CN112380892A

CN112380892A - Image identification method, device, equipment and medium

Info

Publication number: CN112380892A
Application number: CN202010958787.8A
Authority: CN
Inventors: 卢超; 王健; 王雯雯; 史世莲
Original assignee: Hisense TransTech Co Ltd
Current assignee: Hisense TransTech Co Ltd
Priority date: 2020-09-14
Filing date: 2020-09-14
Publication date: 2021-02-19
Anticipated expiration: 2040-09-14
Also published as: CN112380892B

Abstract

The invention discloses an image identification method, device, equipment and medium, which are used for solving the problems that monitoring and identification intellectualization of illegal personnel is not high and identification efficiency of the illegal personnel is low in the prior art. The method comprises the following steps: identifying pedestrians appearing in any image acquired within a preset time length, if the number of times of any pedestrian appearing within the preset time length is larger than a set number threshold, determining the pedestrian as a suspicious pedestrian, and determining the motion trail of the suspicious pedestrian according to the image acquired within the preset time length; if the similarity between the motion track of the suspicious pedestrian and any one of the stored standard tracks meets a preset condition, the suspicious pedestrian is used as the identified target pedestrian, so that the automatic identification of the suspicious pedestrian can be realized, and the target pedestrian is finally determined.

Description

Image identification method, device, equipment and medium

Technical Field

The present invention relates to the field of image processing technologies, and in particular, to an image recognition method, an image recognition device, an image recognition apparatus, and a medium.

Background

With the rapid development of vehicles, people who take taxies, drip express trains and the like are more and more in daily travel, but more and more offenders and offending vehicles attempt to collect passengers illegally to obtain benefits, and more offending vehicles and offending vehicles are combined to collect passengers, so that the difficulty in checking and treating is increased.

In the prior art, the management of the offenders mainly depends on a method for regular renovation and field identification by law enforcement management departments, and the offenders are identified only by the existing offender library recorded manually without a feasible intelligent monitoring method, so that the monitoring and identification intellectualization of the offenders are not high, and the identification efficiency of the offenders is low.

Disclosure of Invention

The invention provides an image identification method, device, equipment and medium, which are used for solving the problems that monitoring and identification intellectualization of illegal personnel is not high and identification efficiency of the illegal personnel is low in the prior art.

In a first aspect, the present invention provides an image recognition method, including:

identifying pedestrians appearing in any image acquired within a preset time length, and if the number of times of any pedestrian appearing within the preset time length is larger than a set number threshold, determining the pedestrian as a suspicious pedestrian;

determining the motion trail of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length;

and if the similarity between the motion trail of the suspicious pedestrian and any one of the stored standard trails meets a preset condition, taking the suspicious pedestrian as the identified target pedestrian.

Further, if the number of occurrences of any pedestrian within the preset time length is greater than a set number threshold, the method further includes:

judging whether an image matched with the face image of the pedestrian exists in a face database according to the face database of the existing suspicious pedestrian and the face database of the working personnel, wherein the face database is stored in advance;

if not, then the subsequent step of taking the pedestrian as the suspicious pedestrian is carried out.

Further, the method further comprises:

and if the occurrence frequency of any pedestrian in the preset time length is not more than a set frequency threshold value, or an image matched with the face image of the pedestrian exists in the face database, determining the pedestrian as a non-suspicious pedestrian.

Further, before determining the motion trajectory of the suspicious pedestrian within the preset time period according to the image acquired within the preset time period, the method further includes:

determining that the target image of the suspicious pedestrian is contained in the images acquired within the preset time length;

determining other pedestrians in the same row with the suspicious pedestrian according to the distance between the suspicious pedestrian and other pedestrians in the target image;

and if the number of other pedestrians on the same row with the suspicious pedestrian within the preset time length is larger than a preset first number threshold, performing a subsequent step of determining the motion trail of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length.

Further, the method further comprises:

and if the number of other pedestrians in the same row with the suspicious pedestrian within the preset time length is not larger than a preset first number threshold, determining the pedestrian as a non-suspicious pedestrian.

Further, the method further comprises:

identifying a vehicle appearing in any image acquired within a preset time span;

if the number of times of any vehicle appearing in the preset time length is larger than a set second number threshold, the vehicle is taken as a suspicious vehicle;

determining that the images acquired within the preset time span contain a target image of the target pedestrian, and determining whether the target pedestrian and the suspicious vehicle in the target image exist at the same time;

and if the number of times that the target pedestrian and the suspicious vehicle coexist within the preset time length is greater than a preset third quantity threshold value, taking the suspicious vehicle as the target vehicle.

Further, if the number of times of occurrence of any vehicle within the preset time period is greater than a set second number threshold, then taking the vehicle as a suspicious vehicle, includes:

judging whether a vehicle matched with the vehicle exists in a vehicle database according to the vehicle database of the existing suspicious vehicle and the vehicle database of the legal operation vehicle which are stored in advance;

if not, the subsequent step of taking the vehicle as a suspicious vehicle is carried out.

In a second aspect, the present invention also provides an image recognition apparatus, comprising:

the determination module is used for identifying pedestrians appearing in any image acquired within a preset time length, and if the number of times of any pedestrian appearing within the preset time length is larger than a set number threshold, determining the pedestrian as a suspicious pedestrian;

the determining module is further configured to determine a motion trajectory of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length;

and the processing module is used for taking the suspicious pedestrian as the identified target pedestrian if the similarity between the motion track of the suspicious pedestrian and any one of the stored standard tracks meets a preset condition.

Further, the determining module is further configured to determine whether an image matched with the face image of the pedestrian exists in the face database according to a pre-stored face database of the existing suspicious pedestrian and the worker; if not, then the subsequent step of taking the pedestrian as the suspicious pedestrian is carried out.

