A Monitoring Data Optimization Method for Saline-Alkali Land Improvement Equipment

Abstract

The present invention provides a method for optimizing monitoring data of a saline-alkali land improvement device, comprising the following steps: S1: Collect salt concentration data through the saline-alkali land improvement device, set the traveling path and data collection frequency of the saline-alkali land improvement device, collect saline-alkali land salt data once every preset distance traveled, and record the position information, time information of the data and the monitoring data of the salt concentration sensor; S2: Optimize the anomaly detection of the collected salt concentration data; S3: Store and upload the filtered valid data. The present invention relates to the technical field of data processing. By integrating spatial junction evaluation and historical trend analysis, it solves the misjudgment problem of the traditional COF algorithm in the saline-alkali land junction area, significantly improves the accuracy of data anomaly detection, and provides efficient and reliable technical support for the precise improvement of saline-alkali land.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method for optimizing monitoring data of a saline-alkali land improvement device. Background Art

[0002] The soil salt condition of saline-alkali land is one of the important indicators for evaluating the effect of improvement measures. Therefore, when a saline-alkali land improvement device conducts land improvement operations, it is necessary to monitor and collect the soil salt concentration data in real time. Usually, the saline-alkali land improvement device will collect data at fixed time or distance intervals along a preset travel path to form a data sequence with temporal and spatial correlations.

[0003] For the real-time collected saline-alkali land monitoring data, the COF (Connectivity-based Outlier Factor) anomaly detection algorithm is often used to identify data anomalies. The COF algorithm evaluates the connectivity distance of each data point in a local temporal window to determine whether the data point is an anomaly. However, the salt in saline-alkali land has regional spatial heterogeneity, that is, there may be obvious differences in salt content in different regions. Especially at the regional junction, the salt concentration changes violently.

[0004] The existing COF anomaly detection method mainly evaluates the connectivity distance between each data point and its local neighborhood points. If the salt concentration difference between the data point and the neighborhood data points is significant, the COF score is high, and this data point is easily determined to be an anomaly point. However, in the actual improvement process of saline-alkali land, there are reasonable significant differences in salt content between different saline-alkali regions. When the saline-alkali land improvement device collects data at the junction of different salt concentration regions, the normal salt change easily causes the COF algorithm to have abnormal misjudgments. This misjudgment causes normal data points to be misidentified as anomaly points, reducing the effectiveness of data collection and the reliability of subsequent data analysis. Summary of the Invention

[0005] In view of this, the present invention aims to propose a method for optimizing monitoring data of a saline-alkali land improvement device, which can effectively identify the data points collected by the saline-alkali land improvement device at the junction of different salt regions to improve the spatial adaptability of anomaly data identification of the saline-alkali land improvement device.

[0006] To achieve the above object, the technical solution of the present invention is realized as follows:

[0007] A method for optimizing monitoring data of a saline-alkali land improvement device, comprising the following steps:

[0008] S1: Collect salinity concentration data through the saline-alkali land improvement equipment, set the travel path and data collection frequency of the saline-alkali land improvement equipment, collect the saline-alkali land salinity data once every preset distance traveled, and record the location information, time information of the data and the monitoring data of the salinity concentration sensor;

[0009] S2: Perform anomaly detection optimization on the collected salinity concentration data;

[0010] The step S2 includes:

[0011] S2.1: Calculate the boundary degree factor of the data point through the mean and standard deviation of the salinity concentration based on the historical data in the spatial neighborhood, , and identify the data points at the junction of different salinity regions;

[0012] S2.2: Establish a trend fitting model for the historical data , based on the trend fitting model of the historical data , combine the residual between the current data point and the historical trend and the time interval sensitivity, and calculate the anomaly degree optimization factor ;

[0013] S2.3: By comprehensively considering the boundary degree factor and the anomaly degree optimization factor , adjust the local average linkage distance to obtain the optimized local average linkage distance , optimize the COF anomaly detection algorithm, and screen out the abnormal data;

[0014] S3: Store and upload the screened valid data.

