Relationship mining method and system based on perception data

Abstract

A relation mining method based on perception data comprises the following steps: accessing a front-end perception device, and constructing a front-end association relation pair according to the longitude and latitude of the front-end perception device; obtaining various types of sensing data according to the front-end sensing equipment, and performing entity trajectory clustering analysis on the sensing data to obtain various types of entity trajectories; according to the front-end association relationship pair and various entity trajectories, calculating an entity daily relationship to obtain an entity association detail pair; constructing an entity association relationship mining model, and inputting the entity association detail pair into the entity association relationship mining model to obtain an entity association relationship result; and optimizing the weights of the front-end association relationship pair and the model parameters according to the entity association relationship result. According to the method, the incidence relation mining model can be automatically constructed, adjusted and optimized, continuous mining, accumulation and iteration of fresh perceptual incidence relations among various entities are achieved, and finally a more accurate, more fresh and richer perceptual incidence relation network among the entities is output.

Description

Technical Field

[0001] The present invention relates to the technical field of data mining, and in particular to a relationship mining method and system based on perception data. Background Art

[0002] With the continuous development of the security industry, various types of perception front-ends are becoming more and more abundant in the construction of smart cities. The existing systems have realized the aggregation of massive perception data such as portrait data, motor vehicle data, and mobile phone data, and portrait feature recognition and motor vehicle feature recognition technologies are also widely used.

[0003] However, the current system only aggregates various types of sensory data to realize functions such as information query, trajectory display, and single-dimensional entity clustering analysis. The mining of associations between various entities has not been fully studied and mined, so the fresh relationship data contained in the sensory data has not been fully applied. Summary of the invention

[0004] In view of the above problems, the present invention is proposed to provide a method and system for relationship mining based on perception data that overcomes the above problems or at least partially solves the above problems.

[0005] In order to solve the above technical problems, the embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention discloses a relationship mining method based on perception data, comprising:

[0007] S100. Access the front-end sensing device and build a front-end association relationship pair according to the latitude and longitude of the front-end sensing device;

[0008] S200. Acquire various types of perception data according to the front-end perception device, perform entity trajectory cluster analysis on the perception data, and obtain various entity trajectories;

[0009] S300. Calculate the daily relationship of entities according to the front-end association relationship pairs and various entity trajectories to obtain entity association detail pairs;

[0010] S400. Construct an entity association relationship mining model, input the entity association details into the entity association relationship mining model, and obtain entity association relationship results;

[0011] S500. According to the entity association result, the weights of the front-end association pairs and model parameters are optimized;

[0012] S600. Repeat S300-S500 according to the actual data collected daily to achieve continuous self-iteration and tuning of the front-end relationship pairs and associated relationship models.

[0013] Further, in S100, a front-end association relationship pair is constructed according to the longitude and latitude of the front-end sensing device, and the specific method includes:

[0014] S101. Collecting data from the front-end sensing device, the data includes at least the latitude and longitude information of the front-end sensing device and other device information, the other device information includes at least the device number, device type, device status and device monitoring object;

[0015] S102. Clean the collected longitude and latitude data to remove duplicate, erroneous or invalid data; then preprocess the data, including at least format unification and coordinate conversion of the data;

[0016] S103. Analyze the space of the front-end device and use geographic information system tools to perform spatial analysis on the latitude and longitude data, including calculating the distance between devices, determining the relative position relationship between devices, and performing spatial clustering operations;

[0017] S104. Based on the results of the spatial analysis, the association relationship between the perception front-end devices is constructed by establishing a network diagram, an adjacency matrix or an association matrix between the devices.

[0018] Furthermore, in S200, various types of perception data are acquired according to the front-end perception device, and the various types of perception data at least include portrait data, motor vehicle data and mobile phone data.

