An activity rule statistical method based on target historical activity data

Abstract

The application provides a kind of activity rule statistics method based on target historical activity data, belongs to intelligence reconnaissance field, specifically in situation analysis stage.The method includes setting conditions such as target of interest, extracting the batch number, activity time, location and other information of target from target historical activity library;According to target batch number and time, form trajectory, classify and aggregate trajectory, and extract activity route;Grid division is carried out on the area of interest, and target position grid conversion is carried out, the number of target positions in grid area is accumulated and normalized, and the grid area is processed based on threshold to obtain key activity area and general activity area;According to time and date, the target activity frequency is counted to form the target activity frequency rule;The activity route, activity area and activity frequency and other information are manually adjusted.The application enriches the type of activity rule statistics, and improves the efficiency and accuracy of activity rule generation by combining manual intervention and situation comparison.

Description

Technical Field

[0001] This invention relates to a statistical method for analyzing activity patterns based on historical activity data of a target, and studies its application in activity pattern mining. It belongs to the field of intelligence reconnaissance and is specifically in the situation analysis stage. Background Technology

[0002] On the battlefield, commanders must quickly analyze the operational intentions of current enemy targets, accurately assess any anomalies, and promptly issue warnings, conduct real-time tracking, and take appropriate action. Relying solely on manual monitoring of battlefield targets to analyze and predict enemy intentions and detect anomalies is unrealistic and would severely impact battlefield analysis efficiency, delaying responses. Therefore, it is necessary to proactively analyze and statistically study the activity patterns of enemy targets, identifying patterns in their activity routes, key activity areas, and activity frequency. Based on an accurate understanding of enemy target activity patterns, commanders must assess possible enemy routes, activity areas, and activity frequencies. If enemy targets deviate from their known routes or key activity areas, or if their activity frequency exceeds the known frequency patterns for a given period, timely warnings must be issued.

[0003] The historical activity data of the targets currently available contains the activity patterns of enemy targets. By statistically analyzing and mining this historical activity data, and extracting information such as target batch number, time, and location from the real target data, and performing track correlation, regional extraction, and activity frequency statistics on the extracted data, we can generate target activity route patterns, activity area patterns, and activity frequency statistics. This can accurately reflect the real activity situation of the targets and provide support for action prediction and threat early warning. Summary of the Invention

[0004] The purpose of this invention is to provide a statistical method for analyzing the activity patterns of enemy targets based on historical activity data, thereby improving the efficiency and accuracy of understanding enemy target activity patterns on the battlefield and providing support for situational assessment such as battlefield target action prediction and threat early warning.

[0005] This invention can be achieved through the following technical means:

[0006] A method for statistically analyzing activity patterns based on historical activity data of a target includes the following steps:

[0007] Users set targets to follow, historical activity time periods, and areas to follow, and extract the target's batch number, activity time, and location from the target's historical activity database;

[0008] Based on the extracted information, historical trajectory information of each target of interest is formed by time correlation. The historical trajectory information is manually removed and preliminarily classified in combination with the situation map, and reference trajectory centerlines are drawn. Trajectory classification is achieved based on distance thresholds, and target activity routes are extracted and formed through fitting methods.

[0009] The set area of ​​interest is divided into grids according to a preset scale. The target location information is converted into grid information. The number of target locations in each grid is accumulated and normalized. Then, the number of target locations corresponding to the normalized grid is compared with a set threshold to obtain the key activity area and the general activity area.

[0010] The frequency of target activities is counted by time and date, and averaged over a set time period to form a pattern of target activity frequency.

[0011] The activity routes, activity areas, and activity frequency information generated by statistics are manually intervened and adjusted, and activity patterns are preserved.

[0012] The process involves: first, establishing historical trajectory information for each target based on extracted information and time correlation; then, manually removing and initially classifying the historical trajectory information using a situation map; finally, defining reference trajectory centerlines; classifying trajectories based on distance thresholds; and extracting target activity routes using a fitting method. The specific steps include:

[0013] Step 2-1: Based on the extracted target batch number, generate a target historical trajectory set T1 = {t1, t2, ..., t3} based on the activity time and location information. k}, where k is the number of tracks, and the track set is managed and displayed on the situation map;

[0014] Step 2-2: Based on the target historical trajectory set T1, and combined with the situation display results, manually screen the data to remove isolated trajectories that deviate from the trajectory set, forming a new trajectory set T2 = {t1, t, ..., t}. m}, where m is the number of updated tracks, m≤k;

[0015] Steps 2-3: Display the new trajectory set T2 on the situation map, and perform preliminary manual classification according to the trajectory trend and distance to generate a trajectory classification set Z = {z1, z2, ..., z...} n}, where n is the number of categories, n < m;

