WO2021022698A1

WO2021022698A1 - Following detection method and apparatus, and electronic device and storage medium

Info

Publication number: WO2021022698A1
Application number: PCT/CN2019/116353
Authority: WO
Inventors: 徐杰
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-08-08
Filing date: 2019-11-07
Publication date: 2021-02-11
Also published as: CN110633636A; CN110633636B

Abstract

A following detection method and apparatus, and an electronic device and a storage medium. The method comprises: when a following detection instruction is received, acquiring a video to be detected (S10); detecting pedestrians in said video, and obtaining a first pedestrian detection box (S11); performing scale transformation on the first pedestrian detection box, and obtaining at least one group of second pedestrian detection boxes (S12); identifying a motion target in said video, and obtaining a first region (S13); comparing each group of the at least one group of second pedestrian detection boxes with the first region, and obtaining a third pedestrian detection box in the first region (S14); calculating the number of the third pedestrian detection boxes (S15); determining the group number of the third pedestrian detection boxes (S16); calculating a ratio of the number of the third pedestrian detection boxes to the group number of the third pedestrian detection boxes (S17); and when the ratio is greater than or equal to a configuration value, determining that a following event occurs in said video (S18). On the basis of image detection, whether the following event occurs in said video is accurately detected under the condition of eliminating interference of other moving objects.

Description

Trailing detection method, device, electronic equipment and storage medium

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on August 8, 2019. The application number is 201910731178.6. The invention title is "Trailer detection method, device, electronic equipment and storage medium". The entire content is incorporated by reference. In this application.

Technical field

This application relates to the field of image processing technology, and in particular to a trailing detection method, device, electronic equipment and storage medium.

Background technique

With the increasing application of turnstiles, the installation of infrared sensors on the turnstile channels to detect whether there is a trailing has become the mainstream of contemporary trailing detection technology. However, the method of detecting whether there is a trailing through infrared sensors is not accurate. high.

Therefore, in order to improve the accuracy of trailing detection, video anti-tailing technology has also emerged. However, in traditional video anti-tailing technology, due to the different angles of the recorded video, the accuracy of trailing detection will be affected. At the same time, moving objects (such as luggage , Pets, etc.) will also cause some interference to the trailing detection of pedestrians.

Summary of the invention

In view of the above, it is necessary to provide a trailing detection method, device, electronic equipment and storage medium, which can accurately detect whether trailing occurs in the video to be detected under the interference of other moving objects, and can also solve the problem of recording video from different angles. Impact on detecting the video to be detected.

A trailing detection method, the method includes:

When receiving the trailing detection instruction, obtain the video to be detected;

Detecting pedestrians in the video to be detected to obtain a first pedestrian detection frame;

Performing scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes: scaling the length and/or width of the first pedestrian detection frame to obtain multiple detection frames, and Acquiring the first pedestrian detection frame that has not undergone scale transformation, and grouping the multiple detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames;

Identifying a moving target in the video to be detected to obtain a first area;

Comparing each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

Calculating the number of the third pedestrian detection frame;

Determining the number of groups of the third pedestrian detection frame;

Calculating the ratio of the number of the third pedestrian detection frame to the group number of the third pedestrian detection frame;

When the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.

According to a preferred embodiment of the present application, the detecting pedestrians in the video to be detected to obtain the first pedestrian detection frame includes:

Use the Faster-RCNN algorithm to extract the feature map of the pedestrian in the video to be detected;

The area candidate network is used to process the feature map to obtain the first pedestrian detection frame.

According to a preferred embodiment of the present application, the scale conversion of the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes one or a combination of the following methods:

Scale the length of the first pedestrian detection frame by using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

Use a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames; and/or

Scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

The first pedestrian detection frame is determined as a group of second pedestrian detection frames.

According to a preferred embodiment of the present application, the identifying the moving target in the to-be-detected video to obtain the first area includes:

Extracting at least one continuous frame of image in the video to be detected by using an OpenCV image processing algorithm;

Performing a difference operation on pixels corresponding to different frames in the at least one frame of image to obtain a grayscale difference;

When the absolute value of the grayscale difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.

According to a preferred embodiment of the present application, each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames is compared with the first area to obtain a third pedestrian detection in the first area The box includes:

Comparing each second pedestrian detection frame in each group of second pedestrian detection frames with the first area;

When a second pedestrian detection frame overlaps the first area, the overlapping second pedestrian detection frame is confirmed as the third pedestrian detection frame.

According to a preferred embodiment of the present application, the method further includes:

When it is determined that a trailing event occurs in the video to be detected, an alarm is issued;

Wherein, the alarm includes an indicator light alarm and/or a speaker alarm.

According to a preferred embodiment of the present application, after the alarm is issued, the method further includes:

When the trailing event has been processed, terminate the alarm; or

When the trailing event is not processed in a preset time interval, a prompt message is sent to the terminal device of the designated person.

A trailing detection device, the device comprising:

The obtaining unit is used to obtain the video to be detected when the trailing detection instruction is received;

The detection unit is configured to detect pedestrians in the video to be detected to obtain the first pedestrian detection frame;

The transformation unit is used to perform scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames, including: scaling the length and/or width of the first pedestrian detection frame to obtain more A detection frame, and obtain the first pedestrian detection frame that has not been scale-transformed, and group the plurality of detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames ；

An identification unit, configured to identify a moving target in the video to be detected to obtain a first area;

A comparing unit, configured to compare each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

A calculation unit for calculating the number of the third pedestrian detection frame;

A determining unit, configured to determine the number of groups of the third pedestrian detection frame;

The calculation unit is further configured to calculate the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame;

The determining unit is further configured to determine that a trailing event occurs in the video to be detected when the ratio is greater than or equal to a configured value.