[0019] Furthermore, in S200, entity trajectory clustering analysis is performed on the perception data to obtain various entity trajectories. The specific method includes:

[0020] S201. Acquire the perception data from the front-end device, remove duplicate, erroneous or invalid data points, process missing values, and organize data points with the same entity identifier into a continuous trajectory sequence;

[0021] S202. Encode the trajectory point sequence using a convolutional neural network to obtain a discriminative feature representation and extract trajectory feature points;

[0022] S203. Use the coordinates of the trajectory feature points to represent the trajectory space, add the timestamps of the trajectory feature points to the trajectory representation, and represent the trajectory time;

[0023] S204. Using a clustering algorithm, cluster the trajectory feature points after time and space representation to obtain various entity trajectories.

[0024] Furthermore, in S300, the front-end perception device obtains attribute information of various types of perception data including at least: target object ID, target object appearance time, target trajectory index, associated object ID, associated object appearance time, and associated object trajectory index.

[0025] Furthermore, in S400, an entity association relationship mining model is constructed, wherein the model is associated with time frequency A, number of associated devices B, associated area span C, and fusion relationship confidence Q. The specific construction method includes:

[0026] S401. Set the association time interval and minimum association number before model building, and calculate the relationship between personnel entities, motor vehicle entities, and mobile phone entities every day. The relationship includes at least person-to-person relationship, person-to-vehicle relationship, person-to-code relationship, vehicle-to-vehicle relationship, vehicle-to-code relationship, and code-to-code relationship. The relationship between the two must meet the requirements of association time interval and minimum association number;

[0027] S402. Obtain the association time frequency weight, the association device number weight, the association area span weight, the analyzed days and the sensitivity adjustment factor, and calculate the daily association relationship parameter d. The calculation formula is:

[0028]

[0029] Among them, pa is the association time frequency weight, A is the association time frequency, pb is the association device number weight, B is the number of associated devices, pc is the association area span weight, C is the association area span, σ is the sensitivity adjustment factor, n is the number of days analyzed and is not equal to 1, Q i is the confidence level that has been analyzed;

[0030] S403. According to the obtained daily correlation parameter d, the fusion relationship confidence Q is calculated, and the calculation formula is:

[0031]

[0032] Among them, d1, d2, ..., dn are the correlation parameters of the 1st day, the 2nd day, and the nth day, respectively; pd is the total confidence weight; pn is the total number of days weight; n is the number of days analyzed and is not equal to 1; and σ is the sensitivity adjustment factor.

[0033] Furthermore, in S500, the weights of the front-end association relationship pairs are optimized according to the entity association relationship results. The specific method includes: performing a perception front-end hit rate analysis based on the mined relationship results, adjusting the weight distribution of the perception front-end association relationship pairs, and eliminating the perception front-end association relationship pairs whose weights are still less than a preset threshold after a preset period.

[0034] Furthermore, in S500, according to the entity association relationship result, the model parameter weights are optimized, and the model parameter weights at least include the association time frequency weight, the association device number weight, the association area span weight, the confidence sum weight and the total day weight.

[0035] In a second aspect, an embodiment of the present invention discloses a relationship mining system based on perception data, including a front-end association relationship pair construction unit, an entity trajectory analysis unit, an entity association detail pair acquisition unit, an entity association relationship mining model construction unit and a weight optimization unit; wherein:

[0036] A front-end association relationship pair building unit, used to access the front-end sensing device and build a front-end association relationship pair according to the latitude and longitude of the front-end sensing device;

[0037] An entity trajectory analysis unit, configured to obtain various types of perception data according to the front-end perception device, perform entity trajectory cluster analysis on the perception data, and obtain various types of entity trajectories;

[0038] An entity association detail pair acquisition unit, configured to calculate the daily entity relationship according to the front-end association relationship pair and various entity trajectories to obtain the entity association detail pair;

[0039] An entity association relationship mining model building unit is used to build an entity association relationship mining model, input the entity association detail pair into the entity association relationship mining model, and obtain an entity association relationship result;

[0040] The weight optimization unit is used to optimize the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results.

[0041] In a third aspect, an embodiment of the present invention discloses an electronic device, including:

[0042] one or more processors;

[0043] A memory for storing one or more programs;

[0044] When the one or more programs are executed by the one or more processors, the one or more processors implement the relationship mining method.