[0015] Furthermore, in the step S2.1, the calculation formula of the boundary degree factor is:

[0016] ;

[0017] Among them, λ is the sensitivity adjustment coefficient, is the salinity concentration value of the th detection data point, is the average salinity concentration of the historical detection data within the spatial neighborhood radius centered on the th detection data point, is the standard deviation of the salinity concentration of the historical data within the spatial neighborhood, : is function, is the historical standard deviation of the salinity concentration representing the entire monitoring area, is a very small positive number.

[0018] Further, in the step S2.2, the anomaly degree optimization factor is calculated by the formula:

[0019] ;

[0020] where is the salt concentration value of the th detection data point, is a very small positive number, is the predicted value of the fitting trend model, is the standard deviation of the historical data fitting residuals, is the time interval between the current data point and the nearest historical data, and △T is the average sampling interval of the historical data.

[0021] Further, in the step S2.3, the optimized local average link distance is adjusted by the following formula:

[0022] ;

[0023] where is the original local average link distance.

[0024] Further, the data point COF outlier factor is evaluated by the optimized average link distance to obtain the outlier factor of the th optimized data point. After linearly normalizing the optimized outlier factor , the anomaly degree value is obtained, and the abnormal data is screened according to the preset threshold .

[0025] Further, the trend fitting model is:

[0026] ;

[0027] where , , are fitted by the least squares method, is the timestamp.

[0028] Compared with the prior art, the present invention has the following advantages:

[0029] The monitoring data optimization method for saline-alkali land improvement equipment according to the present invention can effectively identify the data points collected by the saline-alkali land improvement equipment at the junction of different salinity regions through the evaluation of the junction degree of the detection data points, avoiding the misjudgment of abnormal situations caused by the normal drastic change of the salinity concentration at the spatial junction in the traditional COF algorithm, thereby significantly improving the spatial adaptability of the abnormal data identification of the saline-alkali land improvement equipment.

[0030] By analyzing the change trend of the salinity concentration of the historical data at the sampling position, the matching evaluation of the current data point and the historical trend is carried out, so as to effectively avoid misjudging the data that conforms to the normal improvement trend of the saline-alkali land as abnormal, further improving the accuracy and reliability of the abnormal detection, and ensuring that the data acquisition result is more in line with the actual saline-alkali land improvement process. Brief Description of the Drawings

[0031] The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0032] Figure 1 is a flowchart of the monitoring data optimization method for the saline-alkali land improvement equipment according to the embodiment of the present invention; Detailed Embodiments

[0033] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0034] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "back", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation to the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0035] The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0036] Embodiment 1

[0037] This embodiment relates to a monitoring data optimization method for saline-alkali land improvement equipment, as Figure 1 shown, including the following steps:

[0038] S1: Collect salt concentration data through a saline-alkali land improvement device, set the travel path and data collection frequency of the saline-alkali land improvement device, conduct saline-alkali land salt data collection every time it travels a preset distance, and record the position information, time information of the data and the monitoring data of the salt concentration sensor;

[0039] S2: Optimize the anomaly detection of the collected salt concentration data;

[0040] The step S2 includes:

[0041] S2.1: Calculate the boundary degree factor of the data point through a non-linear mapping function based on the mean and standard deviation of the salt concentration in the spatial neighborhood historical data , and identify the data points at the junction of different salt regions;

[0042] S2.2: Establish a trend fitting model for the historical data , based on the trend fitting model of the historical data , combine the residual between the current data point and the historical trend and the time interval sensitivity, and calculate the anomaly degree optimization factor ;

[0043] S2.3: By comprehensively considering the boundary degree factor and the anomaly degree optimization factor , adjust the local average linkage distance to obtain the optimized local average linkage distance , optimize the COF anomaly detection algorithm, and screen out the abnormal data;

[0044] S3: Store and upload the screened valid data.

[0045] Specifically, in step S1, set the travel path of the saline-alkali land improvement device and set the data collection frequency. In this embodiment, it is set that the saline-alkali land improvement device conducts saline-alkali land salt data collection every 20 meters. For each piece of saline-alkali land salt data collected, record the position information, time information, and monitoring data of the salt concentration sensor of this data. The position information is the coordinate data of the sampling position, and the time information is the time when the data is collected.

