Personnel border crossing early warning method and system based on mine electronic fence
By acquiring and processing behavioral data of underground miners, and using classification and electronic fence models for boundary crossing judgment, the problems of inaccurate data collection and poor early warning timeliness in the existing system have been solved. This has enabled accurate monitoring and timely early warning of boundary crossing behavior, thus ensuring the safety of miners.
Patent Information
- Application Number
- CN202511018661.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-23
- Publication Date
- 2025-11-18
AI Technical Summary
Existing underground mine personnel monitoring systems have shortcomings in the data collection stage. They cannot effectively distinguish between legally entering and illegally crossing boundaries. The location data is inaccurate and lacks systematic integration, resulting in poor timeliness of early warnings and difficulty in ensuring personnel safety.
By acquiring behavioral data of personnel in underground mines, noise reduction and data cleaning are performed. A pre-trained classification model is used for classification and anomaly detection. A mine electronic fence model is constructed to detect boundary crossings and issue an early warning when a boundary crossing is detected.
It enables precise monitoring and timely early warning of personnel crossing boundaries, avoiding safety accidents caused by negligence or accidents, protecting the lives of underground mine workers, and reducing the risk of accidents.
Smart Images
Figure CN120977089A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of mine electronic fence early warning, and in particular to a personnel crossing boundary early warning method and system based on a mine electronic fence. BACKGROUND
[0002] The particularity of the underground mine production environment makes personnel safety control always face severe challenges. Underground roadways crisscross, and dangerous areas (such as goaf, high gas area, blasting warning area, etc.) are complexly distributed. Once personnel cross the boundary in violation of the rules, not only the life safety of the personnel may be directly threatened, but also a chain reaction may be triggered, leading to a larger range of production accidents.
[0003] At present, the underground mine personnel monitoring system has obvious short boards in the data acquisition link. Most systems only rely on single positioning data for personnel position judgment, ignoring the personnel's basic information and activity trajectory characteristics. For example, some systems can only obtain the approximate position coordinates of personnel, but cannot associate the personnel's job type, operation permission, current task, and other basic data, making it difficult to distinguish between work personnel who have entered a specific area legally and personnel who have crossed the boundary in violation of the rules, and prone to misjudgment. In terms of real-time positioning data, traditional positioning technologies such as Bluetooth beacon and ZigBee are severely interfered by the underground environment, and the positioning data often has problems such as jumping and delay, and the data dimension is single, only containing position information, and cannot reflect the moving trend of personnel. The lack of activity trajectory data makes it difficult for the system to perform backtracking analysis and prediction on the moving path of personnel. When personnel approach the boundary, the system cannot determine whether they have the intention to cross the boundary based on their historical trajectory, and can only respond passively after the crossing occurs, greatly reducing the timeliness of the warning.
[0004] At the same time, the data acquisition of the existing system lacks systematic integration. Basic data, positioning data, and trajectory data are often stored in different subsystems, the data formats are not unified, and it is difficult to achieve linked analysis, which may not only issue unnecessary warnings to personnel who have permission to enter, but also may miss the crossing behavior of personnel who do not have permission.
[0005] Therefore, there is an urgent need for a personnel crossing boundary early warning method and system based on a mine electronic fence to solve the above technical problems. SUMMARY
[0006] The present application aims to at least partially solve one of the above technical problems. To this end, the first aspect of the present application aims to propose a personnel crossing boundary early warning method based on a mine electronic fence, which can timely discover whether personnel have entered a dangerous area through monitoring and crossing judgment of personnel behavior data, and management personnel can timely manage and guide the crossing personnel according to the warning information, thereby maximizing the protection of the life safety of personnel in underground production mines.
[0007] The second aspect of the present application aims to provide a personnel crossing border early warning system based on mine electronic fence.
[0008] To achieve the above-mentioned purpose, the first aspect of the present application provides a personnel crossing border early warning method based on mine electronic fence, comprising:
[0009] obtaining behavior data of underground production mine personnel;
[0010] preprocessing the behavior data to obtain preprocessed behavior data;
[0011] inputting the preprocessed behavior data into a mine electronic fence model for border crossing judgment;
[0012] when it is determined that the underground production mine personnel crosses the border, issuing a warning prompt.