[0016] Steps 2-4: Based on the classified trajectory set Z, set the category z for each trajectory. i The reference midline, where i∈[1,n], is set as follows: refer to z iThe direction of the inner trajectory is marked on the situation map with a trajectory line that follows the same direction and is located less than a set value away from the center of the set. This ultimately forms the reference trajectory centerline set S = {s1, s2, ..., s}. n};

[0017] Steps 2-5: Construct a new trajectory classification set P = {p1, p2, ..., p...} n}, where n is the number of categories, and each trajectory t in the trajectory set T2 is traversed. i And each centerline s in the set S of reference trajectory centerlines j Maximum distance,

[0018] That is, d ij =Max(Dis(t) i s j ), where 1≤i≤m, 1≤j≤n

[0019] If d ij If the distance is less than the preset distance threshold △D, then t i Classified into trajectory set p j Continue until the traversal is complete;

[0020] Steps 2-6: Fit each subset in the trajectory classification set P to generate the final trajectory route set R = {r1, r2, ..., r...} n}

[0021] The process involves dividing the designated area of ​​interest into a grid according to a preset scale, converting target location information into grid information, accumulating and normalizing the number of target locations within each grid, and then comparing the number of target locations corresponding to the normalized grid with a set threshold to obtain the key activity area and the general activity area. Specifically, this includes the following steps:

[0022] Step 3-1: Divide the designated area of ​​interest into a grid according to the preset map scale, that is, divide the current area into a grid of m rows and n columns.

[0023]

[0024] Step 3-2: Traverse the target information extracted from the target historical activity database and convert the target location information into the corresponding grid coordinates;

[0025] Step 3-3: Calculate the number of target location points within each grid to obtain the grid target location point count matrix:

[0026]

[0027] Steps 3-4: Obtain the maximum value of the number of locations in the matrix:

[0028] c max =Max(c ij ), where 1≤i≤m, 1≤j≤n,

[0029] The matrix of the number of grid target locations is normalized. The normalized matrix of the number of grid target locations is as follows:

[0030]

[0031] Steps 3-4: Set the threshold value △a for key areas l And the general area threshold value △a s , when a ij ≥△a l When set as an important area marker, when △a s ≤a ij <△a l The time is set as a general region flag, where 1≤i≤m, 1≤j≤n;

[0032] Steps 3-5: Set the quantity value of the important area markers to 1, and the other grid values ​​to 0, to obtain the grid quantity matrix A of the key areas. l Set the value of the general region marker to 1 and the values ​​of other grid cells to 0 to obtain the general region grid number matrix A. s ;

[0033] Steps 3-6: Based on the grid quantity matrix A of key areas l And the general region grid number matrix A s The key area and general area are obtained and converted into latitude and longitude coordinates to obtain the extracted key area and general area.

[0034] This invention statistically analyzes the activity patterns of targets based on their historical activity data. It provides statistical results on activity patterns in three dimensions: target activity routes, target activity areas, and target activity frequencies. Based on real target activity data, and combined with situational awareness and human intervention, it solves the problems of complex parameters and low efficiency of commonly used algorithms such as clustering and mining in previous activity pattern statistics. It can accurately and quickly complete the statistical analysis of activity patterns and is applicable to fields such as battlefield situation analysis and judgment. Attached Figure Description

[0035] Figure 1 This is a flowchart of the statistical analysis of the activity patterns of this invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, please refer to the appendix. Figure 1 The present invention will be further described in detail below.

[0037] This invention provides a method for statistical analysis of activity patterns based on historical activity data of a target, comprising the following steps:

[0038] Users set targets to follow, historical activity time periods, and areas to follow, and extract the target's batch number, activity time, and location from the target's historical activity database.

[0039] Based on the extracted information, historical trajectory information for each target of interest is generated by temporal correlation. This historical trajectory information is then manually filtered and preliminarily categorized using a situation map. Reference trajectory centerlines are drawn, and trajectory classification is achieved based on distance thresholds. Finally, target activity routes are extracted using a fitting method. The specific steps include:

[0040] Step 2-1: Based on the extracted target batch number, generate a target historical trajectory set T1 = {t1, t2, ..., t3} based on the activity time and location information. k}, where k is the number of tracks, and the track set is managed and displayed on the situation map;

[0041] Step 2-2: Based on the target historical trajectory set T1, and combined with the situation display results, manually screen the data to remove isolated trajectories that deviate from the trajectory set, forming a new trajectory set T2 = {t1, t, ..., t}. m}, where m is the number of updated tracks, m≤k;

[0042] Steps 2-3: Display the new trajectory set T2 on the situation map, and perform preliminary manual classification according to the trajectory trend and distance to generate a trajectory classification set Z = {z1, z2, ..., z...} n}, where n is the number of categories, n < m;