According to a preferred embodiment of the present application, the detection unit is specifically configured to:

According to a preferred embodiment of the present application, the transformation unit is specifically configured to:

According to a preferred embodiment of the present application, the identification unit is specifically configured to:

According to a preferred embodiment of the present application, the comparison unit is specifically configured to:

According to a preferred embodiment of the present application, the device further includes:

A sending unit, configured to issue an alarm when it is determined that a trailing event occurs in the video to be detected;

Wherein, the alarm includes an indicator light alarm and/or a speaker alarm.

The termination unit is used to terminate the alarm when the trailing event has been processed after the alarm is issued; or

The sending unit is further configured to send prompt information to the terminal device of the designated person when the trailing event has not been processed in a preset time interval.

An electronic device, which includes:

Memory, storing at least one instruction; and

The processor executes the instructions stored in the memory to implement the trailing detection method.

A non-volatile readable storage medium stores at least one instruction, and the at least one instruction is executed by a processor in an electronic device to implement the trailing detection method.

It can be seen from the above technical solutions that this application can obtain the video to be detected when receiving the trailing detection instruction, detect pedestrians in the video to be detected, obtain the first pedestrian detection frame, and detect the first pedestrian The frame is scaled to obtain at least one group of second pedestrian detection frames, the moving target in the video to be detected is identified, and the first area is obtained, and each group of second pedestrians in the at least one group of second pedestrian detection frames is detected Compare the frame with the first area to obtain the third pedestrian detection frame in the first area, calculate the number of the third pedestrian detection frame, and determine the group number of the third pedestrian detection frame, and further Calculate the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame, and when the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected, not only in Under the interference of other moving objects, it can accurately detect whether there is a trailing in the video to be detected, and it can also solve the impact of recording videos from different angles on the detection of the video to be detected.

Description of the drawings

Fig. 1 is a flowchart of a preferred embodiment of the trailing detection method of the present application.

Fig. 2 is a functional module diagram of a preferred embodiment of the trailing detection device of the present application.

Fig. 3 is a schematic structural diagram of an electronic device implementing a preferred embodiment of a trailing detection method according to the present application.

detailed description

In order to make the objectives, technical solutions, and advantages of the present application clearer, the present application will be described in detail below with reference to the drawings and specific embodiments.

As shown in FIG. 1, it is a flowchart of a preferred embodiment of the trailing detection method of the present application. According to different needs, the order of the steps in the flowchart can be changed, and some steps can be omitted.

The trailing detection method is applied to one or more electronic devices.

S10: When a trailing detection instruction is received, a video to be detected is acquired.

In at least one embodiment of the present application, the trailing detection instruction can be triggered by the user, or can be triggered automatically when a certain condition is met, and this application is not limited.

Wherein, the meeting certain conditions includes, but is not limited to: meeting a preset time, the electronic device detects that someone passes by, etc.

The preset time may include a determined time point, or include a time period, etc., for example: the preset time may be 7 o'clock in the morning every day.

In at least one embodiment of the present application, the video to be detected includes, but is not limited to, one or a combination of the following:

(1) Video recorded by the security inspection device.

Specifically, the security check device can be installed in places such as entrances, company gates, etc. When the security check device detects that a person passes by, the camera device connected to the security check device is controlled to record the entire process of the person passing by.

In this way, the entire process of people passing through the security inspection device can be recorded, which is conducive to follow-up detection of the recorded video, and further prevents people without permission from entering the station, the company and other places.

(2) Video recorded by the camera at the exit.

Specifically, the exits include, but are not limited to: exits of railway stations, exits of subway stations, exits of airports, etc.

In this way, the entire process of people leaving the station can be recorded, which is the basis for subsequent detection of the video recorded by the camera device, so as to detect whether there is a trailing in the video, and further determine whether there is fare evasion, and thus prevent fare evasion. .

S11: Detect pedestrians in the to-be-detected video to obtain a first pedestrian detection frame.

In at least one embodiment of the present application, detecting the pedestrian in the video to be detected by the electronic device to obtain the first pedestrian detection frame includes:

The electronic device uses the Faster-RCNN (Faster-Region-based Convolutional Neural Networks) algorithm to extract the feature map of the pedestrian in the video to be detected. Further, the electronic device uses the regional candidate network to process the feature map to obtain the The first pedestrian detection frame.

Wherein, when the Faster-RCNN algorithm is used to detect pedestrians in the video to be detected, the missed detection rate is low.

Specifically, the electronic device using the Faster-RCNN algorithm to detect pedestrians in the video to be detected includes the following four steps: feature extraction, candidate region generation, candidate region classification, and location refinement.

