WO2021068553A1

WO2021068553A1 - Monitoring method, apparatus and device

Info

Publication number: WO2021068553A1
Application number: PCT/CN2020/097694
Authority: WO
Inventors: 陈庆; 张增东; 丁杉
Original assignee: 华为技术有限公司
Priority date: 2019-10-10
Filing date: 2020-06-23
Publication date: 2021-04-15
Also published as: CN112653832A

Abstract

A monitoring technique, comprising: acquiring, by means of image matching, the current camera that is monitoring a target object at the current moment; then predicting a next-hop camera corresponding to the next monitoring area that the target object is about to enter next; and performing operations, such as acquisition, caching and playing, on multimedia content of the next-hop camera in advance according to a prediction result.

Description

Monitoring method, device and equipment

Technical field

The invention relates to the field of security and protection, in particular to a monitoring method.

Background technique

In a modern society with increasingly developed information technology, smart security systems have penetrated into all aspects of life, and have played an irreplaceable role in the fields of home security, smart city construction, and safe cities. In banks, railways, stadiums, office buildings, markets, etc., there is an urgent need to rely on cameras for monitoring.

A basic function based on suspicious persons monitoring and alerting is: use cameras to photograph pedestrians, identify suspicious persons, and track the movement trajectory of suspicious persons.

However, because people are often on the move. Therefore, under normal circumstances, a suspicious person does not stay within the shooting range of a single camera, and its motion trajectory will switch between the monitoring ranges of multiple cameras. In this scenario, when a suspicious person enters the monitoring range of the next camera from the monitoring range of the current camera, the prior art needs to re-identify the available persons, and start tracking again after the recognition is successful. This is a slow response to the method and it is difficult to continuously monitor the personnel.

Summary of the invention

In the first aspect, the invention provides an embodiment of a monitoring method, including: acquiring a monitoring task, the monitoring task indicating a target object; determining a target camera based on the characteristic information of the target object, the target camera being used for monitoring In the first area, the target object is located in the first area at the current moment; predicting a next-hop camera, and the area monitored by the next-hop camera is the predicted surveillance area that the target object will enter at the next moment; sending The information of the target camera and the information of the next hop camera.

This solution can predict the monitoring range that the target object will enter at the next moment, so as to obtain the information of the next hop camera in advance.

In the first possible implementation method of the first aspect, the predicting the next hop camera includes: predicting the next hop camera according to the information of the target camera and/or the information of the target object. This solution provides a specific solution for predicting the next hop camera.

In the second possible implementation method of the first aspect, the predicting the next-hop camera according to the information of the target camera and/or the information of the target object specifically includes: according to the information of the target camera and / Or the information of the target object, predict the geographic location of the next hop area; and output a list of the next hop cameras according to the geographic location of the next hop area. This solution provides a specific solution for predicting the next hop camera.

In the third possible implementation method of the first aspect, the prediction of the next hop camera based on the information of the target camera and/or the information of the target object specifically includes any one or a combination of the following : The first type: statistics the historical motion trajectory of the object shot by the target camera, and predict the next hop camera based on the statistical historical motion trajectory of the object; the second type: according to the target object leaving the first Predict the next-hop camera based on the position and direction of movement in a region; the third type: Count the historical motion trajectory of the target object leaving the first region, and predict based on the statistical historical motion trajectory of the target object The next-hop camera; the fourth type: predict the next-hop camera based on the geographic location information where the target camera is located. This program provides several specific solutions for predicting the next hop camera.

In the fourth possible implementation method of the first aspect, the number of the next-hop cameras includes at least two; the predicted next-hop cameras include: each next-hop camera and each next-hop that output predictions The confidence level of the camera, which is used to characterize the possibility that each next-hop camera will capture the target object. This solution further provides confidence in the prediction when predicting the next hop camera.

In the fifth possible implementation method of the first aspect, the method further includes: predicting the possible position of the target object after entering the area captured by the next-hop camera; sending the position to at least one of the next-hop cameras An adjustment signal, where the position adjustment signal is used to instruct the next-hop camera that receives the adjustment signal to adjust the possible position to a visible position. This solution further predicts the possible position of the target object after entering the area photographed by the next-hop camera, which can facilitate camera adjustment and detect the target object entering the monitoring range in advance.

In the sixth possible implementation method of the first aspect, the method further includes: according to the information of the target camera and the information of the next hop camera, instructing the video of the target camera and the next hop respectively The video of the camera is played.

In a second aspect, a monitoring device is provided, including: an acquisition module for acquiring a monitoring task that indicates a target object; an analysis module for determining a target camera based on characteristic information of the target object, the target The camera is used to monitor the first area, and the target object is located in the first area at the current moment; the prediction module is used to predict the next hop camera, and the area monitored by the next hop camera is the predicted target object The surveillance area that is entered at the next moment; a sending module is used to send the information of the target camera and the information of the next hop camera.