[0043] Steps 2-4: Based on the classified trajectory set Z, set the category z for each trajectory. i The reference midline, where i∈[1,n], is set as follows: refer to z i The direction of the inner trajectory is marked on the situation map with a trajectory line that follows the same direction and is located less than a set value away from the center of the set. This ultimately forms the reference trajectory centerline set S = {s1, s2, ..., s}. n};

[0044] Steps 2-5: Construct a new trajectory classification set P = {p1, p2, ..., p...} n}, where n is the number of categories, and each trajectory t in the trajectory set T2 is traversed. i And each centerline s in the set S of reference trajectory centerlines j Maximum distance,

[0045] That is, d ij =Max(Dis(t) i sj ), where 1≤i≤m, 1≤j≤n

[0046] If d ij If the distance is less than the preset distance threshold △D, then t i Classified into trajectory set p j Continue until the traversal is complete;

[0047] Steps 2-6: Fit each subset in the trajectory classification set P to generate the final trajectory route set R = {r1, r2, ..., r...} n}

[0048] The designated area of ​​interest is divided into grids according to a preset scale. Target location information is converted into grid information, and the number of target locations within each grid is accumulated and normalized. Then, the number of target locations corresponding to the normalized grid is compared with a set threshold to obtain the key activity area and the general activity area. Specifically, this includes the following steps:

[0049] Step 3-1: Divide the designated area of ​​interest into a grid according to the preset map scale, that is, divide the current area into a grid of m rows and n columns.

[0050]

[0051] Step 3-2: Traverse the target information extracted from the target historical activity database and convert the target location information into the corresponding grid coordinates;

[0052] Step 3-3: Calculate the number of target location points within each grid to obtain the grid target location point count matrix:

[0053]

[0054] Steps 3-4: Obtain the maximum value of the number of locations in the matrix:

[0055] c max =Max(c ij ), where 1≤i≤m, 1≤j≤n,

[0056] The matrix of the number of grid target locations is normalized. The normalized matrix of the number of grid target locations is as follows:

[0057]

[0058] Steps 3-4: Set the threshold value △a for key areas l And the general area threshold value △a s , when a ij ≥△a l When set as an important area marker, when △as ≤a ij <△a l The time is set as a general region flag, where 1≤i≤m, 1≤j≤n;

[0059] Steps 3-5: Set the quantity value of the important area markers to 1, and the other grid values ​​to 0, to obtain the grid quantity matrix A of the key areas. l Set the value of the general region marker to 1 and the values ​​of other grid cells to 0 to obtain the general region grid number matrix A. s ;

[0060] Steps 3-6: Based on the grid quantity matrix A of key areas l And the general region grid number matrix A s The key area and general area are obtained and converted into latitude and longitude coordinates to obtain the extracted key area and general area.

[0061] The frequency of target activities is statistically analyzed by time and date, and then averaged over a set time period to establish a pattern in the frequency of target activities. This includes the following steps:

[0062] Step 4-1: Based on the time of target extraction, count the number of targets by time and date; count by time, that is, count the number of target locations N for each time period from 1:00 to 24:00. t ={n1, ..., n i ,…,n 24}, where i takes values ​​in the range [1, 24]; the statistics are based on the date, specifically the number of target location points N for each day from the 1st to the 31st. d ={n1, ..., n i ,…,n 31}, i takes values ​​in the range [1, 31];

[0063] Step 4-2: Based on the set statistical conditions for the time period, count the number C for each statistical time point (1:00-24:00). t ={c1, ..., c i c 24}, where i takes values ​​in the range [1, 24] and the number of statistical dates (day 1-31) is C. d ={d1, ...,d i , ..., d 31}, i takes values ​​in the range [1, 31];

[0064] Step 4-3: Based on the number N of target location points in each time period t The number of target locations N per day d The number of times at each statistical moment, C tand the number of each statistical date C d Calculate the target activity frequency information, that is, the target activity frequency for each date is... The target activity frequency for each time period is:

[0065] The statistically generated activity routes, activity areas, and activity frequency information are manually intervened and adjusted, and activity patterns are preserved; specifically, this includes the following steps:

[0066] Users can manually intervene and adjust information such as activity routes, activity areas, and activity frequencies based on experience data, and save activity patterns to support the querying and display of activity patterns.

[0067] Step 5-1: Display the statistical trajectory routes, key activity areas, and general activity areas on the situation map. Combine the historical trajectory of the target and the target activity heat map extracted from the database displayed on the situation map with the statistical results, and manually adjust the trajectory route set, key activity areas, and general activity areas.

[0068] Step 5-2: Display the statistical results of the target activity frequency in the form of line graphs, etc. You can combine empirical data to modify the activity frequency values ​​for each time and date.

[0069] Step 5-3: Save the manually confirmed and adjusted trajectory routes, key activity areas, general activity areas, and activity frequencies to generate target activity pattern information such as action routes, activity areas, and activity frequencies.