In at least one embodiment of the present application, detecting the pedestrian in the video to be detected by the electronic device to obtain the first pedestrian detection frame specifically includes:

First, the electronic device uses a series of basic convolution (Convolution, conv), linear rectification function (Rectified Linear Unit, ReLU), and pooling to extract the features of pedestrians in the video to be detected to obtain a feature map, Further, the electronic device uses a regional candidate network (Region Proposal Networks, RPN) to segment the feature map into multiple candidate regions, and identify the foreground in the candidate region and the background in the candidate region, and further Preferably, the electronic device extracts the rough coordinates of the foreground, and accurately returns the rough coordinates of the foreground to obtain accurate foreground coordinates, and cuts the pedestrian from the feature map according to the accurate foreground coordinates Then, the pedestrian detection frame is obtained. Finally, the electronic device performs a pooling operation on the pedestrian detection frame to obtain the first pedestrian detection frame of a fixed size.

Through the foregoing implementation manners, pedestrians in the video to be detected can be accurately detected, and then the first pedestrian detection frame is obtained, which lays a foundation for detecting whether a trailing occurs.

S12: Perform scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames.

In at least one embodiment of the present application, the electronic device performing scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes:

The electronic device scales the length and/or width of the first pedestrian detection frame to obtain a plurality of detection frames, and further acquires the first pedestrian detection frame that has not been scaled, and the electronic device The plurality of detection frames and the first pedestrian detection frame are grouped to obtain the at least one group of second pedestrian detection frames.

Specifically, the electronic device performs scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames including, but not limited to, one or a combination of the following methods:

(1) The electronic device uses a preset ratio to scale the length of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

For example: the length of the first pedestrian detection frame is 3, the width is 5, and the preset ratios are 1.3 and 1.5, then two sets of second pedestrian detection frames are obtained. The aspect ratios of the two second pedestrian detection frames are: 3.9 :5 and 4.5:5.

(2) The electronic device uses a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

For example: the length of the first pedestrian detection frame is 3, the width is 5, and the preset ratios are 1.3 and 1.5, then two sets of second pedestrian detection frames are obtained. The aspect ratios of the two sets of second pedestrian detection frames are: 3 :6.5 and 3:7.5.

(3) The electronic device uses a preset ratio to scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

For example: the length of the first pedestrian detection frame is 3, the width is 5, and the preset ratios are 1.3 and 1.5, then two sets of second pedestrian detection frames are obtained. The aspect ratios of the two second pedestrian detection frames are: 3.9 :6.5 and 4.5:7.5.

(4) The electronic device determines the first pedestrian detection frame as a group of second pedestrian detection frames.

For example, if the length of the first pedestrian detection frame is 3 and the width is 5, no matter what the preset ratio is, the electronic device can obtain a set of second pedestrian detection frames, and the second pedestrian detection frame The aspect ratio is 3:5.

Wherein, the preset ratio may be determined according to the angle at which the video is recorded by the camera device, and according to a large number of experiments, the preset ratio may generally include 1.3 or 0.9, which is not limited in this application.

Further, the present application does not limit the number of configurations of the preset ratio, for example: the number of configurations of the preset ratio may be two.

Through the foregoing implementation manners, by detecting multiple sets of second pedestrian detection frames, the accuracy of detection can also be improved.

In at least one embodiment of the present application, the electronic device determines the number of groups of the second pedestrian detection frame according to the number of the preset ratio.

Specifically, the electronic device multiplies the number of the preset ratio by 3 to obtain a first value, and further, adds one to the first value to obtain the group number of the second pedestrian detection frame.

In general, since the electronic device uses the preset ratio to scale the length and width of the first pedestrian detection frame, the number of groups of the second pedestrian detection frame will be due to the preset ratio. The number of ratios varies.

Through the foregoing implementation manners, the group number of the second pedestrian detection frame can be quickly and accurately obtained, which facilitates the subsequent determination of the group number of the third pedestrian detection frame.

Of course, in other embodiments, the electronic device may also use other methods to group the multiple detection frames and the first pedestrian detection frame, for example: grouping according to the number of pedestrians, this application will not do this here. limit.

S13: Identify the moving target in the to-be-detected video to obtain the first area.

In at least one embodiment of the present application, the electronic device recognizing the moving target in the video to be detected and obtaining the first area includes:

The electronic device uses the OpenCV image processing algorithm to extract at least one continuous frame of image in the to-be-detected video, and performs a differential operation on the pixels corresponding to different frames of the at least one frame of image to obtain the grayscale difference. When the absolute value of the grayscale difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.

Among them, this application does not limit the selection of the threshold.

Of course, in other embodiments, the electronic device recognizes the moving target in the video to be detected, and obtains the first area may also adopt the following methods, which specifically include:

The electronic device selects any frame of image in the video to be detected as a background frame, and further, the electronic device uses an inter-frame difference method to perform differential processing on two adjacent frames of images in the video to be detected to obtain The inter-frame difference value updates the background frame according to the inter-frame difference value, and further uses the difference between the current image and the background frame to obtain the foreground of the moving target, that is, the first region.

Through the foregoing implementation manners, the moving target in the video to be detected can be quickly and accurately identified, and the occurrence of missed detection can be effectively avoided.

Of course, the method of identifying the moving target in the video to be detected is not limited by this application as long as it is legal and reasonable.

S14. Compare each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area.

In at least one embodiment of the present application, each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames is compared with the first area to obtain the first area in the first area. The three-passenger detection frame includes:

The electronic device compares each second pedestrian detection frame in each group of second pedestrian detection frames with the first area, and when a second pedestrian detection frame overlaps the first area, the overlap The second pedestrian detection frame is confirmed as the third pedestrian detection frame.