Further, the determining module is further configured to determine that the pedestrian is a non-suspicious pedestrian if the number of times that any pedestrian appears within the preset time length is not greater than a set number threshold, or an image matching the face image of the pedestrian exists in the face database.

Further, the determining module is further configured to determine that the target image of the suspicious pedestrian is included in the images acquired within the preset time period;

Further, the determining module is further configured to determine the pedestrian as a non-suspicious pedestrian if the number of other pedestrians that are in the same row as the suspicious pedestrian within the preset time duration is not greater than a preset first number threshold.

Further, the apparatus further comprises:

the identification module is specifically used for identifying a vehicle appearing in any image acquired within a preset time span;

the processing module is further configured to take any vehicle as a suspicious vehicle if the number of times that any vehicle appears within the preset time length is greater than a set second number threshold;

the judging module is used for determining that the target image of the target pedestrian is contained in the image acquired within the preset time length and determining whether the target pedestrian and the suspicious vehicle exist simultaneously in the target image;

the processing module is further configured to take the suspicious vehicle as the target vehicle if the number of times that the target pedestrian and the suspicious vehicle coexist within the preset time length is greater than a preset third number threshold.

Further, the identification module is specifically configured to determine whether a vehicle matching the vehicle exists in the vehicle database according to a vehicle database of pre-stored existing suspicious vehicles and legal operating vehicles; if not, the subsequent step of taking the vehicle as a suspicious vehicle is carried out.

In a third aspect, the present invention also provides an electronic device comprising a processor configured to implement the steps of the image recognition method as described in any one of the above when executing a computer program stored in a memory.

In a fourth aspect, the present invention also provides a computer-readable storage medium, which stores a computer program, which when executed by a processor implements the steps of the image recognition method as described in any one of the above.

In the invention, a pedestrian appearing in an image is identified aiming at any image acquired within a preset time length, if the number of times of any pedestrian appearing within the preset time length is greater than a set number threshold, the pedestrian is determined as a suspicious pedestrian, and the motion trail of the suspicious pedestrian is determined according to the image acquired within the preset time length; if the similarity between the motion track of the suspicious pedestrian and any one of the stored standard tracks meets a preset condition, the suspicious pedestrian is used as the identified target pedestrian, so that the automatic identification of the suspicious pedestrian can be realized, and the target pedestrian is finally determined.

Drawings

In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.

FIG. 1 is a schematic diagram of an image recognition process according to some embodiments of the invention;

FIG. 2 is a flow chart of an automatic suspicious pedestrian identification process according to some embodiments of the present invention;

fig. 3 is a flow chart of automatic identification of offenders according to some embodiments of the present invention.

FIG. 4 is a flow chart illustrating automatic identification of suspicious vehicles according to some embodiments of the present invention;

FIG. 5 is a diagram illustrating an exemplary automatic identification process for an offending vehicle in accordance with certain embodiments of the present invention;

FIG. 6 is a physical structure for automatic identification of yellow vehicles and offending vehicles according to some embodiments of the present invention;

FIG. 7 is a physical block diagram of an automatic yellow and offending vehicle identification provided by some embodiments of the present invention;

fig. 8 is a schematic structural diagram of an image recognition apparatus according to some embodiments of the present application;

fig. 9 is a schematic structural diagram of an electronic device according to some embodiments of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be derived from the embodiments of the present invention by a person skilled in the art are within the scope of the present invention.

In order to realize intelligent identification of suspicious pedestrians and improve efficiency of suspicious pedestrian identification, the embodiment of the invention provides an image identification method, an image identification device, image identification equipment and an image identification medium.

Example 1:

fig. 1 is a schematic diagram of an image recognition process according to some embodiments of the present invention, where the process includes the following steps:

s101: and identifying the pedestrian appearing in the image aiming at any image acquired within a preset time length, and if the number of times of any pedestrian appearing within the preset time length is greater than a set number threshold, determining the pedestrian as a suspicious pedestrian.

The image recognition method provided by the embodiment of the invention is applied to the electronic equipment, and the electronic equipment can be a back-end server, and can also be electronic equipment capable of performing image recognition such as a PC (personal computer).

In order to realize the identification of the pedestrian in the image, the image to be identified can be acquired based on the image acquisition device. In order to acquire the information of suspicious pedestrians as much as possible, an image acquisition device is pre-installed in an area needing to be monitored, and the image acquisition device is used for acquiring images of the acquisition area. The image identification method can be used for identifying suspicious pedestrians, wherein the suspicious pedestrians can be illegal persons who illegally pick up passengers, thieves who steal in crowd-dense areas and the like. If the suspicious pedestrian is an illegal person who illegally takes in a passenger in an airport or a railway station, the installation area of the image acquisition equipment is generally selected as an area where the activities of the illegal person frequently appear in a landing exit, a taxi boarding area, a transfer area and a social parking lot.

In the embodiment of the invention, after the image acquisition device acquires each frame of image to be identified, the electronic device can identify the pedestrians existing in any image acquired within the preset time span. Specifically, each pedestrian in the image can be identified by a face recognition technology, which is the prior art and is not described herein again.

Counting the occurrence frequency of each pedestrian in a preset time length according to the pedestrian identified in each image acquired in the preset time length, determining the occurrence frequency of the pedestrian in the preset time length, and determining the pedestrian as a suspicious pedestrian if the occurrence frequency is greater than a frequency threshold. The preset time length is preset, and the preset time length can be set according to requirements, for example, the preset time length can be days or half a month, and it is required to ensure that the image acquisition equipment can fully acquire the image containing the suspicious pedestrian within the preset time length.

