Related personnel determination method based on trans-provincial behavior

Abstract

The invention relates to a related personnel determination method based on trans-provincial behaviors, and belongs to the technical field of big data and security. According to the method for determining the related personnel based on the cross-provincial behaviors, the cross-provincial behavior information of the personnel with the incorrect behaviors is introduced in the related personnel discovery process, so that the flow track information of the personnel with the incorrect behaviors among provinces is considered, more similar related personnel can be discovered in the related personnel discovery process, and the related personnel discovery efficiency is improved. And a debugging condition when only the parking track near the parking point is considered is avoided. According to the method, cross-provincial behavior information is introduced for related personnel discovery problems, the flow track information of incorrect behavior personnel among provinces within a period of time after a crime can be considered by comparing the cross-provincial behavior information of the cross-provincial personnel and the cross-provincial behavior information of the abnormal personnel, and more effective information is introduced, so that the error checking conditions are reduced, and the efficiency is improved. And more accurate related personnel discovery is realized.

Description

Technical Field

[0001] The present invention relates to a method for determining relevant personnel based on cross-provincial behavior, and belongs to the field of big data and security technology. Background Art

[0002] The improper behaviors of people tend to be coordinated and multi-person. At the same time, people with improper behaviors often take the approach of rapid cross-provincial movement to escape after the incident. Therefore, it is of great significance to dig out the related persons of people with improper behaviors and obtain the improper behaviors across provinces.

[0003] At present, the methods for determining related persons with improper behavior mainly include traditional methods and methods based on big data. The existing technology adopts a method based on cell signaling information to determine the cell information and cell users where improper behavior persons have appeared, and analyzes the probability of similarity between the cell users and the cell movement trajectory of the improper behavior person, thereby completing the discovery of related persons with improper behavior. The existing technology adopts a method based on mobile terminal positioning to determine the target location of the improper behavior person and other persons within a preset range centered on the target location, and compares the movement trajectory between the two, thereby completing the discovery of related persons with improper behavior.

[0004] Traditional methods usually consume a lot of manpower and material resources when discovering relevant personnel, and are often time-consuming. Existing big data-based methods only consider the trajectory information of people with incorrect behavior around the parking area when discovering relevant personnel, which easily screens out more normal residents in the area where people with incorrect behavior have stopped, and does not consider the flow trajectory information of people with incorrect behavior between provinces. At the same time, the above work does not consider how to complete the discovery of relevant personnel when trajectory information is missing. Summary of the invention

[0005] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art, consider the behavioral characteristics of cross-provincial behavior of people with incorrect behavior, and propose a method for determining relevant personnel based on cross-provincial behavior to solve the problem that the existing method does not consider the flow trajectory information of people with incorrect behavior between provinces and the possible missing trajectory information.

[0006] The technical solution of the present invention is:

[0007] A method for determining relevant persons based on cross-provincial behavior, the method comprising the steps of:

[0008] The first step is to collect the cross-provincial behavior of people between provinces through monitoring equipment. The cross-provincial behavior information mainly includes the identity information of the person, the starting province information of the cross-province, the ending province information of the cross-province, the time information of leaving the starting province and the time information of arriving at the ending province. The identity information can uniquely identify a person;

[0009] The second step is to set the monitoring time length T1, and store the interprovincial trajectory information collected within the time length T1 into the interprovincial-behavior-repository;

[0010] The third step is to intersect and collide the discovered illegal-person-repository with the interprovincial behavior repository to obtain the illegal-person-interprovincial-behavior-repository.

