Underground intrusion behavior micro-seismic monitoring method based on DBSCAN clustering

Abstract

The invention discloses an underground intrusion behavior micro-seismic monitoring method based on DBSCAN clustering, which comprises the following steps: screening out all micro-seismic events in micro-seismic data, determining the occurrence time and the occurrence position of each micro-seismic event, analyzing and processing the micro-seismic data by adopting a positive and negative DBSCAN clustering method, sequencing according to the occurrence time of each micro-seismic event, and determining the occurrence position of each micro-seismic event according to the sequence of the occurrence time of each micro-seismic event. A forward clustering method and a reverse clustering method are adopted to obtain two groups of cluster data and noise data respectively, the noise data are distinguished, then the two groups of cluster data are subjected to intersection processing, the noise data are further removed, and finally a microseismic event set generated by mining behaviors during underground intrusion is obtained. According to the occurrence position and the occurrence time of each micro-seismic event in the set, realizing continuous monitoring of underground intrusion excavation positions in different time periods; according to the method, a large monitoring range can be ensured, interference microseismic events can be eliminated, and the excavation position of the underground intrusion behavior can be continuously and accurately positioned.

Description

Technical Field

[0001] The invention relates to a method for microseismic monitoring of underground intrusion and excavation behavior, in particular to a method for microseismic monitoring of underground intrusion and excavation behavior based on DBSCAN clustering. Background Art

[0002] Criminals need to cross the national border illegally for illegal purposes such as smuggling and drug trafficking. However, since there are regular patrols by defense personnel around the border, they cannot pass smoothly. As a result, some criminals choose suitable geological conditions to invade through the border from abroad through underground digging and enter the country, thus avoiding defense personnel and illegally crossing the border.

[0003] In order to detect underground intrusions in a timely manner, detection equipment needs to be set up. Currently, microseismic monitoring technology can be used to monitor excavation responses during underground intrusions. However, microseismic monitoring technology may produce false alarms for other underground activities or vibrations. For example, walking, driving, farming, etc. may generate underground vibrations, thereby triggering the alarm of the microseismic monitoring system. These factors may lead to frequent interfering microseismic events, increasing the difficulty of accurately monitoring and analyzing underground intrusions. Therefore, how to provide a new method that can ensure a larger monitoring range and exclude interfering microseismic events when monitoring underground intrusions, so as to accurately locate the excavation position of underground intrusions, is a research direction of this industry. Summary of the invention

[0004] In view of the problems existing in the above-mentioned prior art, the present invention provides an underground intrusion microseismic monitoring method based on DBSCAN clustering, which can ensure a larger monitoring range when monitoring underground intrusion behavior, eliminate interfering microseismic events, and continuously and accurately locate the excavation position of underground intrusion.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a method for underground intrusion microseismic monitoring based on DBSCAN clustering, the specific steps are:

[0006] Step 1: Deploy the microseismic monitoring system: first determine the boundary line position, one side of the boundary line is the monitoring area, and two rows of seismic detector groups are deployed on the boundary line. Each seismic detector in each seismic detector group is connected to the microseismic monitor to complete the deployment of the microseismic monitoring system;

[0007] Step 2: Microseismic monitoring: Start the microseismic monitoring system to continuously monitor the monitoring area and store the obtained microseismic monitoring data;

[0008] Step 3: Processing microseismic data: The microseismic monitoring host first screens the microseismic monitoring data collected in step 2 to obtain the occurrence time T and the occurrence position coordinates X, Y of all microseismic events, and then uses the forward and reverse DBSCAN clustering analysis method to analyze and process all microseismic events, thereby distinguishing the microseismic events generated during the excavation of underground intrusion behavior, and finally obtains the location set of the distinguished microseismic events, which is the microseismic event set of underground intrusion;

[0009] Step 4: Determine the underground intrusion excavation location at different times: Divide the underground intrusion microseismic event set into multiple time periods in chronological order, and average the X and Y coordinates of all microseismic events in each time period, so as to monitor the underground intrusion excavation location in each time period.