[0045] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0046] The present invention discloses a method and system for relationship mining based on perception data, the method comprising: accessing a front-end perception device, constructing a front-end association relationship pair according to the longitude and latitude of the front-end perception device; acquiring various types of perception data according to the front-end perception device, performing entity trajectory clustering analysis on the perception data, and obtaining various types of entity trajectories; calculating the daily relationship of entities according to the front-end association relationship pairs and various types of entity trajectories, and obtaining entity association detail pairs; constructing an entity association relationship mining model, inputting the entity association detail pairs into the entity association relationship mining model, and obtaining entity association relationship results; optimizing the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results. The present invention can automatically construct and tune the association relationship mining model, realize the continuous mining, accumulation, and iteration of fresh perception association relationships between various entities, and finally output a more accurate, fresher, and richer perception association relationship network between entities.

[0047] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0049] Figure 1 This is a flow chart of a relationship mining method based on perception data in Embodiment 1 of the present invention;

[0050] Figure 2 This is a schematic diagram of the structure of an electronic device in Example 3 of the present invention. DETAILED DESCRIPTION

[0051] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0052] In order to solve the problems existing in the prior art, an embodiment of the present invention provides a relationship mining method and system based on perception data.

[0053] Example 1

[0054] The present invention discloses a method for mining relationships based on perception data. Figure 1 ,include:

[0055] S100. Access the front-end sensing device and build a front-end association relationship pair according to the longitude and latitude of the front-end sensing device; specifically, in the security industry, front-end sensing devices play a vital role. They are responsible for collecting, transmitting and parsing various security information and providing basic data support for the entire security system.

[0056] In this embodiment, the front-end sensing device includes at least a camera, an access control system, an alarm and a sensor. Specifically, the camera is one of the most commonly used front-end sensing devices in the security system, which is used to capture and record images and videos on site. Depending on the application scenario and needs, cameras can be divided into many types, such as infrared cameras (suitable for night monitoring), network cameras (supporting remote monitoring), high-definition cameras (providing high image quality for detail recognition), etc. Smart cameras can also perform advanced functions such as abnormal behavior recognition, face and vehicle recognition, further improving the monitoring effect.

[0057] The access control system is a front-end sensing device used to control the entry and exit of people or objects in a certain area. It usually consists of a card reader, a controller, a door body, etc., and supports multiple authentication methods such as swiping cards, passwords, and biometrics (such as fingerprints and facial recognition). The access control system can effectively prevent unauthorized entry, improve the security of entrances and exits, and work with the monitoring system to promptly alarm when abnormal situations occur. The alarm is a front-end sensing device used to detect dangerous or abnormal situations. When dangerous situations such as illegal intrusion, fire, and gas leakage are detected, the alarm will emit alarm sounds, lights, and other signals so that timely countermeasures can be taken. Common types of alarms include infrared alarms, microwave alarms, and vibration alarms. They are usually used in conjunction with devices such as cameras and access control systems to achieve 24-hour uninterrupted monitoring and alarms. A sensor is a front-end sensing device that can sense and detect certain physical quantities (such as temperature, humidity, smoke, light, etc.). In security systems, sensors are often used to monitor environmental parameters and object movement. For example, infrared sensors can detect human movement and trigger alarms; smoke sensors can detect smoke concentrations and sound fire alarms. The sensors are connected to a controller or central processing system to enable rapid response and processing of abnormal situations.

[0058] In S100 of this embodiment, a front-end association relationship pair is constructed according to the longitude and latitude of the front-end sensing device. The specific method includes:

[0059] S101. Collecting data from the front-end sensing device, the data includes at least the latitude and longitude information of the front-end sensing device and other device information, the other device information includes at least the device number, device type, device status and device monitoring object;

[0060] S102. Clean the collected longitude and latitude data to remove duplicate, erroneous or invalid data; then preprocess the data, including at least format unification and coordinate conversion of the data;

[0061] S103. Analyze the space of the front-end device and use geographic information system tools to perform spatial analysis on the latitude and longitude data, including calculating the distance between devices, determining the relative position relationship between devices, and performing spatial clustering operations;

[0062] S104. Based on the results of the spatial analysis, the association relationship between the perception front-end devices is constructed by establishing a network diagram, an adjacency matrix or an association matrix between the devices.