[0011] The fourth step is to intersect and collide the cross-provincial behavior database of people with incorrect behaviors with the total cross-provincial behavior database, and obtain the cross-provincial start-end segment similarity based on the starting and ending provinces of each cross-provincial behavior of the cross-provincial people and the people with incorrect behaviors;

[0012] The fifth step is to intersect and collide the cross-provincial behavior database of people with incorrect behaviors with the total cross-provincial behavior database, and obtain the cross-provincial time segment similarity based on the departure time of the cross-provincial people and the arrival time of the end province of each cross-provincial behavior of the people with incorrect behaviors;

[0013] The sixth step is to calculate the segmented similarity of the cross-provincial behavior of the cross-provincial personnel and the personnel with incorrect behavior based on the cross-provincial start-end segmented similarity and the cross-provincial time segmented similarity of the cross-provincial personnel and the personnel with incorrect behavior, and then calculate the overall similarity of the cross-provincial behavior of the cross-provincial personnel and the personnel with incorrect behavior;

[0014] The seventh step is to determine whether the cross-provincial person is a related person of the person with incorrect behavior based on the overall similarity of the cross-provincial behavior between the cross-provincial person and the person with incorrect behavior. If the overall similarity of the cross-provincial behavior is greater than the set threshold, the cross-provincial person is determined to be a related person of the person with incorrect behavior, and then a database of related persons of the person with incorrect behavior is obtained.

[0015] In the second step, the time length T1 should be selected to be a moderate period of time, generally 7-10 days. When collecting cross-provincial behavior information, only the last time when the cross-provincial personnel left the initial province and the first time when they arrived at the destination province are retained as cross-provincial time information;

[0016] In the third step, the database of personnel with incorrect behavior contains the identity information of the personnel with incorrect behavior. By comparing the identity information in the database of personnel with incorrect behavior with the identity information in the total database of cross-provincial behavior, the cross-provincial behavior information of the cross-provincial personnel with consistent identity information is extracted and added to the cross-provincial behavior database of personnel with incorrect behavior. The cross-provincial behavior database of personnel with incorrect behavior contains the identity information of the personnel with incorrect behavior, the cross-provincial starting province information, the cross-provincial ending province information, the time information of leaving the starting province, and the time information of arriving at the ending province;

[0017] In the fourth step, the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior database of persons with incorrect behaviors are extracted in turn and compared with the overall cross-provincial behaviors of different cross-provincial persons in the cross-provincial behavior database. It is assumed that the overall cross-provincial behavior of a certain person with incorrect behaviors in the cross-provincial behavior database of incorrect behaviors includes m cross-provincial behaviors, and the overall cross-provincial behavior of a certain cross-provincial person in the total cross-provincial behavior database includes n cross-provincial behaviors. The similarity s1 of the cross-provincial start and end segments between the person with incorrect behaviors and the cross-provincial person is calculated. ij :

[0018]

[0019] Among them, 1≤i≤m, 1≤j≤n, P_st i P_des is the starting province of the i-th cross-provincial behavior of the person with improper behavior, i P_st is the destination province of the i-th cross-provincial behavior of the person with improper behavior, j P_des is the starting province of the jth cross-provincial behavior of the cross-provincial person, j The province where the jth segment of the cross-provincial behavior of the cross-provincial person ends;

[0020] In the fifth step, the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior library of persons with incorrect behaviors are extracted in sequence and compared with the overall cross-provincial behaviors of different persons in the cross-provincial behavior library. It is assumed that the overall cross-provincial behavior of a certain person with incorrect behaviors in the cross-provincial behavior library of incorrect behaviors includes m segments of cross-provincial behaviors, and the overall cross-provincial behavior of a certain cross-provincial person in the cross-provincial behavior library includes n segments of cross-provincial behaviors. The cross-provincial time segment similarity s2 of the person with incorrect behaviors and the cross-provincial person is calculated. ij :

[0021]

[0022] Among them, 1≤i≤m, 1≤j≤n, T_st i T_des is the time when the person with incorrect behavior leaves the starting province for the i-th cross-provincial behavior, i T_st is the time when the person with improper behavior arrives at the destination province for the i-th cross-provincial behavior,j T_des is the time when the inter-provincial person leaves the starting province for the jth inter-provincial behavior. j is the time for the cross-provincial person to arrive at the destination province for the jth cross-provincial behavior; ΔT is the time threshold for judging whether the departure time of different cross-provincial people from the starting province or the arrival time of the destination province is similar, which is generally selected to be 5-10 minutes;