The number threshold is preset, the size of the number threshold can be set according to requirements, if a target pedestrian is identified in a short time, the number threshold can be set slightly larger, and if the identification accuracy of a suspicious pedestrian is improved, the number can be set slightly smaller.

S102: and determining the motion trail of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length.

After the suspicious pedestrian is determined, in order to determine whether the suspicious pedestrian is the target pedestrian, the suspicious pedestrian may be further determined, in the embodiment of the present invention, the track of the suspicious pedestrian may be obtained, and based on the track of the suspicious pedestrian, whether the suspicious pedestrian is the target pedestrian is determined.

Specifically, when the track of the suspicious pedestrian is obtained, each image of the suspicious pedestrian appearing in the preset time length is determined according to the image obtained in the preset time length, the position of the suspicious pedestrian in the image is determined according to each obtained image of the suspicious pedestrian appearing, and each specific geographic position of the suspicious pedestrian is determined according to the position of the suspicious pedestrian in the image and the installation position of each image acquisition device. After each specific geographic position where the suspicious pedestrian appears is determined for each image, the motion trail of the suspicious pedestrian is determined according to the time sequence of the acquisition of each image.

The geographic location may be the physical location where the pedestrian is actually located, or may be the installation location of each image capturing device that captures an image of the presence of the suspicious pedestrian.

S103: and if the similarity between the motion trail of the suspicious pedestrian and any one of the stored standard trails meets a preset condition, taking the suspicious pedestrian as the identified target pedestrian.

In order to accurately judge whether a suspicious pedestrian is a target pedestrian, a plurality of standard tracks are stored in the electronic device in advance, wherein the standard tracks are motion tracks which may appear in the target pedestrian within a preset time length.

When determining whether the suspicious pedestrian is the target pedestrian, the similarity between the trajectory of the suspicious pedestrian within the preset time length and any one of the stored standard trajectories can be determined, wherein when determining the similarity, the euclidean distance between the two trajectories can be determined, the determined euclidean distance is used as the similarity between the two trajectories, and the euclidean distance can be specifically determined by the following formula:

wherein the content of the first and second substances,

the movement locus of the suspicious pedestrian is the movement locus,

the ith sample of the suspicious pedestrian is the position information of the suspicious pedestrian collected at the time t of the set time length, and can be specifically replaced by the position information of the monitoring point where the suspicious pedestrian is collected,

the standard trajectory is a standard trajectory in which,

the sample is the ith sample of the pedestrian corresponding to the standard track, and the sample is the position information of the pedestrian collected at the time t of the set time length.

For any standard track, after the similarity between the motion track of the suspicious pedestrian and the standard track is determined, judging whether the similarity meets a preset condition, specifically, whether the similarity is not greater than a preset similarity threshold value Y, and when d is greater than a preset similarity threshold value Y_iWhen the motion track of the suspicious pedestrian is not larger than Y, the motion track of the suspicious pedestrian is determined to be very close to the standard track, the similarity is determined to meet the preset condition, and the suspicious pedestrian is determined to be the target pedestrian.

S104: and if the similarity between the motion trail of the suspicious pedestrian and any one of the stored standard trails does not meet the preset condition, determining the pedestrian as a non-suspicious pedestrian.

Comparing the similarity of the motion trail of the suspicious pedestrian with all standard trails stored in the electronic equipment, and if the similarity of the motion trail of the suspicious pedestrian and all standard trails is larger than a preset similarity threshold value, namely d_i>And Y, determining that the motion trail of the suspicious pedestrian is not close to all standard trails stored in the electronic equipment, determining that the similarity between the motion trail of the suspicious pedestrian and all standard trails stored in the electronic equipment does not meet a preset condition, and determining the pedestrian as a non-suspicious pedestrian.

Example 2:

in order to improve the accuracy of identifying a suspicious pedestrian, on the basis of the above embodiment, in an embodiment of the present invention, after the number of occurrences of any pedestrian within the preset time length is greater than a set number threshold, the method further includes:

judging whether a face image matched with the face image of the pedestrian exists in a face database according to the face database of the existing suspicious pedestrian and the face database of the working personnel, which is stored in advance;

In the embodiment of the invention, in order to improve the accuracy of suspicious pedestrian identification, when whether a pedestrian is a suspicious pedestrian is judged, firstly, whether the number of times of the pedestrian appearing in a preset time length is greater than a set number threshold is judged for any pedestrian appearing in an image, and if the number of times of the pedestrian appearing in the preset time length is determined to be greater than the set number threshold, the pedestrian is determined to frequently appear in a monitoring area and has certain suspicious property, but the pedestrian frequently appears and may be a worker working nearby or a pedestrian already determined as a suspicious person, namely, the suspicious pedestrian exists.

Therefore, in order to determine whether the pedestrian is a newly added suspicious pedestrian, the electronic device according to the embodiment of the present invention pre-stores face databases of the staff and the existing suspicious pedestrian, wherein the face image of the staff may be placed in one face database, the face image of the existing suspicious pedestrian may be placed in another face database, or the face image of the suspicious pedestrian and the face image of the staff may be stored in one face database. After the number of times of occurrence of the pedestrian within the preset time length is determined to be larger than a number threshold, a face image of the pedestrian is obtained based on the image of the pedestrian, whether a face image matched with the face image of the pedestrian exists in the face database is determined based on the face database, if the face image matched with the face image of the pedestrian does not exist, the pedestrian is determined to be a suspicious pedestrian according to the fact that the pedestrian is not a worker or the suspicious pedestrian exists, and then the pedestrian is determined to be the suspicious pedestrian.