[0046] In step S2, first analyze the change pattern of the salt concentration detection data to obtain the boundary degree of the monitoring data ; then analyze the historical change trend of the monitoring data of the saline-alkali land improvement device through the boundary degree of the monitoring data , so as to obtain the anomaly degree optimization factor through trend consistency analysis for the data points with a higher boundary degree ; finally, through the boundary degree of the salt concentration data points and the anomaly degree optimization factor The COF anomaly detection process of data points is optimized to obtain the optimized anomaly degree of each data point, and the salt concentration detection data of saline-alkali land is screened by setting the anomaly degree threshold, and the abnormal data is eliminated, thereby ensuring that the saline-alkali land improvement equipment can accurately monitor the data of the saline-alkali land improvement effect.

[0047] In this embodiment, during the data collection process of the saline-alkali land improvement equipment, when the data collected by the equipment exceeds When the COF anomaly detection method is used, the anomaly detection of the newly added data points can be started. Before the anomaly detection of the data points, the change pattern analysis is first performed through the spatial local historical monitoring data of the monitoring data points to identify the changes in each monitoring data point during the historical detection process. For the historical data of the saline-alkali land improvement process, the spatial location of the monitoring data point that is less than 100 km from the location is selected. All historical data points are taken as neighboring area data points, and the degree of intersection of the data point is evaluated based on the monitoring numerical analysis of the neighboring area data points.

[0048] First test of saline-alkali land improvement equipment data points, the calculation formula for the intersection degree of the data point is:

[0049] ;

[0050] in, : It is the sensitivity adjustment coefficient, which is used to adjust the sensitivity of the algorithm to local data differences.

[0051] :The first equipment for improving saline-alkali land The salt concentration value of each detection data point.

[0052] :For the The detection data point is the center, and the spatial neighborhood radius The average salt concentration of the historical detection data within the area is calculated to reflect the overall level of the neighborhood data.

[0053] : is the standard deviation of the salt concentration of historical data in the spatial neighborhood, reflecting the fluctuation range of the neighborhood salinity value.

[0054] : is a natural constant.

[0055] :for Function is used to measure the relative relationship between the salinity fluctuation characteristics of the local area and the fluctuation characteristics of the overall monitoring area.

[0056] : It is the historical standard deviation of salt concentration in the entire monitoring area, which is used to reflect the background level of global salt fluctuations.

[0057] : is a very small positive number, set to , to prevent the denominator from being .

[0058] Based on the above, first, by evaluating the current data point The historical average salinity of the spatial neighborhood The difference between To indicate the significant difference between the current data point and the surrounding historical data in space. In order to evaluate this difference more flexibly, use The function maps the difference values ​​to form , which allows the magnitude of the difference to be emphasized in a nonlinear form and stabilized within a limited range when the data point differs greatly from the local historical data.

[0059] On the other hand, considering that the salt distribution patterns in different areas of saline-alkali land in actual scenarios are different, some areas are long-term stable, while others have drastic salt fluctuations, and these differences are difficult to express through a single local window. The function is used to measure the relative relationship between the local regional salinity fluctuation characteristics and the overall monitoring area fluctuation characteristics: if the historical salinity data fluctuation in the neighborhood is much higher than the global average fluctuation, it indicates that the salinity condition in this area is likely to be complex, with obvious spatial heterogeneity, and is more likely to belong to the boundary area.

[0060] Combining the above two perspectives, namely the difference in the current data points and the difference in historical regional fluctuations relative to the global level, an optimization factor is formed. The final formula is:

[0061] The first half The function maps local data differences stably in the interval (0,2) and smoothly adjusts the sensitivity to local anomalies;

[0062] Second half Based on the comparison of local fluctuations with overall fluctuations, the degree of intersection is further adjusted to ensure that the strength of the intersection judgment is only increased when the historical fluctuations in the region are large.

[0063] Therefore, the optimization factor The construction of the proposed method takes into account the specific details such as the spatial fluctuation characteristics of data in actual saline-alkali land improvement scenarios, historical regional stability differences, and global salinity fluctuation levels, making the boundary degree assessment more accurate and adaptable, thereby effectively reducing the misjudgment of the traditional COF algorithm at the regional boundary position and improving the accuracy of anomaly detection in saline-alkali land monitoring data.

