CN114612932A - Gait big data retrieval method and system and terminal equipment - Google Patents

Abstract

The invention discloses a gait big data retrieval method, a gait big data retrieval system and terminal equipment, belongs to the technical field of gait recognition, and can realize that a user can use a gait retrieval technology normally and correctly to obtain the optimal retrieval effect, so that target figure track information can be drawn more abundantly and accurately based on a retrieval result, and the problem that the retrieval effect is different from person to person is solved. The method comprises the following steps: s1, acquiring a video image containing a target person, extracting a gait sample of the target person from the video image, and constructing a gait library of the target person according to the gait sample; s2, searching in the monitoring video data of the local area by using the gait library to obtain a local searching result, and updating the gait library by using the local searching result to obtain an updated gait library; s3, searching in the monitoring video data of the whole area by using the updated gait library to obtain a global searching result; and S4, generating the track of the target person according to the local search result and the global search result. The gait retrieval method is used for gait retrieval.

Description

A gait big data retrieval method, system and terminal device thereof

technical field

The invention relates to a gait big data retrieval method, system and terminal equipment, belonging to the technical field of gait recognition.

Background technique

In the AI era, more and more biometric features can be extracted and identified. At present, there are many technologies such as face recognition, fingerprint recognition, voiceprint recognition, iris recognition, palm vein recognition, and gait recognition in the market.

Gait recognition refers to the identification of the target person by analyzing the body structure and walking posture of the target person. Compared with other biometric technologies, gait recognition has the advantages of non-contact long-distance and not easy to camouflage. It is widely used in the tracking of missing children in safe cities, social governance and other fields, the dynamic deployment of target suspects, and key groups of people. dynamic monitoring and other scenarios.

As a new biometric technology in the current security industry, gait recognition has achieved a high recognition accuracy rate in academia. However, in practical applications, when many users perform data retrieval based on gait recognition technology, a large number of retrieval results are missing or the retrieval results are mixed with noise data. This kind of retrieval results are inconsistent due to insufficient operating experience or irregular operation of the users. The problem of not achieving the desired effect is becoming more and more serious with the increase of application landing.

SUMMARY OF THE INVENTION

The present invention provides a gait big data retrieval method, system and terminal equipment, which can realize that users can use the gait retrieval technology in a standardized and correct manner, and obtain the best retrieval effect, so that the target person can be drawn based on the retrieval results in a richer and more accurate manner. Track information to solve the problem that the retrieval effect varies from person to person.

In one aspect, the present invention provides a gait big data retrieval method, the method comprising:

S1, acquiring a video image containing a target person, and extracting a gait sample of the target person in the video image, and constructing a gait library of the target person according to the gait sample;

S2, using the gait database to retrieve the monitoring video data in the local area to obtain a local retrieval result, and using the local retrieval result to update the gait database to obtain an updated gait database;

S3, utilize the updated described gait library to carry out retrieval in the surveillance video data of the whole domain, obtain the global retrieval result;

S4. Generate an action trajectory of the target person according to the local retrieval result and the global retrieval result.

Optionally, the S2 specifically includes:

S21. Obtain the initial time and place where the target character appears as the center point;

S22, using the gait library to perform retrieval in surveillance video data within a preset range around the center point to obtain a local retrieval result;

S23, using the local retrieval result to update the gait database to obtain an updated gait database;

S24, taking the time and place where the target character newly appears in the local retrieval result as the updated center point;

S25. Repeat S22 to S24 until no newly added trajectory information of the target person is retrieved in the local retrieval result or it is retrieved that the target person leaves the local retrieval area.

Optionally, after S3, the method further includes:

S5, determine whether the newly added trajectory information of the target person is retrieved in the global retrieval result; if so, execute step S6, if not, execute step S4;

S6. Use the time and place in the newly added trajectory information of the target person in the global retrieval result as the updated center point, and execute step S22.

Optionally, the S23 specifically includes:

The gait samples that conform to the preset rules in the local retrieval results are added to the gait database as excellent gait samples to obtain an updated gait database.

Optionally, the surveillance video data within a preset range around the center point is all surveillance video data within a range of 30 minutes before and after the center point, radiating 3 kilometers outward from the center point.

Optionally, the S22 is specifically:

S221, using the gait library as a comparison object, and performing gait comparison with the monitoring video data within a preset range around the center point to obtain the comparison similarity of each piece of monitoring video data;

S222. Use the gait samples whose comparison similarity is greater than or equal to a preset threshold and their corresponding time and place as the local retrieval result.

Optionally, the S4 is specifically:

All the time and places where the target person appears in the local retrieval result and the global retrieval result are sorted by time to form the action track of the target person.