For example, when the preset time length is 7 days, a pedestrian appearing in the image is identified for the image acquired within the 7 days. And when the threshold value of the times is set to be 5 times, if the times of occurrence of a certain pedestrian in 7 days is more than 5 times, determining whether a face image matched with the face image of the pedestrian exists in the face database according to the face database of the prestored staff and the existing suspicious pedestrian, and if not, determining the pedestrian as the suspicious pedestrian.

In order to accurately determine a suspicious pedestrian, on the basis of the foregoing embodiments, in an embodiment of the present invention, the method further includes:

and if the occurrence frequency of any pedestrian in the preset time length is not more than a set frequency threshold value, or a face image matched with the face image of the pedestrian exists in the face database, determining the pedestrian as a non-suspicious pedestrian.

In the embodiment of the invention, the face recognition is carried out on the images collected within the preset time length, each pedestrian appearing within the preset time length is determined, the number of times of appearance within the preset time length is counted for each pedestrian, and if the number of times of appearance of any pedestrian within the preset time length is not more than the set threshold value of times, the pedestrian is not frequently appeared and is likely to normally pass through the monitoring scene, so that the pedestrian can be determined to be a normal behavior rather than a suspicious pedestrian.

In addition, if the number of times that a certain pedestrian appears within the preset time length is greater than the set number threshold, it is indicated that the pedestrian is a person frequently appearing in the monitoring scene, and may be a suspicious pedestrian, but may also be a worker working in the monitoring scene, or an existing suspicious pedestrian, in order to reduce the workload of subsequently identifying the suspicious pedestrian, the pedestrian already identified as the suspicious pedestrian may not need to be identified, and in order to ensure the accuracy of identification, the worker is prevented from being identified as the suspicious pedestrian, the face image of the pedestrian may be matched with the face image in the face database saved in advance, and if the face image matched with the face image of the pedestrian exists in the face database, the pedestrian may be determined as a suspicious pedestrian not identified at this time. Specifically, if the face image of the pedestrian is matched with the face image in the face database of the worker, it is determined that the pedestrian is the worker, and if the face image of the pedestrian is matched with the face image in the face database of the suspicious pedestrian, it is determined that the pedestrian is the suspicious pedestrian.

Example 3:

in order to improve the accuracy of identifying a suspicious pedestrian, on the basis of the foregoing embodiments, in an embodiment of the present invention, before determining the motion trajectory of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length, the method further includes:

The scheme provided by the embodiment of the invention is generally used for monitoring illegal persons, namely detected suspicious pedestrians can be illegal persons, the illegal persons can frequently appear in certain areas such as arrival floor exits, taxi boarding areas, transfer areas, social parking lots and the like, the sides of the illegal persons can possibly accompany a plurality of partners, and can possibly talk with a plurality of pedestrians to realize illegal passenger pickup, therefore, in order to improve the identification accuracy, whether the pedestrians are indeed suspicious pedestrians can be further determined according to the number of the pedestrians who accompany the suspicious pedestrians.

Therefore, when a certain pedestrian is determined to be a suspicious pedestrian, in the embodiment of the present invention, a target image including the suspicious pedestrian in the acquired image is determined according to the image acquired within the preset time length. Specifically, the position of the suspicious pedestrian can be the position of a circumscribed rectangle corresponding to the face of the suspicious pedestrian or the position of a circumscribed rectangle corresponding to the body of the suspicious pedestrian, and the positions of other pedestrians in the target image are determined. It is sufficient that the position of the pedestrian is determined in the same manner when the detection is performed.

After the positions of the suspicious pedestrian and each other pedestrian in the target image are determined, the distance is calculated based on the center of the circumscribed rectangle corresponding to each pedestrian, and therefore the distance between the suspicious pedestrian and the other pedestrians in the target image is determined. Of course, the distance between the suspicious pedestrian and other pedestrians may be determined based on the coordinates of other positions of the circumscribed rectangle, for example, the position of the vertex at the upper left corner of the circumscribed rectangle, or the position of the vertex at the lower right corner of the circumscribed rectangle.

Because the position of the image acquisition equipment is fixed, the acquisition range is fixed, and the distance between every two pixel points in the acquired image is also determined. Therefore, when the distance between the suspicious pedestrian and other pedestrians in the image is determined, the actual physical distance between the suspicious pedestrian and other pedestrians can be determined based on the distance.

Based on each target image acquired within a preset time length, the distance between the corresponding number of the suspicious pedestrian and the other pedestrians can be determined according to the number of the suspicious pedestrian and the number of the other pedestrians in each target image. Each distance between the suspicious pedestrian and the other pedestrians is compared with a preset distance threshold value in size.

When the distance between the suspicious pedestrian and other pedestrians is smaller than the preset distance threshold value, the fact that the suspicious pedestrian and other pedestrians are in the same-row relationship is indicated, otherwise, the fact that the suspicious pedestrian and other pedestrians are in the non-same-row relationship is determined, and therefore the number of the other pedestrians in the same row with the suspicious pedestrian and the number of the other pedestrians in the non-same row with the suspicious pedestrian in each target image are determined.

Determining the number of other pedestrians in the same row with the suspicious pedestrian in each target image according to each target image containing the suspicious pedestrian in the images acquired within the preset time length, determining the total number of the other pedestrians in the same row with the suspicious pedestrian in all the target images, comparing the total number with a preset first number threshold, and when the total number is larger than the preset first number threshold, performing the following steps of determining the motion track of the suspicious pedestrian according to the images acquired within the preset time length.

In order to improve the accuracy of identifying suspicious pedestrians, on the basis of the foregoing embodiments, in an embodiment of the present invention, the method further includes:

Determining the number of other pedestrians in the same row with the suspicious pedestrian in each target image according to each target image containing the suspicious pedestrian in the images acquired within the preset time length, determining the total number of the other pedestrians in the same row with the suspicious pedestrian in all the target images, comparing the total number with a preset first number threshold, and determining the pedestrian as a non-suspicious pedestrian when the total number is not greater than the preset first number threshold, namely, the subsequent step of determining the suspicious motion track of the pedestrian according to the images acquired within the preset time length is not necessary.