This solution and various aspects of the solution correspond to each possible implementation method of the first aspect. And has corresponding beneficial effects.

In a third aspect, a computer-readable medium is provided, and the computer-readable storage medium stores instructions. When a computing processor executes the instructions, the first aspect and various possible implementation manners of the first aspect can be executed, and corresponding Technical effect.

In a fourth aspect, a computer program product is provided. The computer program product contains instructions. When a computing processor executes the instructions, the first aspect and various possible implementation manners of the first aspect can be executed, and have corresponding technical effects. .

In a fifth aspect, a monitoring method is provided, including: triggering a monitoring task, the monitoring task indicating a target object; receiving information of the target camera and information of the next hop camera, wherein the target camera is used to monitor a first area, The target object is located in the first area at the current moment, and the area monitored by the at least one next-hop camera is the predicted surveillance area that the target object will enter at the next moment. This program introduces a monitoring method for the predicted next hop camera.

The first possible implementation method of the fifth aspect, further comprising: obtaining the confidence level of each next hop camera, the confidence level being used to characterize the probability that each next hop camera will capture the target object at the next moment . The program further introduces the method of obtaining confidence.

The second possible implementation method of the fifth aspect further includes: respectively acquiring the multimedia data captured by the target camera and the at least one next-hop camera according to the information of the target camera and the information of the at least one next-hop camera. The multimedia data shot by a one-hop camera; the multimedia data shot by the target camera and the multimedia data shot by the at least one next-hop camera are respectively played on the screen. This program introduces a playback program.

The third possible implementation method of the fifth aspect further includes: selecting a display mode of the multimedia data captured by at least the next hop camera according to the confidence of each next hop camera. The program introduces a display mode.

In the fourth possible implementation method of the fifth aspect, according to the confidence of each next hop camera, selecting the display mode of the multimedia data captured by the at least next hop camera specifically includes: Confidence, use the big screen to play the multimedia data captured by the next hop camera with high confidence; or, according to the confidence of each next hop camera, use the small screen to play the multimedia data shot by the next hop camera with low confidence data. This program introduces a display mode based on confidence.

In a sixth aspect, a monitoring device is provided, including: a task module for triggering a monitoring task, the monitoring task indicating a target object; a processing module for receiving information about the target camera and information about the next hop camera, where all The target camera is used to monitor a first area, the target object is located in the first area at the current moment, and the area monitored by the at least one next-hop camera is the predicted surveillance area that the target object will enter at the next moment . This solution and various aspects of the solution correspond to each possible implementation method of the first aspect. And has corresponding beneficial effects.

In a seventh aspect, a computer-readable medium is provided. The computer-readable storage medium stores instructions. When a computing processor executes the instructions, the first aspect and various possible implementation manners of the first aspect can be executed, and corresponding Technical effect.

In an eighth aspect, a computer program product is provided. The computer program product contains instructions. When a computing processor executes the instructions, the first aspect and various possible implementation manners of the first aspect can be executed, and have corresponding technical effects .

Description of the drawings

In order to explain the technical solutions of the embodiments of the present invention more clearly, the following will briefly introduce the figures used in the embodiments and the prior art description. Obviously, the figures in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other graphs can also be obtained from these graphs.

Fig. 1 is a structural diagram of an embodiment of a monitoring system;

Figure 2 is a flowchart of an embodiment of a monitoring method;

Figure 3 is an example of the monitoring range;

Figure 4 is an example of alarm information;

Figure 5 is an example of the confidence of different cameras;

Figure 6(a)-6(b) is a schematic diagram of the target person moving between cameras;

Figure 7 is a schematic diagram showing the method;

Figure 8 is a flowchart of an embodiment of a monitoring method;

Figure 9 is a structural diagram of an embodiment of a monitoring device;

Figure 10 is a structural diagram of an embodiment of a monitoring device;

Figure 11 is a schematic diagram of an embodiment of a monitoring device.

Detailed ways

Fig. 1 is a structural diagram of an embodiment of a monitoring system. The monitoring system includes: a data analysis platform 11, a multimedia data center 12 communicating with the data analysis platform 11, and a display platform 13. The display platform 13 communicates with camera 141 (camera A), camera 142 (camera B), camera 143 (camera C), camera 144 (camera D), and camera 145 (camera E); or, the display platform 13 communicates with The multimedia data center communicates.

Fig. 2 is a flowchart of an embodiment of a monitoring method.