[0064] After obtaining the junction degree of the saline-alkali land improvement equipment, the data points can be screened according to the junction degree of the detection data, so as to evaluate the abnormal degree optimization factor for the data points in the junction area based on the trend consistency of historical changes. For the detection data points in the junction area with different salt distributions in the saline-alkali land, because in the continuous COF anomaly detection process, the local link distance will be relatively large due to different salt distributions, resulting in an increase in the abnormal degree of the data points in the junction area. Therefore, for the data points in the junction area, it is necessary to perform trend fitting through the historical data corresponding to the data points themselves, so as to evaluate the abnormal degree optimization factor through the difference between the newly detected data points and the historical trend fitting. Therefore, in this process, it is first necessary to determine the historical data for fitting, and then perform least squares fitting on the historical data for fitting to obtain the fitting function. Finally, the abnormal degree optimization factor is evaluated through the currently newly detected data points.

[0065] First, determine the historical data window corresponding to the newly detected data points. In this embodiment, the spatial neighborhood historical data set in the above steps is used as the historical data window. For the case of multiple data points corresponding to each time stamp, the average value of the multiple data points is used as the detection value corresponding to the time stamp, so as to ensure that each historical detection time stamp corresponds to only one salt concentration data.

[0066] After determining the data for historical trend fitting, trend fitting can be performed through a regression model. In this embodiment, the fitting trend model is set as where are the regression coefficients fitted by the least squares method respectively, and is the time stamp.

[0067] Determine the polynomial fitting coefficient vector through the least squares method. Fitting the regression coefficients through the least squares method is a prior art, so it will not be elaborated here.

[0068] After obtaining the regression model for trend evaluation, the abnormal degree optimization factor can be evaluated through the difference between the predicted value and the actual measured value of the trend fitting model. For the abnormal degree optimization factor of the th detection data point:

[0069] ;

[0070] where:

[0071] : represents the ​​The salinity concentration values of the detection data points.

[0072] : Represents the natural constant.

[0073] : Represents an extremely small positive number, set to , to prevent the denominator from being .

[0074] : Represents the standard deviation of the historical data fitting residuals.

[0075] : Represents the predicted value at the current detection moment fitted by the regression model. Represents the timestamp of the current data point detection.

[0076] : Represents the time interval between the current data point and the nearest historical data, reflecting the trend stability.

[0077] : Represents the average interval of the historical data sampling time.

[0078] Based on the above, the optimization factor Through the quantitative analysis of the deviation between the historical data trend fitting result and the current measured value, it reflects the consistency between the data point and the historical data trend.

[0079] In particular, two key factors are introduced in the design of this formula:

[0080] Residual sensitivity: Reflects whether the data point significantly deviates from the reasonable change range of the historical trend;

[0081] Time interval sensitivity: Reflects the reliability of the data trend model. The newer the historical sampling data, the higher the reliability of the trend prediction value.

[0082] In this way, it more precisely combines the salinity concentration change trend in the actual saline-alkali land improvement process and the time characteristics of the data monitoring process.

[0083] The anomaly degree optimization factor designed by the present invention , is comprehensively evaluated through two key factors: First, based on the square of the residual between the current data point and the predicted value of the historical trend fitting, a non-linear mapping is performed in the form of a Gaussian exponential function, so that data points with smaller residuals obtain higher values, while data points that significantly deviate from the historical trend obtain lower values. This method effectively quantifies the deviation degree of the newly collected data point from the historical trend; Second, the time interval between the current data point and the nearest historical data collection point is introduced.Perform sensitivity adjustment in terms of time. When the data acquisition time interval is small, it indicates higher reliability of trend prediction, and thus has a higher value. When the interval is large, the weight of trend reliability is moderately reduced, thereby enhancing the timeliness and trend adaptability of the optimization factor.

[0084] Through this dual consideration, this formula can accurately evaluate the degree of consistency between new data points and historical change trends, effectively avoid misjudgment of data at the junction of normal regions, and improve the accuracy and reliability of data anomaly detection for saline-alkali land improvement equipment.