Fig. 2 is a flowchart of automatic suspicious pedestrian identification according to some embodiments of the present invention, including:

s201: the method comprises the steps of identifying a pedestrian appearing in an image according to any image acquired within a preset time length, and determining the number of times that the pedestrian appears in the image within the preset time length.

S202: and judging whether the number of times of the pedestrian appearing in the preset time length is larger than a preset number threshold, if so, performing S203, and otherwise, performing S209.

S203: and judging whether a face image matched with the face image of the pedestrian exists in the face database according to the face database of the existing suspicious pedestrian and the face database of the working personnel, if not, performing S204, otherwise, performing S209.

S204: and determining the pedestrian as a suspicious pedestrian, and determining the number of other pedestrians who are in the same row with the suspicious pedestrian within a preset time length.

S205: and judging whether the number of other pedestrians in the same row with the suspicious pedestrian within the preset time length is larger than a preset first number threshold, if so, performing S206, and if not, performing S209.

S206: and determining the motion trail of the suspicious pedestrian within the preset time length.

S207: and judging whether the similarity between the motion trail of the suspicious pedestrian and any one of the stored standard trails meets a preset condition, if so, performing S208, and otherwise, performing S209.

S208: the pedestrian is taken as the identified target pedestrian.

S209: the pedestrian is regarded as a non-suspicious pedestrian.

If the target pedestrian is an offender, the automatic identification of the offender is explained in detail, and fig. 3 is a flow chart of the automatic identification of the offender provided by some embodiments of the present invention.

The method comprises the steps of identifying pedestrians appearing in an image according to any image acquired within a preset time length, determining the number of times of any pedestrian appearing in the image within the preset time length, and enabling the number of times of any pedestrian appearing within the preset time length to be larger than a threshold value of the number of times and ensuring that the pedestrian is not a worker or is a pedestrian with a suspicious pedestrian as the suspicious pedestrian. Specifically, when the preset time length is 7 days, for the image acquired within the 7 days, the pedestrian appearing in the image is identified, when the set frequency threshold value is 5 times, the face image of the pedestrian meeting the condition that the number of times appearing within the 7 days is greater than 5 times is placed in the face library a, the face image of the worker is placed in the employee library B in advance, the face image of the suspicious person is placed in the suspected violation person library C, the face image of the person meeting the condition that the number of times appearing within the 7 days is greater than 5 but not the worker or the suspicious pedestrian is determined in the identified image is placed in the initial sample library S1, and the face image meeting the condition is placed in the initial sample library S1 ═ a- (a & B) - (a & C).

According to the number Ti of other pedestrians in the same row with the suspicious pedestrian within the preset time length, a first number Y1 is preset, pedestrians of which the face images meet Ti being more than or equal to Y1 in the pedestrians in the initial sample library S1 are counted, and the face images of the pedestrians of which the face images meet Ti being more than or equal to Y1 in the pedestrians corresponding to the face images in the sample library S1 are placed in a sample library S2, namely S2 ═ T_i≥Y1]。

Aiming at any standard offender which is stored in advance according to the motion trail of the suspicious pedestrian within the preset time lengthCalculating the similarity d between the track of the pedestrian in the face image in the sample library S2 and the track of the standard illegal person_iSetting a similarity threshold Y2 in advance, and placing face images satisfying di ≦ Y2 conditions in pedestrians corresponding to the face images in the sample library S2 in the sample library S3, wherein S3 ≦ Y2]The sample library S3 is eventually stored in a suspected offender library.

Example 4:

in order to achieve the identification of the target vehicle, on the basis of the above embodiments, in an embodiment of the present invention, the method further includes:

At present, vehicle identification is mainly applied to identification of illegal vehicles, in the prior art, differences between the illegal vehicles and legal operation vehicles are mainly mined based on big data accumulated by road gates, and the illegal vehicles and the legal operation vehicles are probably legal operation vehicles and illegal operation vehicles, however, the target vehicle identification method cannot provide effective law enforcement evidence, and identification intellectualization is not high.

In order to realize the identification of the vehicle in the image, the image to be identified can be acquired based on the image acquisition device. In order to acquire the information of suspicious vehicles as much as possible, an image acquisition device is pre-installed in an area to be monitored, and the image acquisition device is used for acquiring images of the acquisition area. The image acquisition equipment for acquiring the image of the vehicle can be the same as or different from the image acquisition equipment for identifying the suspicious pedestrian, and can be flexibly set according to the requirements. The installation area of the image acquisition equipment is generally a parking lot, and the area is an area with frequent illegal vehicle activities.

In the embodiment of the invention, after the image acquisition device acquires each frame of image to be identified, the electronic device can identify the vehicle in any image acquired within a preset time span. Specifically, each vehicle in the image can be identified by a vehicle identification technology, which is the prior art and will not be described herein.

According to the vehicle identified in each image collected within the preset time length, counting the occurrence frequency of each vehicle within the preset time length, determining the occurrence frequency of the vehicle within the set time length, and if the occurrence frequency is greater than a second number threshold, determining the vehicle as a suspicious vehicle.

The second quantity threshold is preset, the size of the second quantity threshold can be set according to requirements, if a target vehicle is identified in a short time, the second quantity threshold can be set slightly larger, and if the accuracy of identification of suspicious vehicles is improved, the second quantity threshold can be set slightly smaller.

