or less than -0.05, it is judged that the precipitation range is in the dangerous range, and a disaster warning is issued. One week is counted as one cycle. The precipitation range of the corresponding atmospheric position in the next cycle is predicted based on the precipitation range parameters of three cycles. If the average values ​​of the environmental parameters of the three cycles all exceed the precipitation standard parameters, it is judged that the precipitation range of the corresponding atmospheric position is in the dangerous range, and a danger warning message is issued; The model determination unit is used to perform regional determination between the danger range for issuing the danger warning information and the model range, and obtain the specific location of the danger range through the navigation device; The fusion adjustment unit is used to adjust abnormal data in real time through data fusion abnormality until the data fusion is normal.

Images