For example: there are three second pedestrian detection frames in a certain group of second pedestrian detection frames, namely the second pedestrian detection frame A, the second pedestrian detection frame B, and the second pedestrian detection frame C. There are 2 moving targets in the first area , Respectively, the moving target X and the moving target Y, the second pedestrian detection frame A, the second pedestrian detection frame B, the second pedestrian detection frame C and the moving target in the first area X and the moving target Y are compared, where only the second pedestrian detection frame A overlaps the moving target X, and the others do not overlap each other, the second pedestrian detection frame A is confirmed as the third pedestrian detection frame.

Through the above implementation, the second pedestrian detection frame is compared with the first area, which actually compares the pedestrian with the moving target to determine the moving pedestrian in the video to be detected, that is, the third pedestrian detection frame. , Other moving objects are effectively eliminated.

In addition, since the video recorded at different angles will affect the shape of the moving object, comparing each group of different ratios of the second pedestrian detection frame with the first area can solve the problem of detecting the pending video from different angles. Detect the impact of video.

S15: Calculate the number of the third pedestrian detection frame.

In at least one embodiment of the present application, the number of the third pedestrian detection frames is calculated by accumulating the obtained third pedestrian detection frames one by one.

Through the foregoing implementation manner, the number of the third pedestrian detection frame can be quickly and accurately calculated without human participation.

S16: Determine the number of groups of the third pedestrian detection frame.

In at least one embodiment of the present application, the electronic device determines the number of groups of the third pedestrian detection frame according to the number of groups of the second pedestrian detection frame.

Specifically, since the third pedestrian detection frame is obtained by comparing each group of second pedestrian detection frames with the first area, the number of groups of the third pedestrian detection frame is equal to that of the second pedestrian detection frame. The number of groups of detection frames is equal.

Through the foregoing implementation manner, the number of groups of the third pedestrian detection frame can be quickly determined.

S17: Calculate the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame.

For example, if the number of third pedestrian detection frames is 14, and the number of groups of third pedestrian detection frames is 7, then the ratio of the number of third pedestrian detection frames to the number of groups of third pedestrian detection frames Is 2.

Through the above-mentioned implementation manners, the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame can be quickly and accurately calculated without human operation, which improves efficiency.

S18: When the ratio is greater than or equal to a configured value, it is determined that a trailing event occurs in the video to be detected.

Wherein, the configuration value is configured according to empirical values, such as 1.5.

For example, if the ratio of the number of third pedestrian detection frames to the group number of third pedestrian detection frames is 2, which is greater than the configured value of 1.5, it is determined that there is a trailing in the video to be detected.

Through the foregoing implementation manners, it can be quickly determined that a trailing occurs in the to-be-detected video, so that relevant personnel can handle the trailing phenomenon in a timely manner.

In at least one embodiment of the present application, the method further includes:

When it is determined that a trailing event occurs in the video to be detected, the electronic device issues an alarm, where the alarm includes an indicator light alarm and/or a speaker alarm.

Through the foregoing implementation manners, after detecting a trailing event, an alarm can be issued in a timely manner to remind relevant personnel to ensure that relevant personnel can quickly understand the trailing situation without requiring relevant personnel to stare at the electronic device at all times, which further improves user experience.

In at least one embodiment of the present application, after the alarm is issued, the method further includes:

When the trailing event has been processed, the electronic device terminates the alarm, and when the trailing event has not been processed within a preset time interval, a prompt message is sent to a terminal device of a designated person.

Among them, the designated person includes, but is not limited to: the relevant person in charge of the trailing process, for example: the company's personnel director, the manager of the subway station.

The preset time interval may include 12 hours or 24 hours.

The prompt information may include, but is not limited to: trailing data and so on. Wherein, the trailing data includes, but is not limited to: the time when the trailing event occurred, evidence of the trailing event, photos of the trailing person, and the like.

Through the foregoing implementation manners, it is possible to supervise the relevant personnel of the trailing process, avoid the trailing phenomenon from not being processed in time, and ensure the timeliness of the trailing process.

As shown in FIG. 2, it is a functional module diagram of a preferred embodiment of the trailing detection device of the present application. The trailing detection device 11 includes an acquisition unit 110, a detection unit 111, a transformation unit 112, an identification unit 113, a comparison unit 114, a calculation unit 115, a determination unit 116, a transmission unit 117, and a termination unit 118. The module/unit referred to in this application refers to a series of computer-readable instruction segments that can be executed by the processor 13 and can complete fixed functions, and are stored in the memory 12. In this embodiment, the functions of each module/unit will be described in detail in subsequent embodiments.

When the trailing detection instruction is received, the acquiring unit 110 acquires the video to be detected.

In at least one embodiment of the present application, the trailing detection instruction can be triggered by the user, or it can be triggered automatically when certain conditions are met, which is not limited by this application.

(1) Video recorded by the security inspection device.

(2) Video recorded by the camera at the exit.

The detection unit 111 detects pedestrians in the video to be detected, and obtains a first pedestrian detection frame.

In at least one embodiment of the present application, the detection unit 111 detecting the pedestrian in the video to be detected, and obtaining the first pedestrian detection frame includes:

The detection unit 111 uses the Faster-RCNN (Faster-Region-based Convolutional Neural Networks) algorithm to extract the feature map of the pedestrian in the video to be detected. Further, the detection unit 111 uses a regional candidate network to process the feature map, Obtain the first pedestrian detection frame.