Determining a target image containing a target pedestrian in the acquired image according to the image acquired within the preset time length, determining whether the target pedestrian and the suspicious vehicle appear simultaneously in the target image, and finally determining the number of times that the target pedestrian and the suspicious vehicle appear simultaneously in the target image within the set time length. And presetting a third quantity threshold, determining whether the number of times that the target pedestrian and the suspicious vehicle appear simultaneously in the target image within a set time length is greater than the third quantity threshold, and if so, taking the suspicious vehicle as the target vehicle. Otherwise, the suspect vehicle is determined to be a non-suspect vehicle.

When the number of times that the target pedestrian and the suspicious vehicle appear simultaneously is determined, the number of times that the target pedestrian and the suspicious vehicle appear simultaneously can be counted for each target pedestrian respectively, and judgment is performed for each number of times. However, it may also happen that different target pedestrians use the same suspicious vehicle to execute the violation, that is, multiple violating persons use one violation vehicle to execute the violation, and at this time, the total number of times that all target pedestrians and the suspicious vehicle appear simultaneously may be counted for all target pedestrians, and the determination is performed based on the total number of times.

Wherein, for a certain target pedestrian, the second number threshold is used for comparing with the number of times of appearance of the suspicious vehicle identified in the image within the set time length, and the third number threshold is used for comparing with the number of times of appearance of the target pedestrian and the suspicious vehicle in the target image within the set time length, in this case, the second number threshold is greater than or equal to the third number threshold.

Example 5:

in order to improve the accuracy of suspicious vehicle identification, on the basis of the foregoing embodiments, in an embodiment of the present invention, if the number of times that any vehicle appears within the preset time period is greater than the set second number threshold, the method for identifying a suspicious vehicle before taking the suspicious vehicle as a suspicious vehicle includes:

In the embodiment of the invention, in order to improve the accuracy of suspicious vehicle identification, when judging whether a vehicle is a suspicious vehicle, firstly, aiming at any vehicle appearing in an image, judging whether the number of times of the vehicle appearing in a preset time length is greater than a set second number threshold, and if the number of times of the vehicle appearing in the preset time length is greater than the set second number threshold, determining that the vehicle frequently appears in a monitoring area and has certain suspiciousness, but the frequently appearing vehicle may be a vehicle of a worker working nearby or a suspicious vehicle already determined, namely, the suspicious vehicle is already existed.

Therefore, in order to determine whether the vehicle is a newly added suspicious vehicle, the electronic device according to the embodiment of the present invention may store the vehicle of the operator and the vehicle database of the existing suspicious vehicle in advance, where the image of the vehicle of the operator may be placed in one vehicle database, the image of the existing suspicious vehicle may be placed in another vehicle database, or the image of the existing suspicious vehicle and the image of the vehicle of the operator may be stored in one vehicle database. After the number of times of the vehicle appearing within the preset time length is determined to be larger than a second number threshold, the image of the vehicle is obtained based on the image of the vehicle, whether the image of the vehicle matched with the image of the vehicle exists in the vehicle database is determined based on the vehicle database, if the image of the vehicle matched with the image of the vehicle does not exist, the vehicle is not a staff vehicle or a suspicious vehicle exists, and then the subsequent step of regarding the vehicle as a suspicious vehicle is carried out, namely the vehicle is determined to be the suspicious vehicle.

Fig. 4 is a flowchart of automatic suspicious vehicle identification according to some embodiments of the present invention, including:

s401: the method comprises the steps of identifying a pedestrian appearing in an image according to any image acquired within a preset time length, and determining the number of times that the pedestrian appears in the image within the preset time length.

S402: and judging whether the number of times of the pedestrian appearing in the preset time length is larger than a preset number threshold, if so, performing S403, and otherwise, performing S409.

S403: and judging whether a face image matched with the face image of the pedestrian exists in the face database according to the face database of the suspicious pedestrian and the face database of the working personnel, if not, performing S404, otherwise, performing S409.

S404: and determining the pedestrian as a suspicious pedestrian, and determining the number of other pedestrians who are in the same row with the suspicious pedestrian within a preset time length.

S405: and judging whether the number of other pedestrians in the same row with the suspicious pedestrian within the preset time length is larger than a preset first number threshold, if so, performing S406, and otherwise, performing S409.

S406: and determining the motion trail of the suspicious pedestrian within the preset time length.

S407: and judging whether the similarity between the motion track of the suspicious pedestrian and any one of the stored standard tracks meets a preset condition, if so, performing S408, and otherwise, performing S409.

S408: the pedestrian is taken as the identified target pedestrian.

S409: the pedestrian is regarded as a non-suspicious pedestrian.

S410: the method comprises the steps of identifying vehicles appearing in any image acquired within a preset time length, and determining the number of times that any vehicle appears in the image within the preset time length.

S410: and judging whether the number of times of the vehicle appearing in the preset time length is larger than a preset second number threshold, if so, performing S411, and otherwise, performing S415.

S411: and determining whether the vehicle database has the image of the vehicle matched with the image of the vehicle according to a vehicle database of the existing suspicious vehicle and the vehicle of the staff, if so, performing S412, and otherwise, performing S415.

S412: and taking the vehicle as a suspicious vehicle, and determining the number of times of simultaneous occurrence of the target pedestrian and the suspicious vehicle within a preset time length.

S413: and judging whether the number of times of simultaneous occurrence of the target pedestrian and the suspicious vehicle within the preset time length is greater than a preset third number threshold, if so, performing S414, and otherwise, performing S415.

S414: the vehicle is taken as the identified target vehicle.

S415: the vehicle is treated as a non-suspect vehicle.

If the target vehicle is an illegal vehicle, a method for identifying the illegal vehicle is explained in detail, and fig. 5 is a process diagram for automatically identifying the illegal vehicle according to some embodiments of the present invention.