[0063] Through the above method, effective management and utilization of front-end sensing devices can be achieved.

[0064] S200. Obtain various types of perception data according to the front-end perception device, perform entity trajectory clustering analysis on the perception data, and obtain various types of entity trajectories; in S200 of this embodiment, various types of perception data are obtained according to the front-end perception device, and the various types of perception data at least include portrait data, motor vehicle data and mobile phone data.

[0065] In S200 of this embodiment, entity trajectory clustering analysis is performed on the perception data to obtain various entity trajectories. The specific method includes:

[0066] S201. Acquire the perception data from the front-end device, remove duplicate, erroneous or invalid data points, process missing values, and organize data points with the same entity identifier into a continuous trajectory sequence;

[0067] S202. Encode the trajectory point sequence using a convolutional neural network to obtain a discriminative feature representation and extract trajectory feature points;

[0068] S203. Use the coordinates of the trajectory feature points to represent the trajectory space, add the timestamps of the trajectory feature points to the trajectory representation, and represent the trajectory time;

[0069] S204. Using a clustering algorithm, cluster the trajectory feature points after time and space representation to obtain various entity trajectories.

[0070] S300. According to the front-end association relationship pairs and various entity trajectories, the entity daily relationship is calculated to obtain the entity association detail pair; in S300 of this embodiment, the front-end perception device obtains the attribute information of various perception data including at least: target object ID, target object appearance time, target trajectory index, associated object ID, associated object appearance time, and associated object trajectory index.

[0071] S400. Construct an entity association relationship mining model, input the entity association details into the entity association relationship mining model, and obtain entity association relationship results;

[0072] In S400 of this embodiment, an entity association relationship mining model is constructed. The model is associated with time frequency A, number of associated devices B, associated area span C, and fusion relationship confidence Q. The specific construction method includes:

[0073] S401. Set the association time interval and minimum association number before model building, and calculate the relationship between personnel entities, motor vehicle entities, and mobile phone entities every day. The relationship includes at least person-to-person relationship, person-to-vehicle relationship, person-to-code relationship, vehicle-to-vehicle relationship, vehicle-to-code relationship, and code-to-code relationship. The relationship between the two must meet the requirements of association time interval and minimum association number;

[0074] S402. Obtain the association time frequency weight, the association device number weight, the association area span weight, the analyzed days and the sensitivity adjustment factor, and calculate the daily association relationship parameter d. The calculation formula is:

[0075]

[0076] Among them, pa is the association time frequency weight, A is the association time frequency, pb is the association device number weight, B is the number of associated devices, pc is the association area span weight, C is the association area span, σ is the sensitivity adjustment factor, n is the number of days analyzed and is not equal to 1, Q i is the confidence level that has been analyzed;

[0077] S403. According to the obtained daily correlation parameter d, the fusion relationship confidence Q is calculated, and the calculation formula is:

[0078]

[0079] Among them, d1, d2, ..., dn are the correlation parameters of the 1st day, the 2nd day, and the nth day, respectively; pd is the total confidence weight; pn is the total number of days weight; n is the number of days analyzed and is not equal to 1; and σ is the sensitivity adjustment factor.

[0080] S500. Optimize the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results.

[0081] Specifically, in S500 of this embodiment, the weights of the front-end association relationship pairs are optimized according to the entity association relationship results. The specific method includes: performing a perception front-end hit rate analysis based on the mined relationship results, adjusting the weight distribution of the perception front-end association relationship pairs, and eliminating the perception front-end association relationship pairs whose weights are still less than a preset threshold after a preset period.