Step 21: The data analysis platform 11 obtains a monitoring task. The monitoring task includes a monitoring scope and a monitoring target list. This step can be executed by the task release module of the data analysis platform 11.

The triggering of the monitoring task can be generated on the data analysis platform 11 through user input. For example, the data analysis platform 11 receives the input monitoring range and monitoring list input by the user through a user interface (UI), and starts it in the UI. After the monitored button is clicked, the monitoring task is started. In addition, the monitoring task can also come from other devices, such as a mobile terminal/personal computer/server communicating with the data analysis platform; or, from the display platform 13.

The camera ID that needs to be monitored this time can be determined from the monitoring range. For example, the monitoring range may be a monitoring camera list in which the camera IDs to be monitored are directly recorded; or the monitoring range records the geographic coordinate range, and the camera IDs to be monitored can be determined from the geographic coordinate range. Referring to FIG. 3, the monitoring range in this embodiment directly describes the ID of the monitoring camera. Camera A, camera B, camera C, camera D, and camera E are monitored, and camera F is not monitored.

The camera ID may be the serial number or code of the camera, and may also be the address of the camera (for example, IP address, MAC address). All information that can directly or indirectly distinguish a certain camera from other cameras belongs to the camera ID described in the embodiment of the present invention.

The monitoring list carries: a monitoring target ID. The surveillance target is a moving object that needs to be tracked by the camera. When the monitoring target is a person, the monitoring target ID may be the target person's number, the target person's passport number, the target person's ID card number, and other information that can distinguish the target person from other persons. When the monitoring target is a vehicle, the monitoring target ID may be the license plate of the target vehicle. Optionally, the monitoring list may further carry monitoring target characteristics. The target feature is used to subsequently match the target person. According to actual needs, the monitoring target characteristics may also be stored in a local or non-local storage device for the data analysis platform 11 to use the monitoring target ID for query. Optionally, the server 11 may also store one or more of the name, gender, and age of the target person.

Step 22: The data analysis platform 11 uses the target person characteristics to match the person characteristics of the persons in the video shot by the camera (video person characteristics). Wherein, the multimedia data center 12 communicates with multiple cameras, and can obtain the IDs of the multiple cameras and the images of persons captured by the multiple cameras.

The data analysis platform 11 acquiring the characteristics of the person in the video includes: the data analysis platform 11 sends the monitoring range to the multimedia data center 12, and the multimedia data center 12 follows the camera IDs listed in the monitoring range, The videos taken by these cameras are sent to the data analysis platform 11, and the data analysis platform 11 extracts the video person characteristics from the received videos. This transmission can be real-time. In addition, a camera may also extract the video person characteristics, and then send the extracted person characteristics to the data analysis platform 11 for matching.

The multimedia data center 12 may be integrated into the data analysis platform 11. In this case, the data analysis platform 11 obtains the videos (or the characteristics of the people in the videos) taken by these cameras by itself.

The specific method for matching the characteristics of the target person and the characteristics of the video person may be to compare the similarity between the two. When the value of the similarity between the feature of the video person and the feature of the target person reaches the preset threshold, this means that the matched video person and the target person are more likely to be the same person, and an alarm signal of successful matching can be issued. Optionally, the alarm signal may further include information such as the name, gender, and age described in FIG. 4. Figure 4 is a schematic diagram of an alarm signal, in which the name of the target person is Jason, he appears in the video of camera A, his gender is male and his age is 25 years old.

Assuming that the matched video comes from the camera 141, the data analysis platform 11 may further send the ID of the camera 141 (camera A) to the display platform 13. After the display platform 13 receives the ID of the first camera, it can obtain the video captured by the first camera in real time through the camera 141 (or the multimedia data center 12) for playback.

A method for comparing similarity is: the target person characteristic acquired by the data analysis platform 11 from the target list (or local storage device or remote storage device) is a 256-dimensional floating point (float) array ; The data analysis platform 11 obtains the video frame from the multimedia data center 12, parses the person image in the frame into a 256-dimensional floating-point array (video person characteristics), compares the two floating-point arrays, and compares the similarity of the floating-point arrays The degree of similarity is used as the similarity between the target person and the video person.

Step 23: When the matching is successful, the data analysis platform 11 predicts that when the target person leaves the monitoring range of camera A (the first monitoring range) at the current moment, the monitoring range that the target person will enter at the next moment (the second monitoring range) Range) corresponds to the camera. The instruction display platform 13 extracts the video from the prediction to the camera. When the data analysis platform issues this instruction, the target person has not yet entered the second monitoring range. The time point to start the prediction may be: after the matching is successful, for example: after the matching is successful, and the target person is in the first monitoring range; or after the matching is successful, and the target person has left the first monitoring range, When it has not entered the monitoring range of the next camera.