Specifically, the detection unit 111 using the Faster-RCNN algorithm to detect pedestrians in the video to be detected includes the following four steps: feature extraction, candidate region generation, candidate region classification, and location refinement.

In at least one embodiment of the present application, the detection unit 111 detects pedestrians in the video to be detected to obtain the first pedestrian detection frame, which specifically includes:

First, the detection unit 111 uses a series of basic convolution (Convolution, conv), linear rectification function (Rectified Linear Unit, ReLU), pooling (pooling) to extract the features of pedestrians in the video to be detected to obtain a feature map Further, the detection unit 111 uses a region candidate network (Region Proposal Networks, RPN) to segment the feature map into a plurality of candidate regions, and identify the foreground in the candidate region and the background in the candidate region, Further, the detection unit 111 extracts the rough coordinates of the foreground, and accurately returns the rough coordinates of the foreground to obtain accurate foreground coordinates, and according to the accurate foreground coordinates, the pedestrian is removed from the feature Cut out in the figure to obtain a pedestrian detection frame. Finally, the detection unit 111 performs a pooling operation on the pedestrian detection frame to obtain the first pedestrian detection frame of a fixed size.

The transformation unit 112 performs scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames.

In at least one embodiment of the present application, the transformation unit 112 performs scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames including:

The transformation unit 112 scales the length and/or width of the first pedestrian detection frame to obtain multiple detection frames, and further acquires the first pedestrian detection frame that has not been scaled, the transformation unit 112 group the plurality of detection frames and the first pedestrian detection frame to obtain the at least one group of second pedestrian detection frames.

Specifically, the transformation unit 112 performs scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames including, but not limited to, one or a combination of the following methods:

(1) The transformation unit 112 uses a preset ratio to scale the length of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

(2) The transformation unit 112 uses a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

(3) The transformation unit 112 uses a preset ratio to scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames.

(4) The transformation unit 112 determines the first pedestrian detection frame as a group of second pedestrian detection frames.

For example, if the length of the first pedestrian detection frame is 3 and the width is 5, no matter what the preset ratio is, the transformation unit 112 can obtain a set of second pedestrian detection frames, and the second pedestrian detection frame The aspect ratio of the frame is 3:5.

Further, this application does not limit the number of configurations of the preset ratio, for example: the number of configurations of the preset ratio may be two.

In at least one embodiment of the present application, the determining unit 116 determines the number of groups of the second pedestrian detection frame according to the number of the preset ratio.

Specifically, the determining unit 116 multiplies the number of the preset ratios by 3 to obtain a first value, and further, adds one to the first value to obtain the group number of the second pedestrian detection frame.

In general, since the transformation unit 112 uses the preset ratio to scale the length and width of the first pedestrian detection frame, the number of groups of the second pedestrian detection frame may be due to the preset ratio. Let the number of proportions differ.

Of course, in other embodiments, the multiple detection frames and the first pedestrian detection frame may also be grouped in other ways, for example: grouping according to the number of pedestrians, which is not limited in this application.

The recognition unit 113 recognizes the moving target in the video to be detected, and obtains the first area.

In at least one embodiment of the present application, the recognition unit 113 recognizes the moving target in the to-be-detected video, and obtains the first region includes:

The recognition unit 113 uses the OpenCV image processing algorithm to extract at least one continuous frame of the image to be detected, and performs a differential operation on the pixels corresponding to different frames of the at least one frame of image to obtain the grayscale difference. When the absolute value of the gray level difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.

Among them, this application does not limit the selection of the threshold.

Of course, in other embodiments, the recognition unit 113 recognizes the moving target in the video to be detected, and obtains the first area may also adopt the following methods, which specifically include:

The identification unit 113 selects any one frame of the image in the video to be detected as a background frame, and further, the identification unit 113 uses an inter-frame difference method to perform differential processing on two adjacent frames of the image in the video to be detected To obtain an inter-frame difference value, update the background frame according to the inter-frame difference value, and further use the current image and the background frame to perform a difference to obtain the foreground of the moving target, that is, the first region.

The comparing unit 114 compares each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area.

In at least one embodiment of the present application, the comparing unit 114 compares each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain the first area The third pedestrian detection box in includes:

The comparison unit 114 compares each second pedestrian detection frame in each group of second pedestrian detection frames with the first area, and when a second pedestrian detection frame overlaps the first area, it will overlap The second pedestrian detection frame of is confirmed as the third pedestrian detection frame.

The calculation unit 115 calculates the number of the third pedestrian detection frame.

The determining unit 116 determines the number of groups of the third pedestrian detection frame.

In at least one embodiment of the present application, the determining unit 116 determines the number of groups of the third pedestrian detection frame according to the number of groups of the second pedestrian detection frame.

The calculation unit 115 calculates the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame.

The determining unit 116 determines that a trailing event occurs in the video to be detected when the ratio is greater than or equal to the configured value.

For example, if the ratio of the number of third pedestrian detection frames to the group number of third pedestrian detection frames is 2, which is greater than the configured value 1.5, it is determined that a trailing occurs in the video to be detected.

In at least one embodiment of the present application, when it is determined that a trailing event occurs in the to-be-detected video, the sending unit 117 issues an alarm, where the alarm includes an indicator light alarm and/or a speaker alarm.