[0013] Preferably, the behavior data of the underground production mine personnel includes: basic data of the underground production mine personnel, real-time positioning data and activity trajectory data.
[0014] Preferably, preprocessing the behavior data to obtain preprocessed behavior data comprises:
[0015] de-noising the behavior data to obtain de-noised behavior data;
[0016] cleaning the de-noised behavior data to obtain preprocessed behavior data.
[0017] Preferably, cleaning the de-noised behavior data to obtain preprocessed behavior data comprises:
[0018] inputting the de-noised behavior data into a pre-trained classification model for classification to obtain a plurality of data sets of different categories;
[0019] selecting a data set of a category as a first data set;
[0020] sorting the data points in the first data set based on time sequence to obtain a plurality of sequence data;
[0021] selecting a sequence data as a first sequence data;
[0022] identifying the data points in the first sequence data to obtain a plurality of target data points;
[0023] dividing the first sequence data based on the plurality of target data points to obtain a plurality of second sequence data; wherein each second sequence data contains a target data point;
[0024] Taking any second sequence data as target sequence data, and calculating abnormal possibility index and local fluctuation index corresponding to the target sequence;
[0025] Based on the abnormal possibility index and the local fluctuation index of the target sequence data, determining an abnormal degree value of the target sequence data;
[0026] Comparing the abnormal degree value with a preset abnormal degree threshold value, and determining the target sequence data as to-be-cleaned data when the abnormal degree value is greater than or equal to the preset abnormal degree threshold value;
[0027] Iterating through all second sequences to obtain a plurality of to-be-cleaned data;
[0028] Obtaining a preset data cleaning strategy;
[0029] Based on the preset data cleaning strategy, cleaning the plurality of to-be-cleaned data to obtain preprocessed first sequence data;
[0030] Iterating through all sequence data to obtain a preprocessed first data set;
[0031] Iterating through all categories of data sets to obtain preprocessed behavior data.
[0032] Preferably, the data points in the first sequence data are identified to obtain a plurality of target data points, including:
[0033] Calculating the difference between the data values of any two adjacent data points in the first sequence to obtain a plurality of differences; comparing the plurality of differences with a preset difference threshold value, and marking the two data points corresponding to the difference greater than or equal to the preset difference threshold value as abnormal; and counting the number of abnormal marks of each data point;
[0034] Taking the data point corresponding to the number of abnormal marks greater than or equal to a preset abnormal mark threshold value as the target data point;
[0035] Iterating through all data points in the first sequence data to obtain a plurality of target data points.
[0036] Preferably, the abnormal possibility index corresponding to the target sequence is calculated, including:
[0037] Calculating the similarity between the target sequence data and other second sequence data except the target sequence data to obtain a plurality of similarities; and calculating the mean of the plurality of similarities as a first evaluation index;
[0038] Calculating the standard deviation of the data values corresponding to all data points in the target sequence data as a second evaluation index;
[0039] Based on the first evaluation index and the second evaluation index, the abnormal possibility index of the target sequence data is determined.
[0040] Preferably, the local fluctuation index corresponding to the target sequence is calculated, including:
[0041] The data values of each data point in the target sequence data are obtained, and the maximum data value, the minimum data value and the data mean value of the data points in the target sequence data are determined;
[0042] Based on the maximum data value, the minimum data value and the data mean value, the local fluctuation index of the target sequence data is determined.
[0043] Preferably, the method for constructing the mine electronic fence model comprises:
[0044] An initial mine electronic fence model is constructed;
[0045] A mine electronic fence training data set is obtained;
[0046] The initial mine electronic fence model is iteratively trained based on the mine electronic fence training data set to obtain a target mine electronic fence model;
[0047] A mine electronic fence test data set is obtained;
[0048] The target mine electronic fence model is tested based on the mine electronic fence test data set, and when it is determined that the test result meets the requirements, a trained mine electronic fence model is obtained.