In another aspect, the present invention provides a gait big data retrieval system, the system comprising:

A gait library construction unit, configured to acquire a video image containing a target person, extract a gait sample of the target person from the video image, and construct a gait library of the target person according to the gait sample;

a local retrieval unit, configured to use the gait database to retrieve the monitoring video data in a local area to obtain a local retrieval result, and to update the gait database using the local retrieval result to obtain an updated gait database;

a global retrieval unit, used for using the updated gait library to retrieve the surveillance video data of the whole domain to obtain a global retrieval result;

A trajectory generation unit, configured to generate an action trajectory of the target person according to the local retrieval result and the global retrieval result.

Optionally, the system further includes a data query unit and a data synchronization unit;

The data query unit is configured to query the data in the gait database according to any one or any of several of the database type, creation time, gait sample level and the identity information of the collected person;

The data synchronization unit is used for synchronizing the data in the gait library to an upper-level system or a third-party system according to a preset synchronization cycle or a preset synchronization strategy.

In yet another aspect, the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor;

The processor implements the steps of any of the methods described above when executing the computer program.

The beneficial effects that the present invention can produce include:

(1) The gait big data retrieval method provided by the present invention can dynamically build a database in the case of insufficient gait samples, enrich the number of target person gait samples, and improve the quality of target person gait samples. For example, when the registered target person's gait sample has low pixels and the human figure outline is blurred, other excellent gait samples can be quickly found through local retrieval with a low threshold.

(2) In the gait big data retrieval method provided by the present invention, more time and place information of the appearance of the target person can be quickly obtained through the global retrieval, and the missed trajectory of the target person in the local range can be avoided through the local retrieval, so that the detailed information can be obtained. It can draw a large-scale target character's action trajectory.

Description of drawings

1 is a flowchart of a method for retrieving gait big data provided by an embodiment of the present invention;

2 is a schematic diagram of the principle of a gait big data retrieval method provided by an embodiment of the present invention;

3 is a flowchart of a method for retrieving gait big data provided by another embodiment of the present invention;

FIG. 4 is a schematic diagram of a running process of helper retrieval according to an embodiment of the present invention.

Detailed ways

The present invention will be described in detail below with reference to the examples, but the present invention is not limited to these examples.

The embodiment of the present invention provides a gait big data retrieval method, as shown in FIG. 1 and FIG. 2 , the method includes:

S1. Acquire a video image containing a target person, extract a gait sample of the target person from the video image, and construct a gait library of the target person according to the gait sample.

Among them, the gait sample can be one or more continuous gait image sequences including gait features; the gait database can also be called the gait base library; the function of step S1 is to register the base library, specifically, it will contain target person information The video is uploaded to the system, and the gait samples of the target person are extracted through the gait algorithm to form the gait library of the target person.

S2. Use the gait database to retrieve the surveillance video data in the local area to obtain a local retrieval result, and use the local retrieval result to update the gait database to obtain an updated gait database.

The S2 specifically includes:

S21. Obtain the initial time and place where the target character appears as a center point. The initial time and place where the target person appears may be the time and place where the target person appears with a high probability inferred by the public security personnel based on experience, or the time and place inferred from some relevant evidence materials. This embodiment of the present invention does not limit this.

S22. Use the gait library to search the surveillance video data within a preset range around the center point to obtain a local search result.

The S22 specifically includes:

S221. Using the gait library as a comparison object, perform gait comparison with surveillance video data within a preset range around the center point to obtain the comparison similarity of each piece of surveillance video data.

S222. Use the gait samples whose comparison similarity is greater than or equal to a preset threshold and their corresponding time and place as the local retrieval result.

The preset threshold is a value preset by a person skilled in the art according to an actual situation, which is not limited in this embodiment of the present invention, and may be 80%, 85%, 90%, etc. in an example.

The preset range is a preset time length and geographic range, which can be set by a person skilled in the art or a user according to an actual situation, which is not limited in this embodiment of the present invention. For example, the surveillance video data within a preset range around the center point may be all surveillance video data within a range of 30 minutes before and after the center point radiates 3 kilometers from the center point.

In practical applications, step S22 can also be called helper retrieval; take the time and place where the target person has a high probability of appearing as the center point, the location radiates 3 kilometers outward, and expands 30 minutes before and after the time to select the relevant monitoring video data, and the target person step by step The gait database is used as the comparison object, the threshold of gait comparison is set to 80%, the comparison results are set to be sorted from high to low by gait comparison similarity, and gait retrieval is performed under the above parameter configuration. The above functions can be fixed in the software system to realize one-key retrieval.

S23, updating the gait database by using the local retrieval result to obtain the updated gait database.