[0085] Comprehensively Perform the evaluation of the optimized local average linkage distance:

[0086] ;

[0087] After obtaining the optimized local average linkage distance, the outlier factor of the data point COF can be evaluated through the optimized average linkage distance, and the outlier factor of the th data point can be obtained .

[0088] After obtaining the outlier factor of the detected data point, the degree of abnormality of the data point is obtained through linear normalization mapping.

[0089] ;

[0090] After obtaining the degree of abnormality of the data point, the screening of abnormal data points can be carried out by setting the abnormality degree threshold .

[0091] The screened data is stored and the screened data is uploaded.

[0092] Through the integration of spatial junction evaluation and historical trend analysis, the present invention solves the problem of misjudgment of the traditional COF algorithm in the saline-alkali land junction area, significantly improves the accuracy of data anomaly detection, and provides efficient and reliable technical support for the precise improvement of saline-alkali land.

[0093] Embodiment 2

[0094] A saline-alkali land improvement equipment includes a data acquisition module for collecting salt concentration data according to a preset path and interval; an anomaly detection optimization module for implementing the anomaly data detection method in Embodiment 1; and a data storage and transmission module for storing and uploading the screened valid data. Specifically, the data acquisition module is integrated with a salt concentration sensor, a GPS positioning device, and a clock module.

[0095] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for optimizing monitoring data of saline-alkali land improvement equipment, characterized in that: It includes the following steps: S1: Collect salt concentration data through the saline-alkali land improvement equipment, set the traveling path and data collection frequency of the saline-alkali land improvement equipment, collect saline-alkali land salt data once every preset distance traveled, and record the position information, time information of the data and the monitoring data of the salt concentration sensor; S2: Optimize the anomaly detection of the collected salt concentration data; The step S2 includes: S2.1: Calculate the boundary degree factor of data points through a non - linear mapping function based on the mean and standard deviation of salt concentration in the spatial neighborhood historical data, and identify the data points at the junction of different salt regions. , and identify the data points at the junction of different salt regions. S2.2: Establish a trend fitting model for historical data , the trend fitting model based on historical data , combine the residuals between the current data point and the historical trend and the time interval sensitivity to calculate the anomaly degree optimization factor ; S2.3: By synthesizing the described junction degree factor and the anomaly degree optimization factor , adjust the local average linkage distance to obtain the optimized local average linkage distance , optimize the COF anomaly detection algorithm, and screen out the abnormal data; S3: Store and upload the filtered valid data.

2. The monitoring data optimization method for the saline-alkali land improvement equipment according to claim 1, characterized in that: In the said step S2.1, the junction degree factor has the following calculation formula: ； Where λ is the sensitivity adjustment coefficient, For the The salt concentration value of each detection data point, For the first The detection data point is the center, and the spatial neighborhood radius The average salt concentration of historical test data within The standard deviation of the salt concentration of historical data in the spatial neighborhood, :for function, To represent the historical standard deviation of salt concentration in the entire monitoring area, is a very small positive number.

3. The monitoring data optimization method for the saline-alkali land improvement equipment according to claim 2, characterized in that: In the step S2.2, the abnormal degree optimization factor is calculated by the following formula: ； Among them, is the salt concentration value of the th detection data point, is a very small positive number, is the predicted value of the fitting trend model, is the standard deviation of the historical data fitting residuals, is the time interval between the current data point and the nearest historical data, and △T is the average sampling interval of the historical data.

4. The monitoring data optimization method for the saline-alkali land improvement equipment according to claim 3, characterized in that: In the step S2.3, the optimized local average link distance is adjusted by the following formula: ； Among them, is the original local average link distance.

5. The monitoring data optimization method for the saline-alkali land improvement equipment according to claim 4, characterized in that: Optimized average link distance Evaluate the COF outlier factor of the data points to obtain the outlier factor of the th optimized data point , linearly normalize the optimized outlier factor to obtain the anomaly degree value , and filter out abnormal data according to the preset threshold .

6. The monitoring data optimization method for the saline-alkali land improvement equipment according to claim 1, characterized in that: The trend fitting model is: ； Among them, , , are determined by least squares fitting, is the timestamp.

Images