[0049] Preferably, the initial mine electronic fence model is constructed, including:
[0050] The basic data of the mine is obtained, and the basic data is standardized to obtain standardized data; the basic data includes mining right coordinates, elevation data, mine drawings and safety prohibited coordinates;
[0051] The mining right coordinates are converted into features based on the ArcGIS tool to determine the mining right plane boundary;
[0052] The mining right plane boundary is calibrated based on satellite image registration and DEM terrain correction to obtain a mining plane fence;
[0053] The preset buffer area data is obtained;
[0054] The mining plane fence is set with a buffer area based on the preset buffer area data to determine the buffer area corresponding to the mining plane fence;
[0055] The mining plane fence is stretched according to the elevation data based on the stretching tool to determine the mine cubic model;
[0056] Acquire mine collection roadway data and construct a mine collection roadway CAD model;
[0057] Convert the mine collection roadway CAD model into three dimensions to generate a roadway three-dimensional pipeline model;
[0058] Acquire geological exploration data; generate a goaf voxel model based on the geological exploration data and a voxelization tool;
[0059] Fuse the mine cubic model, the roadway three-dimensional pipeline model and the goaf voxel model to generate an initial mine electronic fence model.
[0060] To achieve the above purpose, the second aspect of the present application provides a personnel out-of-bound early warning system based on a mine electronic fence, comprising:
[0061] An acquisition module is configured to acquire behavior data of underground production mine personnel;
[0062] A preprocessing module is configured to preprocess the behavior data to obtain preprocessed behavior data;
[0063] A judgment module is configured to input the preprocessed behavior data into a mine electronic fence model for out-of-bound judgment;
[0064] An early warning module is configured to issue an early warning prompt when it is determined that the underground production mine personnel are out of bounds.
[0065] The present application provides a personnel out-of-bound early warning method and system based on a mine electronic fence, which can accurately monitor whether personnel enter a dangerous area and timely issue an early warning by acquiring behavior data of underground production mine personnel and making out-of-bound judgment based on a mine electronic fence model, so as to avoid personnel from encountering danger due to negligence or accident and thus protect the life safety of underground production mine personnel; through timely early warning, a series of safety accidents caused by personnel after out-of-bound can be prevented, such as destroying a dangerous operation environment in progress or contacting unstable geological structure to cause collapse and other accidents, thereby effectively reducing the accident risk in mine production.
[0066] Other features and advantages of the present application will be further described in the following specification, and some will become apparent from the specification, or will be learned from the practice of the present application. The purposes and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the written specification and the accompanying drawings.
[0067] The technical solutions of the present application will be further described in detail below with the help of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS
[0068] The accompanying drawings, which are included to provide a further understanding of the application and constitute a part of this specification, illustrate the application and together with the description help to explain the application. In the drawings:
[0069] Figure 1 is a flow chart of a personnel border crossing early warning method based on a mine electronic fence according to an embodiment of the application;
[0070] Figure 2 is a flow chart of a mine electronic fence model construction method according to an embodiment of the application;
[0071] Figure 3 is a block diagram of a personnel border crossing early warning system based on a mine electronic fence according to an embodiment of the application. DETAILED DESCRIPTION
[0072] The preferred embodiments of the application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described here are only used to explain and illustrate the application, and are not used to limit the application.
[0073] Embodiment 1: As shown in a personnel border crossing early warning method based on a mine electronic fence, comprising steps S1-S4: Figure 1
[0074] S1: Obtain behavior data of underground production mine personnel;
[0075] S2: Preprocess the behavior data to obtain preprocessed behavior data;
[0076] S3: Input the preprocessed behavior data into a mine electronic fence model for border crossing judgment;
[0077] S4: When it is determined that the underground production mine personnel cross the border, an early warning prompt is issued.
[0078] The working principle and beneficial effects of the above technical solution are: by obtaining behavior data of underground production mine personnel and making border crossing judgment based on a mine electronic fence model, the personnel can be accurately monitored whether to enter a dangerous area, and an early warning can be timely issued to avoid personnel from encountering danger due to negligence or accident, thereby protecting the life safety of underground production mine personnel; through timely early warning, a series of safety accidents caused by personnel crossing the border can be prevented, such as damaging the ongoing dangerous operation environment or contacting unstable geological structure to cause collapse and other accidents, and the accident risk in mine production is effectively reduced.
[0079] Embodiment 2: The behavior data of underground production mine personnel includes: basic data, real-time positioning data and activity trajectory data of underground production mine personnel.