Through the foregoing implementation manners, after detecting a trailing event, an alarm can be issued in a timely manner to remind related personnel to ensure that the related personnel can quickly understand the trailing situation without requiring the related personnel to always stare at the electronic device, which further improves the user experience.

In at least one embodiment of the present application, after an alarm is issued, when the trailing event has been processed, the terminating unit 118 terminates the alarm, and when the trailing event has not been processed within a preset time interval, the The sending unit 117 sends prompt information to the terminal device of the designated person.

The preset time interval may include 12 hours or 24 hours.

As shown in FIG. 3, it is a schematic structural diagram of an electronic device implementing a preferred embodiment of the trailing detection method of the present application.

The electronic device 1 is a device that can automatically perform numerical calculation and/or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC) ), programmable gate array (Field-Programmable Gate Array, FPGA), digital processor (Digital Signal Processor, DSP), embedded equipment, etc.

The electronic device 1 can also be, but is not limited to, any electronic product that can interact with the user through a keyboard, a mouse, a remote control, a touch panel, or a voice control device, for example, a personal computer, a tablet computer, or a smart phone. , Personal Digital Assistant (PDA), game consoles, interactive network TV (Internet Protocol Television, IPTV), smart wearable devices, etc.

The electronic device 1 may also be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.

The network where the electronic device 1 is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (Virtual Private Network, VPN), etc.

In an embodiment of the present application, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer readable instructions stored in the memory 12 and running on the processor 13 , Such as the trailing detection program.

Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 1 and does not constitute a limitation on the electronic device 1. It may include more or less components than those shown in the figure, or a combination of certain components, or different components. Components, for example, the electronic device 1 may also include input and output devices, network access devices, buses, and the like.

The processor 13 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (ASICs), Ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor 13 is the computing core and control center of the electronic device 1 and connects the entire electronic device with various interfaces and lines. Each part of 1, and executes the operating system of the electronic device 1, and various installed applications, program codes, etc.

The processor 13 executes the operating system of the electronic device 1 and various installed applications. The processor 13 executes the application program to implement the steps in the foregoing embodiments of the trailing detection method, such as steps S10, S11, S12, S13, S14, S15, S16, S17, and S18 shown in FIG. 1.

Alternatively, the processor 13 implements the functions of the modules/units in the foregoing device embodiments when executing the computer-readable instructions, for example: when a trailing detection instruction is received, obtain the video to be detected; detect the video to be detected Pedestrians in, obtain a first pedestrian detection frame; perform scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames, including: lengthening the first pedestrian detection frame and/ Or wide zoom to obtain multiple detection frames, and obtain the first pedestrian detection frame that has not been scale-transformed, and group the multiple detection frames and the first pedestrian detection frame to obtain the at least A group of second pedestrian detection frames; identify the moving target in the video to be detected to obtain the first area; combine each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area Compare to obtain the third pedestrian detection frame in the first area; calculate the number of the third pedestrian detection frame; determine the group number of the third pedestrian detection frame; calculate the third pedestrian detection frame The ratio of the number to the group number of the third pedestrian detection frame; when the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.

Exemplarily, the computer-readable instructions may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 12 and executed by the processor 13 to Complete this application. The one or more modules/units may be a series of computer-readable instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer-readable instructions in the electronic device 1. For example, the computer-readable instructions may be divided into an acquisition unit 110, a detection unit 111, a transformation unit 112, an identification unit 113, a comparison unit 114, a calculation unit 115, a determination unit 116, a transmission unit 117, and a termination unit 118.

The memory 12 may be used to store the computer-readable instructions and/or modules. The processor 13 executes or executes the computer-readable instructions and/or modules stored in the memory 12 and calls the computer-readable instructions and/or modules stored in the memory 12 The data inside realizes various functions of the electronic device 1. The memory 12 may mainly include a storage program area and a storage data area, where the storage program area may store an operating system, an application program required by at least one function (such as a trailing detection function, a sound playback function, an image playback function, etc.), etc.; The data storage area can store data created according to the use of the electronic device. In addition, the memory 12 may include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a Secure Digital (SD) card, a flash memory card (Flash Card), At least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device.

The memory 12 may be an external memory and/or an internal memory of the electronic device 1. Further, the memory 12 may be a circuit with a storage function without a physical form in an integrated circuit. Alternatively, the memory 12 may also be a memory in physical form, such as a memory stick, a TF card (Trans-flash Card), and so on.

If the integrated module/unit of the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer readable storage medium. Based on this understanding, this application implements all or part of the processes in the above-mentioned embodiments and methods, and can also be completed by instructing relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile memory. In the read storage medium, when the computer-readable instructions are executed by the processor, the steps of the foregoing method embodiments can be implemented.

Wherein, the computer-readable instruction includes computer-readable instruction code, and the computer-readable instruction code may be in the form of source code, object code, executable file, or some intermediate form. The non-volatile computer-readable medium may include: any entity or device capable of carrying the computer-readable instruction code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM , Read-Only Memory).