The method comprises the steps of identifying vehicles appearing in any image acquired within a preset time length, determining the number of times of any vehicle appearing in the image within the preset time length by combining entrance and exit data of a social parking lot, presetting a second quantity threshold, and enabling the number of times of any vehicle appearing within the preset time length to be larger than the second quantity threshold and ensuring that the vehicle is not a worker vehicle or an existing suspicious vehicle or a vehicle operating legally as a suspicious vehicle. Specifically, when the preset time length is 7 days, the images are acquired within the 7 days, the vehicles appearing in the images are identified, and when the second number threshold is set to 5 times, the images of the vehicles satisfying the vehicles appearing more than 5 times within the 7 days are placed in the vehicle library a, placing the image of the vehicle of the staff vehicle in the staff vehicle library B in advance, placing the image of the vehicle operating legally in the vehicle library C, placing the image of the vehicle of the existing suspicious vehicle in the existing suspicious violation vehicle library D, determining that the number of times of occurrence of the vehicle identified in the image is more than 5 times within 7 days but not the image of the staff vehicle or the existing suspicious vehicle or the vehicle operating legally is satisfied, placing the image of the vehicle satisfying the condition in an initial sample library S4, wherein S4 ═ a- (a & B) - (a & C) - (a & D).

Counting the number of times that pedestrians corresponding to the face images in the sample library 3 and vehicles corresponding to the images of the vehicles in the sample library S4 appear simultaneously within a preset time length, namely counting the number of times that identified offenders and vehicles corresponding to the images of the vehicles in the sample library S4 appear simultaneously, presetting a third quantity threshold T, counting the vehicles of which the images meet Ts ≧ T in the vehicles in the initial sample library S4, placing the images of the vehicles of which Ts ≧ T in the vehicles corresponding to the images of the vehicles stored in the sample library S4 in the sample library S5, namely S5 [ Ts ≧ T ], and finally storing the sample library S5 in a suspected offending vehicle library.

The following description will be made of recognition of an offender and an offending vehicle, taking an example in which the target pedestrian is an offender and the target vehicle is an offending vehicle.

Fig. 6 is a physical structure of automatic identification of a yellow car and an illegal vehicle according to some embodiments of the present invention, and fig. 7 is a physical structure diagram of automatic identification of a yellow car and an illegal vehicle according to some embodiments of the present invention. Now that

The following description is made with reference to fig. 6 and 7:

firstly, a plurality of image acquisition devices are arranged in a distribution control area, the image acquisition devices acquire videos, and each frame of image in the videos acquired by the image acquisition devices is sent to an electronic device, wherein the electronic device can be a back-end processing server, and the back-end processing server respectively adopts a face recognition technology and a vehicle recognition technology to recognize pedestrians and vehicles. After the illegal person is identified, the newly identified illegal person is expanded into a suspected illegal person library, after the illegal person is identified, all images of the illegal person appearing within a preset time length are used as target images, when the number of times that the illegal person and a suspicious illegal vehicle appear simultaneously in the target images is larger than a preset third data threshold value, the vehicle is determined as an illegal vehicle, after the illegal person and the illegal vehicle are finally identified, resources such as videos and pictures of the illegal person and/or the illegal vehicle appear in a correlated mode, the resources are stored and subjected to law enforcement and evidence obtaining, a back-end processing server can control an alarm device to give an alarm after the illegal person and the illegal person appear, and field operation and maintenance personnel go to the alarm and are separated, so that the personal safety of passengers is maintained.

Example 6;

fig. 8 is a schematic structural diagram of an image recognition apparatus according to some embodiments of the present invention, where the apparatus includes:

a determining module 801, configured to identify, for any image acquired within a preset time length, a pedestrian appearing in the image, and if the number of times of any pedestrian appearing within the preset time length is greater than a set number threshold, determine that the pedestrian is a suspicious pedestrian;

the determining module 801 is further configured to determine a motion trajectory of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length;

the processing module 802 is configured to, if the similarity between the motion trajectory of the suspicious pedestrian and any stored standard trajectory meets a preset condition, use the suspicious pedestrian as the identified target pedestrian.

In a possible implementation manner, the determining module 801 is further configured to determine whether an image matching with the face image of the pedestrian exists in a face database according to a face database of an existing suspicious pedestrian and a face database of a working person, which is stored in advance; if not, then the subsequent step of taking the pedestrian as the suspicious pedestrian is carried out.

In a possible implementation manner, the determining module 801 is further configured to determine that any pedestrian is a non-suspicious pedestrian if the number of occurrences of the pedestrian within the preset time length is not greater than a set number threshold, or an image matching the face image of the pedestrian exists in the face database.

In a possible implementation manner, the determining module 801 is further configured to determine that a target image of the suspicious pedestrian is included in the images acquired within the preset time period; determining other pedestrians in the same row with the suspicious pedestrian according to the distance between the suspicious pedestrian and other pedestrians in the target image; and if the number of other pedestrians on the same row with the suspicious pedestrian within the preset time length is larger than a preset first number threshold, performing a subsequent step of determining the motion trail of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length.

In a possible implementation manner, the determining module 801 is further configured to determine that the pedestrian is a non-suspicious pedestrian if the similarity between the motion trajectory of the suspicious pedestrian and any one of the stored standard trajectories does not meet a preset condition.

In a possible embodiment, the apparatus further comprises:

an identifying module 803, configured to identify a vehicle appearing in any one of the images acquired within a preset time period;

the processing module 802 is further configured to, if the number of times that any vehicle appears within the preset time length is greater than a set second number threshold, take the vehicle as a suspicious vehicle;

a determining module 804, configured to determine that a target image including the target pedestrian exists in the image acquired within the preset time period, and determine whether the target pedestrian and the suspicious vehicle in the target image exist at the same time;

the processing module 802 is further configured to, if the number of times that the target pedestrian and the suspicious vehicle coexist within the preset time length is greater than a preset third number threshold, take the suspicious vehicle as the target vehicle.