[0080] Embodiment 3: preprocessing the behavior data to obtain preprocessed behavior data, including:
[0081] de-noising the behavior data to obtain de-noised behavior data;
[0082] cleaning the de-noised behavior data to obtain the preprocessed behavior data.
[0083] The working principle and beneficial effects of the above technical solution are: in a mine environment, the personnel behavior data collected by the sensor may be affected by various interference factors, and the behavior data is de-noised to eliminate the noise data caused by these interference factors; on the basis of de-noising, data cleaning is performed so as to be subsequently input into a mine electronic fence model for out-of-boundary judgment; through de-noising and data cleaning, impurities in the behavior data can be removed, so that the data more accurately reflects the actual behavior of underground production mine personnel, and accurate behavior data helps the mine electronic fence model to more accurately judge whether the personnel out-of-boundary and the specific situation of out-of-boundary.
[0084] Embodiment 4: cleaning the de-noised behavior data to obtain preprocessed behavior data, including:
[0085] inputting the de-noised behavior data into a pre-trained classification model for classification to obtain a plurality of data sets of different categories;
[0086] taking a data set of a category as a first data set;
[0087] sorting data points in the first data set based on time series to obtain a plurality of sequence data;
[0088] taking one sequence data as a first sequence data;
[0089] identifying data points in the first sequence data to obtain a plurality of target data points;
[0090] dividing the first sequence data based on the plurality of target data points to obtain a plurality of second sequence data; each second sequence data contains a target data point;
[0091] taking one second sequence data as a target sequence data, and calculating an abnormality likelihood index and a local fluctuation index corresponding to the target sequence;
[0092] determining an abnormality degree value of the target sequence data based on the abnormality likelihood index and the local fluctuation index of the target sequence data;
[0093] The abnormality degree value is compared with a preset abnormality degree threshold value, and when it is determined that the abnormality degree value is greater than or equal to the preset abnormality degree threshold value, the target sequence data is taken as to-be-cleaned data.
[0094] All the second sequences are traversed to obtain a plurality of to-be-cleaned data.
[0095] A preset data cleaning strategy is obtained.
[0096] The plurality of to-be-cleaned data is subjected to data cleaning based on the preset data cleaning strategy to obtain preprocessed first sequence data.
[0097] All the sequence data are traversed to obtain a preprocessed first data set.
[0098] All the data sets of the categories are traversed to obtain preprocessed behavior data.
[0099] In this embodiment, the first sequence data is divided based on a plurality of target data points to obtain a plurality of second sequence data. The principle is that when the sequence data is divided, the abnormal data points are ensured to be in the middle part of each second sequence data. In this way, the subdivided sequence data helps to improve the accuracy of the secondary identification of the second sequence data in the later stage. Unlike the prior art, which divides based on the position of the abnormal data points, it is difficult to accurately identify the abnormality degree of the abnormal data points and the overall influence degree of the abnormal data points relative to the data segment, thereby resulting in poor data cleaning effect.
[0100] In this embodiment, the abnormality possibility index and the local fluctuation index corresponding to the target sequence data are mainly investigated from the similarity between the target sequence data and other sequence data and the internal of the target sequence data, and the target data points in the target sequence data are evaluated from multiple dimensions to improve the accuracy of the evaluation.
[0101] In this embodiment, the preset data cleaning strategy is constructed based on experimental data, and different data cleaning strategies can be used for different types of data.
[0102] The working principle and beneficial effects of the above technical solution are: first, the noise-reduced behavior data is input into a pre-trained classification model for classification. This classification model is trained based on a large amount of existing behavior data and can divide the data into several different categories of data sets according to different characteristics or patterns; for a randomly selected category data set (first data set), the data points in it are sorted based on time sequence to form several sequence data. Then, a randomly selected one (first sequence data) is selected from these sequence data, and the data points in it are identified to find several target data points; then, based on these target data points, the first sequence data is divided to obtain several second sequence data containing only one target data point; for a randomly selected one (target sequence data), the corresponding abnormal possibility index and local fluctuation index are calculated; by traversing all the second sequence data, several to-be-cleaned data are obtained. A preset data cleaning strategy is obtained, which may include correction of abnormal data points.