Specifically: adding the gait samples that conform to the preset rules in the local retrieval results as excellent gait samples into the gait database to obtain the updated gait database.

The preset rule is a preset determination rule, which can be set by those skilled in the art according to actual conditions, which is not limited in this embodiment of the present invention. For example, the preset rules for selecting excellent gait samples may be that the gait samples show the natural walking state of a complete humanoid, the pixels of the gait samples are greater than or equal to 60*100, the number of frames of the gait samples is greater than or equal to 15, and the gait samples include 2 The above gait cycle, the humanoid outline in the gait sample has a clear boundary

S24. Use the time and place where the target person newly appears in the partial retrieval result as the updated center point.

S25. Repeat S22 to S24 until no newly added trajectory information of the target person is retrieved in the local retrieval result or until the target person leaves the local retrieval area.

Take the time and place where the target person has a high probability of appearing as the center point to select all relevant surveillance video data for helper retrieval, select excellent gait samples from the retrieval results and add them to the gait database of the target person, and add the time and place of the new appearance of the target person in the retrieval result. As a new center point, the updated target person gait library is used as a new target comparison object, and all relevant surveillance video data are selected again for helper retrieval. When there is no new search result (that is, no new track information) in the local area through the helper search, or the target person is found to leave by means of a vehicle through the search result, it is determined that the partial search is over.

S3. Use the updated gait database to perform retrieval in the surveillance video data of the whole domain to obtain a global retrieval result; wherein, the surveillance video data of the whole domain is to select all relevant surveillance video data in full.

Further, after S3, the method further includes:

S5. Determine whether the newly added trajectory information of the target person is retrieved in the global retrieval result; if yes, execute step S6, if not, execute step S4.

S6. Use the time and place in the newly added trajectory information of the target person in the global retrieval result as the updated center point, and execute step S22.

Taking the updated gait database as the target comparison object, all relevant surveillance video data are selected for gait retrieval, and the new time and place where the target person appears in the retrieval result is used as the entry point for further local retrieval.

S4. Generate an action trajectory of the target person according to the local retrieval result and the global retrieval result.

Specifically: sort all the times and places where the target person appears in the local retrieval result and the global retrieval result in chronological order to form the action trajectory of the target person.

Another embodiment of the present invention provides a gait big data retrieval method, as shown in FIG. 3 and FIG. 4 , the method includes:

S301. Acquire a video image containing a target person, extract a gait sample of the target person from the video image, and construct a gait library of the target person according to the gait sample.

S302: Obtain the initial time and place where the target character appears as the center point.

S303 , using the gait library to search the surveillance video data within a preset range around the center point to obtain a local search result.

S304 , adding the gait samples conforming to the preset rules in the local retrieval results as excellent gait samples into the gait database to obtain an updated gait database.

S305: Determine whether the local retrieval result includes the time and place where the target person newly appears; if so, go to step 306; if not, go to step 307.

S306 , take the time and place where the target person newly appears in the partial retrieval result as the updated center point, and execute step 303 .

S307 , using the updated gait database to perform retrieval in the surveillance video data of the whole domain to obtain a global retrieval result.

S308: Determine whether the global retrieval result includes the time and place where the target person newly appears; if yes, go to step 309; if not, go to step 310.

S309 , take the time and place where the target person newly appears in the global retrieval result as the updated center point, and execute step 303 .

S310: Sort all the times and places where the target person appears in the local retrieval result and the global retrieval result by time to form an action track of the target person.

The present invention can dynamically build a database in the case of insufficient gait samples, enrich the quantity of target person's gait samples, and improve the quality of target person's gait samples. For example, when the registered target person's gait sample has low pixels and the human figure outline is blurred, other excellent gait samples can be quickly found through local retrieval with a low threshold. At the same time, more time and place information of the target person's appearance can be quickly obtained through global retrieval, and the missed trajectory of the target person in the local range can be avoided through local retrieval, so that the movement trajectory of a large-scale target person can be refined and drawn.

Another embodiment of the present invention provides a gait big data retrieval system, the system includes:

The gait library construction unit is used to obtain a video image containing a target person, extract a gait sample of the target person from the video image, and construct a gait library of the target person according to the gait sample.

The local retrieval unit is used to use the gait database to retrieve the surveillance video data in the local area to obtain the local retrieval result, and use the local retrieval result to update the gait database to obtain the updated gait database.

The global retrieval unit is used to retrieve the global surveillance video data by using the updated gait library to obtain the global retrieval result.

The trajectory generation unit is used to generate the action trajectory of the target person according to the local retrieval result and the global retrieval result.