[0010] Furthermore, in step 3, the forward and reverse DBSCAN cluster analysis method is used to analyze and process all microseismic events. The specific process is as follows:

[0011] ① Forward time clustering: In the DBSCAN clustering process, clustering starts from the earliest point of all microseismic events, and neighborhood diffusion only selects points that occur later or slightly earlier than the core point for clustering;

[0012] ② Reverse time clustering: In the DBSCAN clustering process, clustering starts from the point with the latest occurrence time among all microseismic events, and neighborhood diffusion only selects points that occur earlier or slightly later than the core point for clustering;

[0013] ③ Intersection of clustering results: the largest cluster after removing noise from the forward time clustering in step ① is intersected with the largest cluster after removing noise from the reverse time clustering in step ②. The microseismic events at the intersection of the two are the microseismic events generated by excavation during underground intrusion.

[0014] Furthermore, the forward time clustering in step ① is specifically as follows:

[0015] 1) Parameter setting: Set three parameters: spatial neighborhood radius ε s and the temporal neighborhood ε t , and the minimum number of neighborhood points MinPts;

[0016] The spatial neighborhood radius ε s :Defines the neighborhood of a point on the plane (X, Y), that is, the point is the center and the radius is ε s The circular area of

[0017] Time Neighborhood ε t :Defines a time neighborhood range of T. When performing forward clustering, the occurrence time of other events is no earlier than the occurrence time of the microseismic event at this point minus the time neighborhood ε t; In reverse clustering, the occurrence time of other events is no later than the occurrence time of the microseismic event at this point plus the time neighborhood ε t ;

[0018] Minimum number of neighborhood points MinPts: defines the core point in the spatial neighborhood radius ε s The minimum number of points required to simultaneously satisfy the temporal neighborhood conditions;

[0019] 2) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters;

[0020] Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point:

[0021] T α >T E -ε t

[0022] Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E;

[0023] Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point;

[0024] Noise point: a point that is neither a core point nor a boundary point;

[0025] 3) Initial selection:

[0026] Select the earliest point among all microseismic events;

[0027] 4) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point;

[0028] Mark the initial point as visited;

[0029] Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows;

[0030] 5) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster;

[0031] 6) Processing noise points: If the initial point is a noise point, mark it as a noise point;

[0032] 7) Repeat the process: Repeat steps 3) to 6) until all points have been visited;

[0033] 8) Output results: Output all formed clusters and noise points.

[0034] Furthermore, the reverse time clustering in step ② is specifically as follows:

[0035] 1) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters;

[0036] Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point:

[0037] T α <T E +ε t

[0038] Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E;

[0039] Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point;

[0040] Noise point: a point that is neither a core point nor a boundary point;

[0041] 2) Initial selection:

[0042] Select the point with the latest occurrence time among all microseismic events;

[0043] 3) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point;

[0044] Mark the initial point as visited;

[0045] Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows;

[0046] 4) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster;

[0047] 5) Processing noise points: If the initial point is a noise point, mark it as a noise point;

[0048] 6) Repeat the process: Repeat steps 2) to 5) until all points have been visited;

[0049] 7) Output results: Output all formed clusters and noise points.

[0050] The principle of determining the excavation location of underground intrusion behavior is as follows: In terms of microseismicity, the inventors have found that during the underground intrusion excavation process, a large number of microseismic events will be continuously generated near the excavation location, so the microseismic events will present a dense "invasion path" in the monitoring area, and interference events such as walking, driving, and farming will also generate microseismic events as noise data; the traditional DBSCAN method is used to perform cluster analysis on the X and Y parameters of the excavated microseismic events, and the results can remove noise data to a certain extent and extract real microseismic events; the image obtained is as follows Figure 1 As shown in Figure 1, the effect is good from a plane perspective. However, if the time axis of the microseismic occurrence is added, it will be found that this method still retains some interference events. These interference events are very close to the excavation path in plane position (X, Y), and the DBSCAN method that only clusters X and Y cannot eliminate them. Figure 2 If the average of microseismic events within the time period of these interference events is retained, the location of the cavity will be distorted; and the actual excavation of underground intrusion may be discontinuous in time (that is, in order to prevent the underground intrusion from being discovered, the excavation is generally not continuous, but intermittent). If the traditional DBSCAN with time T is used (X, Y, T clustering), the same excavation intrusion path will be clustered into two tunnels, as shown in Figure 3 As shown; based on this, the inventors further studied and obtained a forward and reverse DBSCAN clustering method that can realize temporal and spatial clustering of such temporally disconnected and spatially continuous microseismic events, and ultimately more accurately determine the location of real-time underground intrusion.