[0103] Embodiment 5: identifying the data points in the first sequence data to obtain several target data points, comprising:
[0104] calculating the difference between the data values of any two adjacent data points in the first sequence to obtain several difference values; comparing the several difference values with a preset difference value threshold, and marking the two data points corresponding to the difference value greater than or equal to the preset difference value threshold as abnormal; and counting the number of abnormal marks for each data point;
[0105] the data point corresponding to the number of abnormal marks greater than or equal to a preset abnormal mark threshold is taken as a target data point;
[0106] traversing all the data points in the first sequence data to obtain several target data points.
[0107] The working principle and beneficial effects of the above technical solution are: in a normal behavior data sequence, the data value change between adjacent data points is usually relatively smooth, fluctuating within a certain range; if the difference between adjacent data points is too large, it may imply that the data is abnormal; then compare these differences with the preset difference threshold value, when the difference is greater than or equal to the preset difference threshold value, mark the corresponding two data points as abnormal; this preset difference threshold value is set in advance according to historical data, statistical characteristics of data or domain knowledge, which defines the range of normal data change; count the number of abnormal marks for each data point; since a data point may be compared with multiple adjacent data points, there may be multiple abnormal marks; when the number of abnormal marks of a data point is greater than or equal to the preset abnormal mark threshold value, the data point is taken as the target data point; by analyzing the difference between adjacent data points to identify abnormalities, local abnormal changes in the data sequence can be well captured; by identifying and screening target data points, these abnormal data points can be specially processed in subsequent data processing steps, thereby improving the overall quality of the data; high-quality data is very critical for mine electronic fence model to accurately determine personnel behavior out-of-bound situations and other applications, reducing the risk of misjudgment caused by abnormal data interference.
[0108] Embodiment 6: Calculate the abnormal possibility index corresponding to the target sequence, comprising:
[0109] Calculate the similarity between the target sequence data and other second sequence data except the target sequence data, to obtain a plurality of similarities; calculate the mean of the plurality of similarities as a first evaluation index;
[0110] Calculate the standard deviation of the data values corresponding to all data points in the target sequence data as a second evaluation index;
[0111] Determine the abnormal possibility index of the target sequence data based on the first evaluation index and the second evaluation index.
[0112] In this embodiment, the ratio of the second evaluation index to the first evaluation index is taken as the abnormal possibility index of the target sequence data; the smaller the first evaluation index value, the lower the approximation degree, the greater the abnormal possibility; the greater the second evaluation index, the greater the fluctuation of the data values of the data points in the target sequence data, the greater the abnormal possibility.
[0113] The working principle and beneficial effects of the above technical solution are: by comparing the similarity of the target sequence with other sequences, the position of the target sequence in the overall data distribution can be understood, and after calculating a plurality of similarities, the mean of the similarities is calculated as the first evaluation index; then the standard deviation of the data values corresponding to all data points in the target sequence data is calculated. The standard deviation is a statistical quantity for measuring the degree of data dispersion; finally, the abnormal possibility index of the target sequence data is determined based on the first evaluation index and the second evaluation index; by calculating the first evaluation index (based on the similarity with other sequences) and the second evaluation index (based on the standard deviation of the data points), the abnormal possibility index is determined, which comprehensively considers the relationship between the target sequence and other sequences and the dispersion degree of the data itself; this multi-factor comprehensive evaluation can more comprehensively and accurately measure the abnormal possibility of the target sequence data, rather than relying on a single feature; for different types of behavior data, the data distribution and fluctuation characteristics may be different, both the relative position of the target sequence in the overall data and the data change within itself are considered, thereby improving the quality of the overall data.
[0114] Embodiment 7: calculating the local fluctuation index corresponding to the target sequence, comprising:
[0115] obtaining the data values of each data point in the target sequence data, determining the maximum data value, the minimum data value and the data mean value of the data points in the target sequence data;
[0116] determining the local fluctuation index of the target sequence data based on the maximum data value, the minimum data value and the data mean value.