Further, the system also includes a data query unit and a data synchronization unit;

The data query unit is used to query the data in the gait database according to any one or several of the database type, creation time, gait sample level and the identity information of the collected person, so as to browse the identity of the personnel in the base library, gait samples, etc. information.

The data synchronization unit is used to synchronize the data in the gait library to the upper-level system or the third-party system according to the preset synchronization cycle or the preset synchronization strategy.

The retrieval system provided by the embodiment of the present invention has functions such as helper retrieval, gait retrieval, gait comparison, retrieval result summary, trajectory drawing, and gait base library quality assessment.

The retrieval system provided by the embodiment of the present invention has the functions of user management, role management, and organization management; has different use rights management functions of ordinary users and administrators, and has a relevant authorization mechanism in operations such as gait registration, deletion, and query. At the same time, the retrieval system of the present invention also has a log management function. For each event, the log record should include event occurrence time, event type, user, event execution result or failure reason, log valid time, and the like.

Yet another embodiment of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements any of the above methods when executing the computer program A step of.

The above are only a few embodiments of the present application, and are not intended to limit the present application in any form. Although the present application is disclosed as above with preferred embodiments, it is not intended to limit the present application. Without departing from the scope of the technical solution of the present application, any changes or modifications made by using the technical content disclosed above are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims (10)

1. a gait big data retrieval method, is characterized in that, described method comprises: S1, acquiring a video image containing a target person, and extracting a gait sample of the target person in the video image, and constructing a gait library of the target person according to the gait sample; S2, using the gait database to retrieve the monitoring video data in the local area to obtain a local retrieval result, and using the local retrieval result to update the gait database to obtain an updated gait database; S3, utilize the updated described gait library to carry out retrieval in the surveillance video data of the whole domain, obtain the global retrieval result; S4. Generate an action trajectory of the target person according to the local retrieval result and the global retrieval result. 2. The method according to claim 1, wherein the S2 specifically comprises: S21. Obtain the initial time and place where the target character appears as the center point; S22, using the gait library to perform retrieval in surveillance video data within a preset range around the center point to obtain a local retrieval result; S23, using the local retrieval result to update the gait database to obtain an updated gait database; S24, taking the time and place where the target character newly appears in the local retrieval result as the updated center point; S25. Repeat S22 to S24 until no newly added trajectory information of the target person is retrieved in the local retrieval result or it is retrieved that the target person leaves the local retrieval area. 3. The method according to claim 2, wherein after S3, the method further comprises: S5, determine whether the newly added trajectory information of the target person is retrieved in the global retrieval result; if so, execute step S6, if not, execute step S4; S6. Use the time and place in the newly added trajectory information of the target person in the global retrieval result as the updated center point, and execute step S22. 4. The method according to claim 2, wherein the S23 specifically comprises: The gait samples that conform to the preset rules in the local retrieval results are added to the gait database as excellent gait samples to obtain an updated gait database. 5. The method according to claim 2, wherein the monitoring video data in the preset range around the center point is the center point, the location radiates 3 kilometers outward, and the time is extended for 30 minutes before and after. All surveillance video data within range. 6. The method according to claim 2, wherein the S22 is specifically: S221, using the gait library as a comparison object, and performing gait comparison with the monitoring video data within a preset range around the center point to obtain the comparison similarity of each piece of monitoring video data; S222. Use the gait samples whose comparison similarity is greater than or equal to a preset threshold and their corresponding time and place as the local retrieval result. 7. The method according to claim 1, wherein the S4 is specifically: All the time and places where the target person appears in the local retrieval result and the global retrieval result are sorted by time to form the action track of the target person. 8. A gait big data retrieval system, wherein the system comprises: A gait library construction unit, configured to acquire a video image containing a target person, extract a gait sample of the target person from the video image, and construct a gait library of the target person according to the gait sample; a local retrieval unit, configured to use the gait database to retrieve the monitoring video data in a local area to obtain a local retrieval result, and to update the gait database using the local retrieval result to obtain an updated gait database; a global retrieval unit, used for using the updated gait library to retrieve the surveillance video data of the whole domain to obtain a global retrieval result; A trajectory generation unit, configured to generate an action trajectory of the target person according to the local retrieval result and the global retrieval result. 9. The system according to claim 8, wherein the system further comprises a data query unit and a data synchronization unit; The data query unit is configured to query the data in the gait database according to any one or several of the database type, creation time, gait sample level and the identity information of the collected person; The data synchronization unit is used for synchronizing the data in the gait library to an upper-level system or a third-party system according to a preset synchronization cycle or a preset synchronization strategy. 10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the method according to any one of claims 1 to 7 when executing the computer program.

Images