In a possible implementation manner, the identification module 803 is specifically configured to determine whether a vehicle matching the vehicle exists in the vehicle database according to a vehicle database of pre-stored existing suspicious vehicles and legally operated vehicles; if not, the subsequent step of taking the vehicle as a suspicious vehicle is carried out.

Example 7:

on the basis of the foregoing embodiments, some embodiments of the present invention further provide an electronic device, as shown in fig. 9, including: a processor 901, a communication interface 902, a memory 903 and a communication bus 904, wherein the processor 901, the communication interface 902 and the memory 903 are communicated with each other through the communication bus 904.

The memory 903 has stored therein a computer program which, when executed by the processor 801, causes the processor 901 to perform the steps of:

Further, the processor 901 is further configured to determine whether an image matching the face image of the pedestrian exists in the face database according to a pre-stored face database of the existing suspicious pedestrian and the worker; if not, then the subsequent step of taking the pedestrian as the suspicious pedestrian is carried out.

Further, the processor 901 is further configured to determine that any pedestrian is a non-suspicious pedestrian if the number of occurrences of the pedestrian in the preset time length is not greater than a set number threshold, or an image matching the face image of the pedestrian exists in the face database.

Further, the processor 901 is further configured to determine that the target image of the suspicious pedestrian is included in the images acquired within the preset time period; determining other pedestrians in the same row with the suspicious pedestrian according to the distance between the suspicious pedestrian and other pedestrians in the target image; and if the number of other pedestrians on the same row with the suspicious pedestrian within the preset time length is larger than a preset first number threshold, performing a subsequent step of determining the motion trail of the suspicious pedestrian within the preset time length according to the image acquired within the preset time length.

Further, the processor 901 is further configured to determine that the pedestrian is a non-suspicious pedestrian if the number of other pedestrians that are in the same row with the suspicious pedestrian within the preset time length is not greater than a preset first number threshold.

Further, the processor 901 is further configured to determine the pedestrian as a non-suspicious pedestrian if the similarity between the motion trajectory of the suspicious pedestrian and any stored standard trajectory does not meet a preset condition.

Further, the processor 901 is further configured to identify, for any image acquired within a preset time period, a vehicle appearing in the image; if the number of times of any vehicle appearing in the preset time length is larger than a set second number threshold, the vehicle is taken as a suspicious vehicle; determining that the images acquired within the preset time span contain a target image of the target pedestrian, and determining whether the target pedestrian and the suspicious vehicle in the target image exist at the same time; and if the number of times that the target pedestrian and the suspicious vehicle coexist within the preset time length is greater than a preset third quantity threshold value, taking the suspicious vehicle as the target vehicle.

Further, the processor 901 determines whether a vehicle matching the vehicle exists in the vehicle database according to a vehicle database of pre-stored existing suspicious vehicles and legal operation vehicles; if not, the subsequent step of taking the vehicle as a suspicious vehicle is carried out.

The communication bus mentioned in the above server may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.

The communication interface 902 is used for communication between the electronic apparatus and other apparatuses.

The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Alternatively, the memory may be at least one memory device located remotely from the processor.

The Processor may be a general-purpose Processor, including a central processing unit, a Network Processor (NP), and the like; but may also be a Digital instruction processor (DSP), an application specific integrated circuit, a field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like.

Example 8:

on the basis of the foregoing embodiments, an embodiment of the present invention further provides a computer-readable storage medium, in which a computer program executable by an electronic device is stored, and when the program is run on the electronic device, the electronic device is caused to execute the following steps:

the memory having stored therein a computer program that, when executed by the processor, causes the processor to perform the steps of:

Further, the method further comprises:

and if the similarity between the motion trail of the suspicious pedestrian and any one of the stored standard trails does not meet the preset condition, determining the pedestrian as a non-suspicious pedestrian.

Further, the method further comprises:

In the invention, the pedestrian appearing in the image is identified aiming at any image acquired within the preset time length, if the appearing times of any pedestrian within the preset time length are more than the set time threshold, the pedestrian is determined as the suspicious pedestrian, and the motion trail of the suspicious pedestrian is determined according to the image acquired within the preset time length; if the similarity between the motion track of the suspicious pedestrian and any one of the stored standard tracks meets a preset condition, the suspicious pedestrian is used as the identified target pedestrian, so that the automatic identification of the suspicious pedestrian can be realized, and the target pedestrian is finally determined.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims

1. An image recognition method, comprising:

2. The method according to claim 1, wherein if the number of occurrences of any pedestrian within the preset time length is greater than a set threshold number of occurrences, the pedestrian is determined to be a suspicious pedestrian, and the method further comprises:

3. The method of claim 2, further comprising:

4. The method according to claim 1, wherein before determining the motion trajectory of the suspicious pedestrian within the preset time period according to the image acquired within the preset time period, the method further comprises:

5. The method of claim 4, further comprising:

6. The method of claim 1, further comprising:

7. The method according to claim 6, wherein if the number of occurrences of any vehicle within the preset time period is greater than a set second number threshold, then the vehicle is considered as a suspicious vehicle, and the method comprises:

8. An image recognition apparatus, characterized in that the apparatus comprises:

9. An electronic device, characterized in that the electronic device comprises a processor for implementing the steps of the method according to any of claims 1-7 when executing a computer program stored in a memory.

10. A computer-readable storage medium, characterized in that it stores a computer program which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.