With reference to FIG. 1, the memory 12 in the electronic device 1 stores multiple instructions to implement a trailing detection method, and the processor 13 can execute the multiple instructions to realize: when a trailing detection instruction is received, Obtain the video to be detected; detect pedestrians in the video to be detected to obtain a first pedestrian detection frame; perform scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames, including: The first pedestrian detection frame is scaled in length and/or width to obtain multiple detection frames, and the first pedestrian detection frame that has not been scale-transformed is obtained, and the multiple detection frames and the first The pedestrian detection frames are grouped to obtain the at least one set of second pedestrian detection frames; the moving target in the video to be detected is identified to obtain the first area; each group of the at least one second pedestrian detection frame is The second pedestrian detection frame is compared with the first area to obtain the third pedestrian detection frame in the first area; the number of the third pedestrian detection frame is calculated; the group of the third pedestrian detection frame is determined Calculate the ratio of the number of the third pedestrian detection frame to the group number of the third pedestrian detection frame; when the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.

Specifically, for the specific implementation method of the processor 13 for the foregoing instructions, reference may be made to the description of the relevant steps in the embodiment corresponding to FIG. 1, which is not repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed system, device, and method may be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

In addition, it is obvious that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claims can also be implemented by one unit or device through software or hardware. The second class words are used to indicate names, and do not indicate any specific order.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the application and not to limit them. Although the application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the application can be Make modifications or equivalent replacements without departing from the spirit and scope of the technical solution of the present application.

Claims

A trailing detection method, characterized in that the method includes:

When receiving the trailing detection instruction, obtain the video to be detected;

Detecting pedestrians in the video to be detected to obtain a first pedestrian detection frame;

Performing scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes: scaling the length and/or width of the first pedestrian detection frame to obtain multiple detection frames, and Acquiring the first pedestrian detection frame that has not undergone scale transformation, and grouping the multiple detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames;

Identifying a moving target in the video to be detected to obtain a first area;

Comparing each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

Calculating the number of the third pedestrian detection frame;

Determining the number of groups of the third pedestrian detection frame;

Calculating the ratio of the number of the third pedestrian detection frame to the group number of the third pedestrian detection frame;

When the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.
The trailing detection method according to claim 1, wherein the detecting the pedestrian in the video to be detected to obtain the first pedestrian detection frame comprises:

Use the Faster-RCNN algorithm to extract the feature map of the pedestrian in the video to be detected;

The area candidate network is used to process the feature map to obtain the first pedestrian detection frame.
The trailing detection method according to claim 1, wherein the scaling of the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes one or a combination of the following methods:

Scale the length of the first pedestrian detection frame by using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

Use a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames; and/or

Scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

The first pedestrian detection frame is determined as a group of second pedestrian detection frames.
The trailing detection method according to claim 1, wherein the identifying the moving target in the to-be-detected video to obtain the first area comprises:

Extracting at least one continuous frame of image in the video to be detected by using an OpenCV image processing algorithm;

Performing a difference operation on pixels corresponding to different frames in the at least one frame of image to obtain a grayscale difference;

When the absolute value of the grayscale difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.
The trailing detection method according to claim 1, wherein each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames is compared with the first area to obtain the first area The third pedestrian detection frame in an area includes:

Comparing each second pedestrian detection frame in each group of second pedestrian detection frames with the first area;

When a second pedestrian detection frame overlaps the first area, the overlapping second pedestrian detection frame is confirmed as the third pedestrian detection frame.
The trailing detection method according to claim 1, wherein the method further comprises:

When it is determined that a trailing event occurs in the video to be detected, an alarm is issued;

Wherein, the alarm includes an indicator light alarm and/or a speaker alarm.
The trailing detection method according to claim 6, wherein after an alarm is issued, the method further comprises:

When the trailing event has been processed, terminate the alarm; or

When the trailing event is not processed in a preset time interval, a prompt message is sent to the terminal device of the designated person.
A trailing detection device, characterized in that the device comprises:

The obtaining unit is used to obtain the video to be detected when the trailing detection instruction is received;

The detection unit is configured to detect pedestrians in the video to be detected to obtain the first pedestrian detection frame;

The transformation unit is used to perform scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames, including: scaling the length and/or width of the first pedestrian detection frame to obtain more A detection frame, and obtain the first pedestrian detection frame that has not been scale-transformed, and group the plurality of detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames ；

An identification unit, configured to identify a moving target in the video to be detected to obtain a first area;

A comparing unit, configured to compare each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

A calculation unit for calculating the number of the third pedestrian detection frame;

A determining unit, configured to determine the number of groups of the third pedestrian detection frame;

The calculation unit is further configured to calculate the ratio of the number of the third pedestrian detection frame to the number of groups of the third pedestrian detection frame;

The determining unit is further configured to determine that a trailing event occurs in the video to be detected when the ratio is greater than or equal to a configured value.
An electronic device, characterized in that, the electronic device includes:

The memory stores at least one computer readable instruction; and

The processor executes the at least one computer-readable instruction to implement the following steps:

When receiving the trailing detection instruction, obtain the video to be detected;

Detecting pedestrians in the video to be detected to obtain a first pedestrian detection frame;

Performing scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes: scaling the length and/or width of the first pedestrian detection frame to obtain multiple detection frames, and Acquiring the first pedestrian detection frame that has not undergone scale transformation, and grouping the multiple detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames;

Identifying a moving target in the video to be detected to obtain a first area;

Comparing each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

Calculating the number of the third pedestrian detection frame;

Determining the number of groups of the third pedestrian detection frame;