[0117] In this embodiment, the difference between the maximum data value and the minimum data value is calculated, and the ratio of the difference to the data mean value is taken as the local fluctuation index of the target sequence data; if the difference between the maximum data value and the minimum data value is relatively large compared to the data mean value, the local fluctuation index will be higher, indicating that the data fluctuates violently in the local range; on the contrary, if the difference is relatively small compared to the data mean value, the local fluctuation index will be lower, indicating that the data fluctuates less.
[0118] The working principle and beneficial effects of the above technical solution are: first, the data values of each data point in the target sequence data are obtained; then the maximum data value, the minimum data value and the data mean value in the target sequence data are determined from these data values; the maximum data value and the minimum data value reflect the boundary conditions of the target sequence data in the local range, i.e. the extreme points of the data; the data mean value is the average level of all data points, which represents the central tendency of the target sequence data; the local fluctuation index of the target sequence data is determined based on the maximum data value, the minimum data value and the data mean value, which is different from some global fluctuation indexes, and it pays more attention to the fluctuation between the internal data points of the target sequence. In many cases, local fluctuation is very important for understanding the characteristics of data.
[0119] Embodiment 8: As shown in the figure, the method for constructing a mine electronic fence model comprises steps S31-S35: Figure 2
[0120] S31: Construct an initial mine electronic fence model;
[0121] S32: Obtain a mine electronic fence training data set;
[0122] S33: Based on the mine electronic fence training data set, iteratively train the initial mine electronic fence model to obtain a target mine electronic fence model;
[0123] S34: Obtain a mine electronic fence test data set;
[0124] S35: Based on the mine electronic fence test data set, test the target mine electronic fence model, and when it is determined that the test result meets the requirements, obtain a trained mine electronic fence model.
[0125] Embodiment 9: Constructing an initial mine electronic fence model comprises:
[0126] Obtain the basic data of the mine and perform standardization processing on the basic data to obtain standardized data; the basic data includes mining right coordinates, elevation data, mine drawings, and safety prohibited coordinates;
[0127] Based on the ArcGIS tool, the mining right coordinates are converted to the surface, and the mining right plane boundary is determined;
[0128] Based on satellite image registration and DEM terrain correction, the mining right plane boundary is calibrated to obtain a mining plane fence;
[0129] Obtain preset buffer area data;
[0130] Based on the preset buffer area data, set the buffer area of the mining plane fence to determine the corresponding buffer area of the mining plane fence;
[0131] Based on the stretching tool, the mining plane fence is stretched according to the elevation data to determine the mine cubic model;
[0132] Obtain the mine collection roadway data and construct a mine collection roadway CAD model;
[0133] The mine collection roadway CAD model is converted into three dimensions to generate a roadway three-dimensional pipeline model;
[0134] Obtain the geological exploration data; based on the geological exploration data and the voxel tool, generate a goaf voxel model;
[0135] The mine cube model, the roadway three-dimensional pipeline model and the goaf voxel model are fused to generate an initial mine electronic fence model.
[0136] In this embodiment, the mine right coordinate represents the plane range of the mine electronic fence; and the elevation data represents the vertical range of the mine electronic fence.
[0137] In this embodiment, the ArcGIS product line provides a scalable and comprehensive GIS platform for users, arranges the mine right turning point coordinates into a table format, imports the coordinates into ArcGIS, opens ArcGIS Pro, clicks the "Map" tab, selects the coordinate table, right-clicks the table, selects "Display XY Data", opens the "Geoprocessing" panel, searches for the tool point set to line, creates a boundary line, searches for the tool feature to surface, clicks "Run" to generate a plane boundary polygon, and completes the line feature to surface.
[0138] In this embodiment, the satellite image registration can accurately correspond the plane boundary of the mine to the actual geographical image, and the DEM terrain correction considers the influence of the terrain factor on the plane boundary, so that a more accurate mine plane fence is obtained.