[0082] Specifically, in S500 of this embodiment, the model parameter weights are optimized according to the entity association relationship results, and the model parameter weights include at least the association time frequency weight, the association device number weight, the association area span weight, the confidence sum weight and the total day weight.

[0083] S600. Repeat S300-S500 according to the actual data collected daily to achieve continuous self-iteration and tuning of the front-end relationship pairs and associated relationship models.

[0084] The present embodiment discloses a method for relationship mining based on perception data, including: accessing a front-end perception device, and constructing a front-end association relationship pair according to the longitude and latitude of the front-end perception device; acquiring various types of perception data according to the front-end perception device, and performing entity trajectory clustering analysis on the perception data to obtain various types of entity trajectories; calculating the daily relationship of entities according to the front-end association relationship pairs and various types of entity trajectories to obtain entity association detail pairs; constructing an entity association relationship mining model, and inputting the entity association detail pairs into the entity association relationship mining model to obtain entity association relationship results; optimizing the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results. The present invention can automatically construct and tune the association relationship mining model, realize the continuous mining, accumulation, and iteration of fresh perception association relationships between various entities, and finally output a more accurate, fresher, and richer perception association relationship network between entities.

[0085] Example 2

[0086] Based on the same inventive concept, the embodiment of the present disclosure also provides a relationship mining system based on perception data, including a front-end association relationship pair construction unit, an entity trajectory analysis unit, an entity association detail pair acquisition unit, an entity association relationship mining model construction unit and a weight optimization unit; wherein:

[0087] A front-end association relationship pair building unit, used to access the front-end sensing device and build a front-end association relationship pair according to the latitude and longitude of the front-end sensing device;

[0088] An entity trajectory analysis unit, configured to obtain various types of perception data according to the front-end perception device, perform entity trajectory cluster analysis on the perception data, and obtain various types of entity trajectories;

[0089] An entity association detail pair acquisition unit, configured to calculate the daily entity relationship according to the front-end association relationship pair and various entity trajectories to obtain the entity association detail pair;

[0090] An entity association relationship mining model building unit is used to build an entity association relationship mining model, input the entity association detail pair into the entity association relationship mining model, and obtain an entity association relationship result;

[0091] The weight optimization unit is used to optimize the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results.

[0092] Among them, the specific working methods of the front-end association relationship pair construction unit, the entity trajectory analysis unit, the entity association detail pair acquisition unit, the entity association relationship mining model construction unit and the weight optimization unit have been recorded in detail in Example 1, and will not be repeated in this embodiment.

[0093] Example 3

[0094] Based on the same inventive concept, an embodiment of the present disclosure also provides an electronic device. Figure 2 FIG. 1 is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Figure 2 As shown, an embodiment of the present disclosure provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement any optimization method in the above embodiments; the one or more I / O interfaces 103 are connected between the processor and the memory, and are configured to implement information interaction between the processor and the memory.

[0095] Among them, the processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU), etc.; the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH); the I / O interface (read-write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus), etc.

[0096] In some embodiments, the processor 101 , the memory 102 , and the I / O interface 103 are connected to each other via a bus 104 , and further connected to other components of the computing device.

[0097] In some embodiments, the one or more processors 101 include a field programmable gate array.

[0098] According to an embodiment of the present disclosure, a computer-readable medium is further provided, wherein a computer program is stored on the computer-readable medium, wherein when the program is executed by a processor, the steps in any optimization method in the above-mentioned embodiment are implemented.

[0099] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process can be rearranged without departing from the scope of protection of the present disclosure. The attached method claims present the elements of the various steps in an exemplary order and are not intended to be limited to the specific order or hierarchy described.

[0100] In the above detailed description, various features are grouped together in a single embodiment to simplify the disclosure. This method of disclosure should not be interpreted as reflecting an intention that the embodiments of the claimed subject matter require more features than are clearly stated in each claim. On the contrary, as reflected in the appended claims, the invention is in a state of having less than all the features of the disclosed individual embodiments. Therefore, the appended claims are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate preferred embodiment of the invention.