[0023] In the sixth step, according to the similarity s1 between the start and end segments of the inter-provincial personnel and the personnel with incorrect behavior ij Similarity with cross-provincial time segment s2 ij , calculate the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors ij :

[0024] s ij =s1 ij ·s2 ij

[0025] Then, according to the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors, ij , calculate the overall similarity s between the cross-provincial behaviors of cross-provincial personnel and those with incorrect behaviors:

[0026]

[0027] In the seventh step, whether the cross-provincial person is a related person of the person with incorrect behavior is determined based on the overall similarity s between the cross-provincial person and the person with incorrect behavior. If s is greater than or equal to the similarity threshold S, the cross-provincial person is determined to be a related person of the person with incorrect behavior. The calculation method of the similarity threshold S' is:

[0028]

[0029] Among them, 1≤p≤min(m,n), p is the number of similar cross-provincial behavior segments between cross-provincial personnel and personnel with incorrect behavior.

[0030] Beneficial Effects

[0031] 1. A method for determining related personnel based on cross-provincial behavior of the present invention introduces cross-provincial behavior information of personnel with incorrect behavior in the process of related personnel discovery, so as to consider the flow trajectory information of personnel with incorrect behavior between various provinces. This enables the related personnel discovery process to discover more similar related personnel and avoids the error detection situation when only considering the stop trajectory near the stop point.

[0032] 2. When determining the segmented similarity of the cross-provincial behaviors of cross-provincial personnel and illegal personnel, the cross-provincial start-end segmented similarity and the cross-provincial time segmented similarity of the two are considered and integrated at the same time; when determining the overall similarity of the cross-provincial behaviors of cross-provincial personnel and personnel with incorrect behaviors, the cross-provincial behaviors of the two are first calculated in segments and then integrated. This avoids the complexity of directly determining the overall similarity of the cross-provincial behaviors of the two, and it can also be determined when there is missing information on the cross-provincial behaviors of the two.

[0033] 3. Aiming at the problem of discovering relevant personnel, the present invention introduces cross-provincial behavior information. By comparing the cross-provincial behavior information of cross-provincial personnel with that of abnormal personnel, it is possible to consider the flow trajectory information of personnel with incorrect behavior between various provinces within a period of time after the incident, and introduce more effective information, thereby reducing the number of errors and achieving more accurate discovery of relevant personnel. DETAILED DESCRIPTION

[0034] The following is a detailed description of a method for determining related personnel based on cross-provincial behavior of the present invention in conjunction with an embodiment:

[0035] A method for determining relevant persons based on cross-provincial behavior, the method comprising the steps of:

[0036] The first step is to collect the cross-provincial behavior of people between provinces through monitoring equipment. The cross-provincial behavior information mainly includes the identity information of the person, the starting province information of the cross-province, the ending province information of the cross-province, the time information of leaving the starting province, and the time information of arriving at the ending province;

[0037] The second step is to set the monitoring time length T1, and store the cross-provincial trajectory information collected within the T1 time length into the cross-provincial behavior database. T1 should be a relatively moderate period of time, which is set to 7 days in this example;

[0038] The third step is to intersect and collide the discovered improper behavior personnel database with the cross-provincial behavior database based on personnel identity information to obtain the cross-provincial behavior database of improper behavior personnel;

[0039] The fourth step is to extract the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior database of persons with incorrect behaviors and compare them with the overall cross-provincial behaviors of different cross-provincial persons in the cross-provincial behavior database in sections. It is assumed that the overall cross-provincial behavior of a person with incorrect behaviors in the cross-provincial behavior database of incorrect behaviors includes 4 sections of cross-provincial behaviors, and the start and end situations of his cross-provincial behaviors are shown in Table 1. The overall cross-provincial behavior of a cross-provincial person in the cross-provincial behavior database includes 3 sections of cross-provincial behaviors, and the start and end situations of his cross-provincial behaviors are shown in Table 2. Therefore, the cross-provincial start and end segment similarity s1 between the person with incorrect behaviors and the cross-provincial person can be calculated. ij :