[0139] The working principle and beneficial effects of the technical scheme are as follows: firstly, the basic data of the mine are acquired, including mining right coordinates, elevation data, mine drawings and safety prohibited coordinates and the like. The basic data are subjected to standardized processing, so that the data of different sources and different magnitudes are comparable. The mining right coordinates are subjected to element surface conversion operation through the ArcGIS tool, which is the process of converting discrete coordinate points into planar boundary of the mining right in a surface shape; the ownership range of the mine on the plane can be determined; the planar boundary of the mining right is calibrated based on satellite image registration and DEM topographic correction; the satellite image registration can make the planar boundary of the mine accurately correspond to the actual geographical image, and the DEM topographic correction considers the influence of the topographic factor on the planar boundary, so that a more accurate mining planar fence is obtained; then, preset buffer area data are acquired, and a buffer area corresponding to the mining planar fence is determined according to the data. The setting of the buffer area can take into account some special areas around the mine, such as areas that may be affected by mining, and the like; the mining planar fence is stretched according to the elevation data by using a stretching tool, so that a mine cubic model with a three-dimensional space concept is constructed from the planar mining right range, and the model reflects the basic shape and range of the mine in the three-dimensional space; mine collection roadway data are acquired and a mine collection roadway CAD model is constructed, and then the CAD model is subjected to three-dimensional conversion to generate a three-dimensional pipeline model of the roadway, and this process is the process of converting two-dimensional roadway design data into a roadway model conforming to the three-dimensional space. Meanwhile, geological exploration data are acquired, and a goaf voxel model is generated based on a voxelization tool, and the goaf voxel model can accurately represent the goaf area formed by the mining completed in the mine. Finally, the mine cubic model, the three-dimensional pipeline model of the roadway and the goaf voxel model are fused to generate an initial mine electronic fence model; in the construction process, the influence of geographical factors on the mine model is considered through satellite image registration, DEM topographic correction and the like. This makes the model not only accurate on the plane, but also in conformity with the actual topography in the three-dimensional space, and improves the accuracy and practicability of the model.
[0140] To achieve the above object, the second aspect of the present application proposes a personnel crossing boundary early warning system based on mine electronic fence, comprising:
[0141] The acquisition module is configured to acquire behavior data of the personnel in the underground production mine.
[0142] The preprocessing module is configured to preprocess the behavior data to obtain preprocessed behavior data.
[0143] The judgment module is configured to input the preprocessed behavior data into the mine electronic fence model for boundary crossing judgment.
[0144] The warning module is configured to issue a warning prompt when it is determined that the personnel in the underground production mine crosses the boundary.
[0145] The working principle and beneficial effects of the technical solution are as follows: by acquiring the behavior data of the underground production mine personnel and making the out-of-bound judgment based on the mine electronic fence model, it can accurately monitor whether the personnel enter the dangerous area, timely issue a warning, avoid the personnel from encountering danger due to negligence or accident, and thus protect the life safety of the underground production mine personnel; through the timely warning, a series of safety accidents caused by the personnel after crossing the boundary can be prevented, such as destroying the dangerous operation environment being carried out or contacting the unstable geological structure to cause collapse and other accidents, and the accident risk in mine production is effectively reduced.
[0146] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the claims of the present application and the equivalent technologies thereof, the present application also intends to include these modifications and variations.
Claims
1. A method for early warning of personnel crossing boundaries based on electronic fences in mines, characterized in that, include: Obtain behavioral data on personnel in underground mines; The behavioral data is preprocessed to obtain preprocessed behavioral data; The preprocessed behavioral data is input into the mine electronic fence model for boundary crossing judgment; When it is determined that personnel in underground mining operations have crossed the boundary, an early warning notice will be issued.
2. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 1, characterized in that, The behavioral data of underground mining personnel includes: basic data, real-time location data, and activity trajectory data of underground mining personnel.
3. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 1, characterized in that, The behavioral data is preprocessed to obtain preprocessed behavioral data, including: The behavioral data is denoised to obtain denoised behavioral data; The denoised behavioral data is cleaned to obtain preprocessed behavioral data.
4. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 3, characterized in that, The denoised behavioral data is cleaned to obtain preprocessed behavioral data, including: The denoised behavioral data is input into a pre-trained classification model for classification, resulting in several datasets of different categories. Choose any dataset from one category as the first dataset; The data points in the first dataset are sorted based on the time series, resulting in several sequence data. Choose any sequence of data as the first sequence of data; The data points in the first sequence of data are identified to obtain several target data points; The first sequence of data is divided based on several target data points to obtain several second sequence data points; each second sequence data point contains one target data point. Take any second sequence data as the target sequence data, and calculate the anomaly probability index and local volatility index corresponding to the target sequence; Based on the anomaly probability index and local volatility index of the target sequence data, the degree of anomaly of the target sequence data is determined. The anomaly level value is compared with a preset anomaly level threshold. When the anomaly level value is determined to be greater than or equal to the preset anomaly level threshold, the target sequence data is taken as data to be cleaned. Traverse all the second sequences to obtain a number of data to be cleaned; Obtain the preset data cleaning strategy; Based on a preset data cleaning strategy, several data to be cleaned are cleaned to obtain the first sequence of preprocessed data. Iterate through all the sequence data to obtain the first preprocessed dataset; Iterate through the datasets of all categories to obtain preprocessed behavioral data.
5. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 4, characterized in that, The data points in the first sequence of data are identified to obtain several target data points, including: Calculate the difference between the data values of any two adjacent data points in the first sequence to obtain several differences; compare each of the several differences with a preset difference threshold, and mark the two data points corresponding to the differences that are greater than or equal to the preset difference threshold as anomalies; count the number of anomaly marks for each data point; The data points corresponding to the number of anomaly markers being greater than or equal to the preset anomaly marker threshold are used as target data points; Traverse all data points in the first sequence of data to obtain several target data points.
6. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 5, characterized in that, Calculate the anomaly probability index corresponding to the target sequence, including: Calculate the similarity between the target sequence data and other second sequence data besides the target sequence data to obtain several similarity scores; calculate the mean of several similarity scores as the first evaluation index. Calculate the standard deviation of the data values corresponding to all data points in the target sequence data, and use it as the second evaluation index; Based on the first and second evaluation indicators, the anomaly probability index of the target sequence data is determined.
7. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 6, characterized in that, Calculate the local volatility index corresponding to the target sequence, including: Obtain the data values of each data point in the target sequence data, and determine the maximum data value, minimum data value, and mean data value of each data point in the target sequence data; Based on the maximum data value, minimum data value, and data mean, the local fluctuation index of the target sequence data is determined.
8. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 1, characterized in that, The method for constructing a mine electronic fence model includes: Construct an initial electronic fence model for the mine; Obtain the training dataset for the mine electronic fence; The initial mine electronic fence model was iteratively trained based on the mine electronic fence training dataset to obtain the target mine electronic fence model. Obtain the test dataset for the mine electronic fence; The target mine electronic fence model is tested based on the mine electronic fence test dataset. When the test results meet the requirements, the trained mine electronic fence model is obtained.
9. The method for early warning of personnel crossing boundaries based on mine electronic fences as described in claim 8, characterized in that, Constructing an initial mine electronic fence model, including: Acquire basic data of the mine and standardize the basic data to obtain standardized data; the basic data includes mining rights coordinates, elevation data, mine maps and safety prohibition coordinates; Based on ArcGIS tools, the coordinates of mining rights are converted into polygons to determine the planar boundaries of mining rights. Based on satellite image registration and DEM terrain correction, the planar boundary of the mining rights is calibrated to obtain the mining planar fence; Retrieve data from the preset buffer area; Based on the preset buffer area data, the buffer area of the mining plane fence is set to determine the buffer zone corresponding to the mining plane fence. The mining plane fence is stretched according to the elevation data using the stretching tool to determine the mining cube model; Acquire data from mine acquisition roadways and construct CAD models of the mine acquisition roadways; The CAD model of the mine tunnel is converted into a 3D model to generate a 3D pipeline model of the tunnel. Acquire geological exploration data; generate voxel models of goaf areas based on geological exploration data and voxelization tools; The initial mine electronic fence model is generated by merging the mine cube model, the roadway 3D pipeline model, and the goaf voxel model.
10. A personnel boundary crossing early warning system applying the personnel boundary crossing early warning method based on a mine electronic fence as described in any one of claims 1-9, characterized in that, include: The acquisition module is used to acquire behavioral data of personnel in underground mines. The preprocessing module is used to preprocess the behavioral data to obtain preprocessed behavioral data; The judgment module is used to input the preprocessed behavioral data into the mine electronic fence model for boundary judgment. The early warning module is used to issue an early warning when it is determined that personnel in underground mining operations have crossed the boundary.