[0051] Compared with the prior art, the present invention is based on the inventor's research, firstly screens out all microseismic events in the microseismic data, and determines the occurrence time and location of each microseismic event, then uses the forward and reverse DBSCAN clustering method to analyze and process the microseismic data, sorts them according to the occurrence time of each microseismic event, uses the forward clustering method and the reverse clustering method to obtain two groups of cluster data and noise data respectively, distinguishes the noise data, and then performs intersection processing on the two groups of cluster data to further remove the noise data, thereby finally obtaining a set of microseismic events generated by excavation during underground invasion, and according to the occurrence location and time of each microseismic event in the set, realizes continuous monitoring of the excavation location of underground invasion behavior in different time periods. This method can not only ensure a larger monitoring range, but also exclude interfering microseismic events, and can continuously and accurately locate the excavation location of underground invasion. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a plane diagram of the traditional DBSCAN method before and after XY clustering of microseismic data positions at a certain moment;

[0053] in, Figure 1 a is the plane diagram of the original microseismic data before clustering; Figure 1 b is the plane graph after clustering;

[0054] Figure 2 This is the data graph before and after the XY clustering of microseismic data positions with time axis by the traditional DBSCAN method;

[0055] in, Figure 2 a is the original microseismic data before clustering; Figure 2 b is the data graph after clustering;

[0056] Figure 3 This is the data graph before and after the traditional DBSCAN method clusters the microseismic data by location XY and time T;

[0057] in, Figure 3 a is the original microseismic data before clustering; Figure 3 b is the data graph after clustering;

[0058] Figure 4 It is a data graph after the positive and negative DBSCAN methods of the present invention cluster the microseismic data by position XY and time T;

[0059] in, Figure 4 a is the data graph after forward clustering; Figure 4 b is the data graph after reverse clustering; Figure 4 c is the data graph after the intersection of positive and negative clustering. DETAILED DESCRIPTION

[0060] The present invention will be further described below.

[0061] The specific steps of the present invention are:

[0062] Step 1: Deploy the microseismic monitoring system: first determine the boundary line position, one side of the boundary line is the monitoring area, and two rows of seismic detector groups are deployed on the boundary line. Each seismic detector in each seismic detector group is connected to the microseismic monitor to complete the deployment of the microseismic monitoring system;

[0063] Step 2: Microseismic monitoring: Start the microseismic monitoring system to continuously monitor the monitoring area and store the obtained microseismic monitoring data;

[0064] Step 3: Microseismic data processing: The microseismic monitoring host first screens the microseismic monitoring data collected in step 2 to obtain the occurrence time T and the occurrence position coordinates X, Y of all microseismic events, and then uses the forward and reverse DBSCAN clustering analysis method to analyze and process all microseismic events, so as to distinguish the microseismic events generated during the excavation of underground intrusion behavior, and finally obtains the location set of the distinguished microseismic events, which is the microseismic event set of underground intrusion. The specific process is as follows:

[0065] ① Forward time clustering: In the DBSCAN clustering process, clustering starts from the earliest point of all microseismic events, and neighborhood diffusion only selects points that occur later or slightly earlier than the core point for clustering, specifically:

[0066] 1) Parameter setting: Set three parameters: spatial neighborhood radius ε s and the temporal neighborhood ε t , and the minimum number of neighborhood points MinPts;