[0040]

[0041] Among them, 1≤i≤m, 1≤j≤n, m=4, n=3, P_st i P_des is the starting province of the i-th cross-provincial behavior of the person with improper behavior, i P_st is the destination province of the i-th cross-provincial behavior of the person with improper behavior, j P_des is the starting province of the jth cross-provincial behavior of the cross-provincial person, j The province where the jth segment of the cross-provincial behavior of the cross-provincial person ends;

[0042] Table 1 The beginning and end of a person's cross-provincial behavior

[0043] Identity information of persons with inappropriate behavior Information about the province where the cross-province entry begins Interprovincial destination province information 0001 Province A Province B 0001 Province B Province C 0001 Province C Province D 0001 Province D Province E

[0044] Table 2 The start and end of a person's cross-provincial behavior

[0045] Identity information of people across provinces Information about the province where the cross-province entry begins Interprovincial destination province information 0010 Province A Province B 0010 Province B Province C 0010 Province D Province E

[0046] The fifth step is to extract the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior database of persons with incorrect behaviors and compare them with the overall cross-provincial behaviors of different persons in the cross-provincial behavior database in sections. It is assumed that the overall cross-provincial behavior of a person with incorrect behaviors in the cross-provincial behavior database of incorrect behaviors includes 4 sections of cross-provincial behaviors, and the time situation of his cross-provincial behaviors is shown in Table 3. The overall cross-provincial behavior of a cross-provincial person in the cross-provincial behavior database includes 3 sections of cross-provincial behaviors, and the time situation of his cross-provincial behaviors is shown in Table 4. Therefore, the cross-provincial time segment similarity s2 between the person with incorrect behaviors and the cross-provincial person can be calculated. ij :

[0047]

[0048] Among them, 1≤i≤m, 1≤j≤n, m=4, n=3, T_st i T_des is the time when the person with incorrect behavior leaves the starting province for the i-th cross-provincial behavior, i T_st is the time when the person with improper behavior arrives at the destination province for the i-th cross-provincial behavior, j T_des is the time when the inter-provincial person leaves the starting province for the jth inter-provincial behavior. j The time it takes for the inter-provincial person to reach the destination province for the jth inter-provincial behavior. ΔT is set to 5 minutes in this example.

[0049] Table 3 Time of cross-provincial behavior of a person with improper behavior

[0050] Identity information of persons with inappropriate behavior Time information for leaving the starting province Time information of arrival at destination province 0001 2020-07-21 21:00 2020-07-22 03:00 0001 2020-07-23 11:00 2020-07-23 19:00 0001 2020-07-25 13:00 2020-07-25 20:00 0001 2020-07-27 22:00 2020-07-28 05:00

[0051] Table 4 The time of a person's interprovincial behavior

[0052] Identity information of people across provinces Time information for leaving the starting province Time information of arrival at destination province 0010 2020-07-21 21:02 2020-07-22 02:59 0010 2020-07-23 10:59 2020-07-23 19:01 0010 2020-07-27 21:58 2020-07-28 05:02

[0053] Step 6: Based on the similarity s1 between the start and end segments of the cross-province personnel and the personnel with incorrect behavior ij Similarity with cross-provincial time segment s2 ij , calculate the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors ij :

[0054] s ij =s1 ij ·s2 ij

[0055] In this example, the segmented similarity s between the cross-provincial behaviors of the cross-provincial personnel and those with incorrect behaviors ij The calculation results are shown in Table 5:

[0056] Table 5 Calculation results of segmented similarity of cross-provincial behaviors between cross-provincial personnel and personnel with incorrect behaviors

[0057]