[0101] Those skilled in the art will also appreciate that the various illustrative logic blocks, modules, circuits, and algorithmic steps described in conjunction with the embodiments herein can all be implemented as electronic hardware, computer software, or a combination thereof. In order to clearly illustrate the interchangeability between hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above around their functions. Whether such functions are implemented as hardware or software depends on specific applications and the design constraints imposed on the entire system. A skilled person can implement the described functions in an alternative manner for each specific application, but such implementation decisions should not be interpreted as departing from the scope of protection of the present disclosure.

[0102] The steps of the method or algorithm described in conjunction with the embodiments herein may be directly embodied as hardware, a software module executed by a processor, or a combination thereof. The software module may be located in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a mobile disk, a CD-ROM, or any other form of storage medium well known in the art. An exemplary storage medium is connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Of course, the storage medium may also be an integral part of the processor. The processor and the storage medium may be located in an ASIC. The ASIC may be located in a user terminal. Of course, the processor and the storage medium may also be present in a user terminal as discrete components.

[0103] For software implementation, the techniques described in this application can be implemented with modules (e.g., procedures, functions, etc.) that perform the functions described in this application. These software codes can be stored in a memory unit and executed by a processor. The memory unit can be implemented within the processor or outside the processor. In the latter case, it is coupled to the processor in a communication manner via various means, which are well known in the art.

[0104] The above description includes examples of one or more embodiments. Of course, it is impossible to describe all possible combinations of components or methods for the purpose of describing the above embodiments, but it should be recognized by those skilled in the art that the various embodiments may be further combined and arranged. Therefore, the embodiments described herein are intended to cover all such changes, modifications and variations that fall within the scope of protection of the appended claims. In addition, with respect to the term "comprising" used in the specification or claims, the word is covered in a manner similar to the term "including", just as "including," is explained as a transitional word in the claims. In addition, any term "or" used in the specification of the claims is intended to mean "non-exclusive or".

Claims

1. A relationship mining method based on perception data, characterized in that: include: S100. Access the front-end sensing device and build a front-end association relationship pair according to the latitude and longitude of the front-end sensing device; S200. Acquire various types of perception data according to the front-end perception device, perform entity trajectory cluster analysis on the perception data, and obtain various entity trajectories; S300. Calculate the daily relationship of entities according to the front-end association relationship pairs and various entity trajectories to obtain entity association detail pairs; S400. Construct an entity association relationship mining model, input the entity association details into the entity association relationship mining model, and obtain entity association relationship results; S500. According to the entity association result, the weights of the front-end association pairs and model parameters are optimized; S600. Repeat S300-S500 according to the actual data collected daily to achieve continuous self-iteration and tuning of the front-end relationship pairs and associated relationship models.

2. A method for relation mining based on perception data as described in claim 1, characterized in that: In S100, a front-end association relationship pair is constructed according to the longitude and latitude of the front-end sensing device. The specific method includes: S101. Collecting data from the front-end sensing device, the data includes at least the latitude and longitude information of the front-end sensing device and other device information, the other device information includes at least the device number, device type, device status and device monitoring object; S102. Clean the collected longitude and latitude data to remove duplicate, erroneous or invalid data; then preprocess the data, including at least format unification and coordinate conversion of the data; S103. Analyze the space of the front-end device and use geographic information system tools to perform spatial analysis on the latitude and longitude data, including calculating the distance between devices, determining the relative position relationship between devices, and performing spatial clustering operations; S104. Based on the results of the spatial analysis, the association relationship between the perception front-end devices is constructed by establishing a network diagram, an adjacency matrix or an association matrix between the devices.

3. A method for relation mining based on perception data as described in claim 1, characterized in that: In S200, various types of perception data are acquired according to the front-end perception device, and the various types of perception data at least include portrait data, motor vehicle data and mobile phone data.