[0067] The spatial neighborhood radius ε s :Defines the neighborhood of a point on the plane (X, Y), that is, the point is the center and the radius is ε s The circular area of

[0068] Time Neighborhood ε t :Defines a time neighborhood range of T. When performing forward clustering, the occurrence time of other events is no earlier than the occurrence time of the microseismic event at this point minus the time neighborhood ε t ; In reverse clustering, the occurrence time of other events is no later than the occurrence time of the microseismic event at this point plus the time neighborhood ε t ;

[0069] Minimum number of neighborhood points MinPts: defines the core point in the spatial neighborhood radius ε s The minimum number of points required to simultaneously satisfy the temporal neighborhood conditions;

[0070] 2) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters;

[0071] Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point:

[0072] T α >T E -ε t

[0073] Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E;

[0074] Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point;

[0075] Noise point: a point that is neither a core point nor a boundary point;

[0076] 3) Initial selection:

[0077] Select the earliest point among all microseismic events;

[0078] 4) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point;

[0079] Mark the initial point as visited;

[0080] Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows;

[0081] 5) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster;

[0082] 6) Processing noise points: If the initial point is a noise point, mark it as a noise point;

[0083] 7) Repeat the process: Repeat steps 3) to 6) until all points have been visited;

[0084] 8) Output results: Output all formed clusters and noise points such as Figure 4 As shown in a.

[0085] ② Reverse time clustering: In the DBSCAN clustering process, clustering starts from the point with the latest occurrence time among all microseismic events, and neighborhood diffusion only selects points that occur earlier or slightly later than the core point for clustering, specifically:

[0086] 1) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters;

[0087] Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point:

[0088] T α <T E +ε t

[0089] Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E;

[0090] Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point;

[0091] Noise point: a point that is neither a core point nor a boundary point;

[0092] 2) Initial selection:

[0093] Select the point with the latest occurrence time among all microseismic events;

[0094] 3) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point;

[0095] Mark the initial point as visited;

[0096] Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows;

[0097] 4) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster;

[0098] 5) Processing noise points: If the initial point is a noise point, mark it as a noise point;

[0099] 6) Repeat the process: Repeat steps 2) to 5) until all points have been visited;

[0100] 7) Output results: Output all formed clusters and noise points such as Figure 4 As shown in b.

[0101] ③ Intersection of clustering results: The largest cluster after removing noise from the forward time clustering in step ① is intersected with the largest cluster after removing noise from the reverse time clustering in step ②. The microseismic events at the intersection of the two are the microseismic events generated by excavation during underground intrusion. Figure 4 As shown in c.

[0102] Step 4: Determine the underground intrusion excavation location at different times: Divide the underground intrusion microseismic event set into multiple time periods in chronological order, and average the X and Y coordinates of all microseismic events in each time period, so as to monitor the underground intrusion excavation location in each time period.

[0103] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for microseismic monitoring of underground intrusion behavior based on DBSCAN clustering, characterized in that: The specific steps are: Step 1: Deploy the microseismic monitoring system: first determine the boundary line position, one side of the boundary line is the monitoring area, and two rows of seismic detector groups are deployed on the boundary line. Each seismic detector in each seismic detector group is connected to the microseismic monitor to complete the deployment of the microseismic monitoring system; Step 2: Microseismic monitoring: Start the microseismic monitoring system to continuously monitor the monitoring area and store the obtained microseismic monitoring data; Step 3: Processing microseismic data: The microseismic monitoring host first screens the microseismic monitoring data collected in step 2 to obtain the occurrence time T and the occurrence position coordinates X, Y of all microseismic events, and then uses the forward and reverse DBSCAN clustering analysis method to analyze and process all microseismic events, thereby distinguishing the microseismic events generated during the excavation of underground intrusion behavior, and finally obtains the location set of the distinguished microseismic events, which is the microseismic event set of underground intrusion; Step 4: Determine the underground intrusion excavation location at different times: Divide the underground intrusion microseismic event set into multiple time periods in chronological order, and average the X and Y coordinates of all microseismic events in each time period, so as to monitor the underground intrusion excavation location in each time period.