Calculating the ratio of the number of the third pedestrian detection frame to the group number of the third pedestrian detection frame;

When the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.
The electronic device according to claim 9, wherein the processor executes at least one computer-readable instruction to realize the detection of the pedestrian in the video to be detected, and when the first pedestrian detection frame is obtained, the following steps are included: step:

Use the Faster-RCNN algorithm to extract the feature map of the pedestrian in the video to be detected;

The area candidate network is used to process the feature map to obtain the first pedestrian detection frame.
The electronic device of claim 9, wherein the processor executes at least one computer-readable instruction to implement the scale transformation of the first pedestrian detection frame to obtain at least one set of second pedestrian detection Frame, including one or a combination of the following methods:

Scale the length of the first pedestrian detection frame by using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

Use a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames; and/or

Scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

The first pedestrian detection frame is determined as a group of second pedestrian detection frames.
9. The electronic device of claim 9, wherein the processor executes at least one computer readable instruction to realize the recognition of the moving target in the video to be detected, and when the first region is obtained, the method comprises the following steps:

Extracting at least one continuous frame of image in the video to be detected by using an OpenCV image processing algorithm;

Performing a difference operation on pixels corresponding to different frames in the at least one frame of image to obtain a grayscale difference;

When the absolute value of the grayscale difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.
The electronic device according to claim 9, wherein the processor executes at least one computer-readable instruction to implement the combination of each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames and When the first area is compared to obtain the third pedestrian detection frame in the first area, the following steps are included:

Comparing each second pedestrian detection frame in each group of second pedestrian detection frames with the first area;

When a second pedestrian detection frame overlaps the first area, the overlapping second pedestrian detection frame is confirmed as the third pedestrian detection frame.
9. The electronic device of claim 9, wherein the processor executes at least one computer-readable instruction to further implement the following steps:

When it is determined that a trailing event occurs in the video to be detected, an alarm is issued;

Wherein, the alarm includes an indicator light alarm and/or a speaker alarm.
A non-volatile readable storage medium, characterized in that: the non-volatile readable storage medium stores at least one computer readable instruction, and the at least one computer readable instruction is used by a processor in an electronic device Perform the following steps:

When receiving the trailing detection instruction, obtain the video to be detected;

Detecting pedestrians in the video to be detected to obtain a first pedestrian detection frame;

Performing scale transformation on the first pedestrian detection frame to obtain at least one set of second pedestrian detection frames includes: scaling the length and/or width of the first pedestrian detection frame to obtain multiple detection frames, and Acquiring the first pedestrian detection frame that has not undergone scale transformation, and grouping the multiple detection frames and the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames;

Identifying a moving target in the video to be detected to obtain a first area;

Comparing each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames with the first area to obtain a third pedestrian detection frame in the first area;

Calculating the number of the third pedestrian detection frame;

Determining the number of groups of the third pedestrian detection frame;

Calculating the ratio of the number of the third pedestrian detection frame to the group number of the third pedestrian detection frame;

When the ratio is greater than or equal to the configured value, it is determined that a trailing event occurs in the video to be detected.
The storage medium according to claim 15, wherein the at least one computer-readable instruction is executed by the processor to realize the detection of the pedestrian in the video to be detected, and when the first pedestrian detection frame is obtained, it includes The following steps:

Use the Faster-RCNN algorithm to extract the feature map of the pedestrian in the video to be detected;

The area candidate network is used to process the feature map to obtain the first pedestrian detection frame.
The storage medium of claim 15, wherein the at least one computer-readable instruction is executed by a processor to implement the scale transformation of the first pedestrian detection frame to obtain at least one group of second pedestrians When detecting the frame, it includes one or a combination of the following methods:

Scale the length of the first pedestrian detection frame by using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

Use a preset ratio to scale the width of the first pedestrian detection frame to obtain the at least one set of second pedestrian detection frames; and/or

Scale the length of the first pedestrian detection frame and the width of the first pedestrian detection frame using a preset ratio to obtain the at least one set of second pedestrian detection frames; and/or

The first pedestrian detection frame is determined as a group of second pedestrian detection frames.
The storage medium according to claim 15, wherein the at least one computer-readable instruction is executed by the processor to realize the recognition of the moving target in the video to be detected, and when the first area is obtained, the following steps are included: :

Extracting at least one continuous frame of image in the video to be detected by using an OpenCV image processing algorithm;

Performing a difference operation on pixels corresponding to different frames in the at least one frame of image to obtain a grayscale difference;

When the absolute value of the grayscale difference is greater than or equal to the threshold, the pixel is confirmed as a moving target in the video to be detected, and the first area is obtained.
The storage medium according to claim 15, wherein the at least one computer-readable instruction is executed by a processor to implement the interaction between each group of second pedestrian detection frames in the at least one group of second pedestrian detection frames and the When the first area is compared to obtain the third pedestrian detection frame in the first area, the following steps are included:

Comparing each second pedestrian detection frame in each group of second pedestrian detection frames with the first area;

When a second pedestrian detection frame overlaps the first area, the overlapping second pedestrian detection frame is confirmed as the third pedestrian detection frame.
The storage medium of claim 15, wherein the at least one computer readable instruction is executed by the processor to further implement the following steps:

When it is determined that a trailing event occurs in the video to be detected, an alarm is issued;

Wherein, the alarm includes an indicator light alarm and/or a speaker alarm.