4. A method for relation mining based on perception data as described in claim 1, characterized in that: In S200, cluster analysis of entity trajectories is performed on the perception data to obtain various entity trajectories. The specific method includes: S201. Acquire the perception data from the front-end device, remove duplicate, erroneous or invalid data points, process missing values, and organize data points with the same entity identifier into a continuous trajectory sequence; S202. Encode the trajectory point sequence using a convolutional neural network to obtain a discriminative feature representation and extract trajectory feature points; S203. Use the coordinates of the trajectory feature points to represent the trajectory space, add the timestamps of the trajectory feature points to the trajectory representation, and represent the trajectory time; S204. Using a clustering algorithm, cluster the trajectory feature points after time and space representation to obtain various entity trajectories.

5. A method for relation mining based on perception data as described in claim 1, characterized in that: In S300, the front-end sensing device obtains attribute information of various types of sensing data, including at least: target object ID, target object appearance time, target trajectory index, associated object ID, associated object appearance time, and associated object trajectory index.

6. A method for relation mining based on perception data as described in claim 1, characterized in that: In S400, an entity association relationship mining model is constructed. The model is associated with time frequency A, number of associated devices B, associated area span C, and fusion relationship confidence Q. The specific construction method includes: S401. Set the association time interval and minimum association number before model building, and calculate the relationship between personnel entities, motor vehicle entities, and mobile phone entities every day. The relationship includes at least person-to-person relationship, person-to-vehicle relationship, person-to-code relationship, vehicle-to-vehicle relationship, vehicle-to-code relationship, and code-to-code relationship. The relationship between the two must meet the requirements of association time interval and minimum association number; S402. Obtain the association time frequency weight, the association device number weight, the association area span weight, the analyzed days and the sensitivity adjustment factor, and calculate the daily association relationship parameter d. The calculation formula is: Among them, pa is the association time frequency weight, A is the association time frequency, pb is the association device number weight, B is the number of associated devices, pc is the association area span weight, C is the association area span, σ is the sensitivity adjustment factor, n is the number of days analyzed and is not equal to 1, Q i is the confidence level that has been analyzed; S403. According to the obtained daily correlation parameter d, the fusion relationship confidence Q is calculated, and the calculation formula is: Among them, d1, d2, ..., dn are the correlation parameters of the 1st day, the 2nd day, and the nth day, respectively; pd is the total confidence weight; pn is the total number of days weight; n is the number of days analyzed and is not equal to 1; and σ is the sensitivity adjustment factor.

7. A method for relation mining based on perception data as described in claim 1, characterized in that: In S500, according to the entity association relationship results, the weights of the front-end association relationship pairs are optimized. The specific method includes: performing a perception front-end hit rate analysis based on the mined relationship results, adjusting the weight distribution of the perception front-end association relationship pairs, and eliminating the perception front-end association relationship pairs whose weights are still less than a preset threshold after a preset period.

8. A method for relation mining based on perception data as claimed in claim 1, characterized in that: In S500, according to the entity association result, the model parameter weights are optimized, and the model parameter weights at least include the association time frequency weight, the association device number weight, the association area span weight, the confidence sum weight and the total day weight.

9. A relationship mining system based on perception data, characterized in that: It includes a front-end association relationship pair construction unit, an entity trajectory analysis unit, an entity association detail pair acquisition unit, an entity association relationship mining model construction unit and a weight optimization unit; wherein: A front-end association relationship pair building unit, used to access the front-end sensing device and build a front-end association relationship pair according to the latitude and longitude of the front-end sensing device; An entity trajectory analysis unit, configured to obtain various types of perception data according to the front-end perception device, perform entity trajectory cluster analysis on the perception data, and obtain various types of entity trajectories; An entity association detail pair acquisition unit, configured to calculate the daily entity relationship according to the front-end association relationship pair and various entity trajectories to obtain the entity association detail pair; An entity association relationship mining model building unit is used to build an entity association relationship mining model, input the entity association detail pair into the entity association relationship mining model, and obtain an entity association relationship result; The weight optimization unit is used to optimize the weights of the front-end association relationship pairs and model parameters according to the entity association relationship results.

10. An electronic device comprising: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the relationship mining method.

Images