2. The underground intrusion behavior microseismic monitoring method based on DBSCAN clustering according to claim 1 is characterized in that: In step 3, the forward and reverse DBSCAN cluster analysis method is used to analyze and process all microseismic events. The specific process is as follows: ① Forward time clustering: In the DBSCAN clustering process, clustering starts from the earliest point of all microseismic events, and neighborhood diffusion only selects points that occur later or slightly earlier than the core point for clustering; ② Reverse time clustering: In the DBSCAN clustering process, clustering starts from the point with the latest occurrence time among all microseismic events, and neighborhood diffusion only selects points that occur earlier or slightly later than the core point for clustering; ③ Intersection of clustering results: the largest cluster after removing noise from the forward time clustering in step ① is intersected with the largest cluster after removing noise from the reverse time clustering in step ②. The microseismic events at the intersection of the two are the microseismic events generated by excavation during underground intrusion.

3. The underground intrusion behavior microseismic monitoring method based on DBSCAN clustering according to claim 2 is characterized in that: The forward time clustering in step ① is specifically as follows: 1) Parameter setting: Set three parameters: spatial neighborhood radius ε s and the temporal neighborhood ε t , and the minimum number of neighborhood points MinPts; The spatial neighborhood radius ε s :Defines the neighborhood of a point on the plane (X, Y), that is, the point is the center and the radius is ε s Circular area of Time Neighborhood ε t :Defines a time neighborhood range of T. When performing forward clustering, the occurrence time of other events is no earlier than the occurrence time of the microseismic event at this point minus the time neighborhood ε t ; In reverse clustering, the occurrence time of other events is no later than the occurrence time of the microseismic event at this point plus the time neighborhood ε t ; Minimum number of neighborhood points MinPts: defines the core point in the spatial neighborhood radius ε s The minimum number of points required to simultaneously satisfy the temporal neighborhood conditions; 2) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters; Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point: T α >T E -ε t Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E; Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point; Noise point: a point that is neither a core point nor a boundary point; 3) Initial selection: Select the earliest point among all microseismic events; 4) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point; Mark the initial point as visited; Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows; 5) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster; 6) Processing noise points: If the initial point is a noise point, mark it as a noise point; 7) Repeat the process: Repeat steps 3) to 6) until all points have been visited; 8) Output results: Output all formed clusters and noise points.

4. The underground intrusion behavior microseismic monitoring method based on DBSCAN clustering according to claim 3 is characterized in that: The reverse time clustering in step ② is specifically as follows: 1) Classify data points: Classify data points into core points, boundary points and noise points according to the set parameters; Core point: If the spatial neighborhood radius ε of a point E s If there are at least MinPts microseismic event points that meet the following time neighborhood conditions, then this point is a core point: T α <T E +e t Among them, T α is the occurrence time of a microseismic event point in the spatial domain, T E is the occurrence time of the microseismic event point E; Boundary point: If the spatial neighborhood radius of a point is ε s There are less than MinPts points that meet the above time neighborhood conditions, but the point is within the spatial neighborhood radius ε of other core points. s If the point is within the time range of other core points and meets the time neighborhood conditions of other core points, then the point is a boundary point; Noise point: a point that is neither a core point nor a boundary point; 2) Initial selection: Select the point with the latest occurrence time among all microseismic events; 3) Expanding the cluster: If the initial point is a core point, a new cluster is expanded with this point as the starting point; Mark the initial point as visited; Set the spatial neighborhood radius ε of the initial point s All core points and boundary points that meet the above time neighborhood conditions are added to the cluster; for each core point added to the cluster, step 4) is repeated to expand the cluster until the cluster no longer grows; 4) Marking boundary points: If the initial point is a boundary point, mark it as visited but do not expand the cluster; 5) Processing noise points: If the initial point is a noise point, mark it as a noise point; 6) Repeat the process: Repeat steps 2) to 5) until all points have been visited; 7) Output results: Output all formed clusters and noise points.

Images