[0058] Then, according to the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors, ij , calculate the overall similarity s between the cross-provincial behaviors of cross-provincial personnel and those with incorrect behaviors:

[0059]

[0060] In this case, the overall similarity s between the cross-provincial behaviors of the cross-provincial personnel and those with incorrect behaviors is calculated to be 1;

[0061] In the seventh step, whether the cross-provincial person is a related person of the person with incorrect behavior is determined based on the overall similarity s between the cross-provincial person and the person with incorrect behavior. If s is greater than or equal to the similarity threshold S, the cross-provincial person is determined to be a related person of the person with incorrect behavior. The calculation method of the similarity threshold S' is:

[0062]

[0063] Among them, p is the number of similar cross-provincial behavior segments between the cross-provincial personnel and the personnel with incorrect behavior, which is set to 2 in this example, and the value of S' is approximately 0.67. Therefore, in this example, the overall similarity s between the cross-provincial behavior of the cross-provincial personnel and the personnel with incorrect behavior is greater than the similarity threshold S. Therefore, it can be determined that the cross-provincial personnel is a related person of the personnel with incorrect behavior and is added to the database of related personnel of the personnel with incorrect behavior.

[0064] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining relevant personnel based on cross-provincial behavior, characterized in that The steps of the method include: The first step is to collect cross-provincial behaviors of people between provinces through monitoring equipment; The second step is to store the cross-provincial trajectory information collected during the monitoring period into the cross-provincial behavior database; The third step is to intersect and collide the discovered database of improperly behaved personnel with the total database of cross-provincial behaviors to obtain a database of cross-provincial behaviors of improperly behaved personnel; The fourth step is to intersect and collide the cross-provincial behavior database of people with incorrect behaviors with the total cross-provincial behavior database, and obtain the cross-provincial start-end segment similarity based on the starting and ending provinces of each cross-provincial behavior of the cross-provincial people and the people with incorrect behaviors; The fifth step is to intersect and collide the cross-provincial behavior database of people with incorrect behaviors with the total cross-provincial behavior database, and obtain the cross-provincial time segment similarity based on the departure time of the cross-provincial people and the arrival time of the end province of each cross-provincial behavior of the people with incorrect behaviors; Step 6: Calculate the segmented similarity of the cross-provincial behaviors of the cross-provincial personnel and the personnel with incorrect behaviors based on the cross-provincial start-end segmented similarity and the cross-provincial time segmented similarity, and then calculate the overall similarity of the cross-provincial behaviors of the cross-provincial personnel and the personnel with incorrect behaviors based on the segmented similarity; The seventh step is to determine whether the cross-provincial person is a related person of the person with incorrect behavior based on the overall similarity of the cross-provincial behavior between the cross-provincial person and the person with incorrect behavior. If the overall similarity of the cross-provincial behavior is greater than the set threshold, the cross-provincial person is determined to be a related person of the person with incorrect behavior, and a database of related persons of the person with incorrect behavior is obtained. If the overall similarity of the cross-provincial behavior is not greater than the set threshold, the cross-provincial person is determined not to be a related person of the person with incorrect behavior.

2. According to claim 1, a method for determining related personnel based on cross-provincial behavior is characterized by: In the first step, the cross-provincial behavior information includes the person's identity information, cross-provincial starting province information, cross-provincial ending province information, time information of leaving the starting province, and time information of arriving at the ending province, wherein the identity information can uniquely identify a person.

3. A method for determining related personnel based on cross-provincial behavior according to claim 1 or 2, characterized in that: In the second step, the monitoring time length T1 is 7-10 days. When collecting cross-provincial behavior information, only the last time when the cross-provincial personnel leave the initial province and the first time when they arrive at the destination province are retained as the cross-provincial time information.