[0070] This invention proposes a statistical method for analyzing activity patterns based on historical target activity data. It statistically analyzes activity patterns from three dimensions: target activity route, target activity area, and target activity frequency, enriching the types of activity pattern statistics. Furthermore, by combining manual intervention and situational comparison, it improves the efficiency and accuracy of activity pattern generation. This method can be used in battlefield situational analysis to mine enemy target activity patterns, supporting the prediction of enemy behavioral intentions and threat early warning analysis.

Claims

1. An activity rule statistics method based on target historical activity data, characterized in that, The process comprises the following steps: The user sets a target, a historical activity time period, and a target area range, and extracts a target batch number, an activity time, and a location from a target historical activity library; According to the extracted information, historical trajectory information of each target is formed according to time association, the historical trajectory information is manually removed and preliminarily classified in combination with a situation map, a reference trajectory center line is drawn, trajectory classification is realized according to a distance threshold, and a target activity route is extracted and formed through a fitting method; The set target area is divided into grids according to a preset scale, target location information is converted into grid information, the number of target locations in each grid is accumulated and normalized, and then the number of target locations corresponding to the normalized grid is compared with a set threshold to obtain a key activity area and a general activity area; The number of target activities is counted according to time and date respectively, and is averaged according to a set time period to form a target activity frequency rule; The activity route, the activity area, and the activity frequency information formed through statistics are manually intervened and adjusted, and the activity rule is saved; According to the extracted information, historical trajectory information of each target is formed according to time association, the historical trajectory information is manually removed and preliminarily classified in combination with a situation map, a reference trajectory center line is drawn, trajectory classification is realized according to a distance threshold, and a target activity route is extracted and formed through a fitting method; specifically comprising the following steps: Step 2-1, according to the extracted target batch number, generate a target historical trajectory set T1 = {t1, t2, …, t k} based on activity time and location information, manage the number of trajectories k, and display on the situation map; Step 2-2, on the basis of the target historical trajectory set T1, combined with the situation display result, manually screening out isolated and deviated trajectory set trajectories, forming a new trajectory set T2 = {t1, t, …, t m}m is the number of updated tracks, m≤k; Step 2-3, display the new track set T2 on the situation map, and manually preliminarily classify the tracks according to the trends and distances of the tracks to generate a track classification set Z = {z1, z2, …, zn}, where n is the number of classifications, and n < m. n},n is the number of classifications, and n < m. Step 2-4, according to the classified trajectory set Z, set the reference center line of each trajectory classification z i , wherein i∈[1,n], the specific setting method is: reference z i The trajectory of the inner trajectory, draw a trajectory line on the situation diagram which is consistent with the trajectory and the position distance from the set center is less than the set value, and finally form the reference trajectory center line set S={s1,s2,…,s n} Step 2-5, construct a new trajectory classification set P = {p1, p2, …, p n}, n is the number of classifications, respectively traverse each trajectory t i in the trajectory set T2 and the maximum distance of each centerline s j in the reference trajectory centerline set S, i.e. d ij = Max(Dis(t i , s j )), where 1≤i≤m, 1≤j≤n If d ij is smaller than a preset distance threshold ΔD, t i is classified into the trajectory set p j until the traversal is completed. Step 2-6, fit each subset in the set of classified trajectories P, generating the final set of trajectory routes R = {r1, r2,..., r n}.

2. The method of claim 1, wherein, The set target area is divided into grids according to a preset scale, target location information is converted into grid information, the number of target locations in each grid is accumulated and normalized, and then the number of target locations corresponding to the normalized grid is compared with a set threshold to obtain a key activity area and a general activity area; Specifically comprising the following steps: Step 3-1, according to a preset map scale, the set target area is divided into grids, that is, the current area is divided into m rows and n columns of grids, Step 3-2, the target information extracted from the target historical activity library is traversed, and the target location information is converted into corresponding grid coordinates; Step 3-3, the number of target location points in each grid is calculated to obtain a grid target location point number matrix: Step 3-4, the maximum value of the number in the location point number matrix is obtained: c max = Max(c ij ), where 1≤i≤m, 1≤j≤n, The grid target location point number matrix is normalized, and the normalized grid target location point number matrix is: Step 3-4, setting the important area threshold value Δa l and the general area threshold value Δa s , when a ij ≥ Δa l , it is set as an important area flag, when Δa s ≤ a ij < Δa l , it is set as a general area flag, wherein 1≤i≤m, 1≤j≤n; Step 3-5, set the value of the important area marker to 1 and the value of other grids to 0 to obtain the important area grid number matrix A l ; set the value of the general area marker to 1 and the value of other grids to 0 to obtain the general area grid number matrix A s ; Step 3-6, the number of grids matrix A based on the key area l and the number of grids matrix A based on the general area s The key area and the general area are obtained and converted into latitude and longitude coordinates to obtain the extracted key area and general area.

Images