4. According to claim 1, a method for determining related personnel based on cross-provincial behavior is characterized by: In the third step, the database of persons with incorrect behavior contains the identity information of persons with incorrect behavior. By comparing the identity information in the database of persons with incorrect behavior with the identity information in the total database of cross-provincial behavior, the cross-provincial behavior information of cross-provincial persons with consistent identity information is extracted and added to the cross-provincial behavior database of persons with incorrect behavior. The cross-provincial behavior database of persons with incorrect behavior contains the identity information of persons with incorrect behavior, cross-provincial starting province information, cross-provincial ending province information, time information of leaving the starting province, and time information of arriving at the ending province.

5. The method for determining related personnel based on cross-provincial behavior according to claim 1 is characterized by: In the fourth step, the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior database of persons with incorrect behaviors are extracted in turn and compared with the overall cross-provincial behaviors of different cross-provincial persons in the cross-provincial behavior database. It is assumed that the overall cross-provincial behavior of a certain person with incorrect behaviors in the cross-provincial behavior database of incorrect behaviors includes m cross-provincial behaviors, and the overall cross-provincial behavior of a certain cross-provincial person in the total cross-provincial behavior database includes n cross-provincial behaviors. The similarity s1 of the cross-provincial start and end segments between the person with incorrect behaviors and the cross-provincial person is calculated. ij : Among them, 1≤i≤m, 1≤j≤n, P_st i P_des is the starting province of the i-th cross-provincial behavior of the person with improper behavior, i P_st is the destination province of the i-th cross-provincial behavior of the person with improper behavior, j P_des is the starting province of the jth cross-provincial behavior of the cross-provincial person, j It is the destination province of the jth cross-provincial behavior of the cross-provincial person.

6. The method for determining related personnel based on cross-provincial behavior according to claim 1 is characterized by: In the fifth step, the overall cross-provincial behaviors of different persons with incorrect behaviors in the cross-provincial behavior library of persons with incorrect behaviors are extracted in sequence and compared with the overall cross-provincial behaviors of different persons in the cross-provincial behavior library. It is assumed that the overall cross-provincial behavior of a certain person with incorrect behaviors in the cross-provincial behavior library of incorrect behaviors includes m segments of cross-provincial behaviors, and the overall cross-provincial behavior of a certain cross-provincial person in the cross-provincial behavior library includes n segments of cross-provincial behaviors. The cross-provincial time segment similarity s2 of the person with incorrect behaviors and the cross-provincial person is calculated. ij : Among them, 1≤i≤m, 1≤j≤n, T_st i T_des is the time when the person with incorrect behavior leaves the starting province for the i-th cross-provincial behavior, i T_st is the time when the person with improper behavior arrives at the destination province for the i-th cross-provincial behavior, j T_des is the time when the inter-provincial person leaves the starting province for the jth inter-provincial behavior. j is the time for the inter-provincial person to arrive at the destination province for the jth inter-provincial behavior; ΔT is the time threshold for judging whether the time for different inter-provincial people to leave the starting province or arrive at the destination province is similar, and 5-10 minutes is selected.

7. The method for determining related personnel based on cross-provincial behavior according to claim 1 is characterized by: In the sixth step, according to the similarity s1 between the start and end segments of the inter-provincial personnel and the personnel with incorrect behavior ij Similarity with cross-provincial time segment s2 ij , calculate the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors ij : s ij =s1 ij ·s2 ij Then, according to the segmented similarity s between the cross-provincial personnel and the personnel with incorrect behaviors, ij , calculate the overall similarity s between the cross-provincial behaviors of cross-provincial personnel and those with incorrect behaviors:

8. The method for determining related personnel based on cross-provincial behavior according to claim 1 is characterized by: In the seventh step, whether the cross-provincial person is a related person of the person with incorrect behavior is determined based on the overall similarity s between the cross-provincial person and the person with incorrect behavior. If s is greater than or equal to the similarity threshold S, the cross-provincial person is determined to be a related person of the person with incorrect behavior. The calculation method of the similarity threshold S' is: Among them, 1≤p≤min(m,n), p is the number of similar cross-provincial behavior segments between cross-provincial personnel and personnel with incorrect behavior.

Images