Here is the relationship between the current moment and the next moment. The current moment refers to the point in time when the matching is performed; the next moment refers to the point in time when the target person leaves the current monitoring range and immediately enters the monitoring range of the next camera. Between the current moment and the next moment, the target person does not enter the monitoring range of other cameras. For example, during matching, the target person is in the monitoring range of camera A, and it is expected that the target person will enter the monitoring range of camera B and the monitoring range of camera F in sequence. Then the monitoring area that the target person enters at the next moment refers to the monitoring range of camera B (not the monitoring range of camera F).

Due to the uncertainty in the future, it is predicted that the number of cameras may be more than one, and the confidence that the target person enters the monitoring range of different cameras is different. Refer to Figure 5, after the target person Jason leaves the monitoring range of camera A, the cameras corresponding to the monitoring range that will be entered include: camera B, camera C, camera D, and camera E, but it is possible for Jason to enter these monitoring ranges Different, he has the greatest probability of entering camera C, reaching 60%, while the probability of entering the monitoring range of camera E is only 3%. The data analysis platform 11 has multiple options for sending videos to the cameras of the display platform. For example: select the cameras with the top k confidence levels, and instruct the display platform 13 to obtain the videos of these cameras; select the cameras with the confidence levels greater than the confidence threshold to send the videos to the display platform.

Optionally, the data analysis platform 11 may further predict the possible position of the target person entering the next camera monitoring range. Referring to Fig. 6(a), the monitoring range of the camera C is the monitoring range 31; the monitoring range of the camera A is the monitoring range 32. According to the information obtained from camera A, the data analysis platform 11 can obtain the current geographic location of the target person Jensen (75° east longitude, 31.5° north latitude), which is within the monitoring range 32 of camera A, and camera A shoots The direction of movement to Jensen is north. According to his movement trajectory, he will reach a new coordinate (75° east longitude, 32.5° north latitude). This new coordinate belongs to the monitoring range 31 of the camera C, specifically at the boundary of the monitoring range 31. In other words, it is expected that Jason will enter the surveillance range of camera C from the south of camera C. However, the visible range of the camera at the current moment is the range 311, which means that even if Jensen enters the monitoring range 31, the camera C cannot immediately monitor him.

To this end, in this embodiment, the data analysis platform 11 sends adjustment information to the camera C. The adjustment information may include the position of the target person, that is, (75° east longitude, 32.5° north latitude); or, the adjustment information may include an adjustment direction (adjust the visible range of the camera C to the south). Referring to Figure 6(b), when the camera C adjusts its visual range according to this adjustment information, (75°E, 32.5°N) is included in its monitoring range, which means that after Jensen enters the monitoring range 31, it will Was monitored immediately. Therefore, this embodiment has better monitoring timeliness. According to a similar method, the data analysis platform 11 can generate adjustment information corresponding to the camera B and the camera D, and send it to the camera B and the camera D respectively, so that when Jensen enters the monitoring range of the camera B or the camera D, he Can be monitored as soon as possible.

Several confidence calculation methods are provided below.

Method 1: When the target person leaves the first area this time, record the position when he leaves and the direction of movement when he leaves. Because the walking rule of the person is to continue moving in the original direction in most cases, it can be estimated The subsequent movement direction of the personnel is taken out, so that the position reached by continuing to advance in the original direction (this position belongs to the second area) is taken as the possible position of the entered monitoring range. That is, it predicts the monitoring range that the target person is about to enter, and predicts the possible location of entering the monitoring range.

Method 2: Obtain the historical movement track of the target person in the first area from a database, and count the position and movement direction of the target person leaving the first area. When the target person leaves the first area from the same position this time, the movement direction obtained by statistics is used as the most likely movement direction of the target person this time, so as to predict the monitoring that the target person is about to enter. Range and possible locations to enter the monitoring range. Or, the next monitoring range obtained by statistics is taken as the monitoring range that may enter this time, so as to predict the monitoring range that the target person is about to enter and the possible location of entering the monitoring range. Optional statistics based on time period. For example: the target person will enter the monitoring range of camera D from the monitoring range of camera A in 55% of the cases from 8:00 to 10:00 in the morning, then the following prediction can be made: In this monitoring, the The target person will enter the monitoring range of camera D from the monitoring range of camera A in 55% of the cases.

Method 3: Calculate the movement laws of all persons in the first monitoring range recorded in the database, and obtain the proportion of these persons entering the monitoring range of other cameras, thereby predicting the monitoring range that the target person is about to enter. For example, in the past, 1,000 people left the first monitoring range, and 400 of these 1,000 people entered the monitoring range of camera B next. Then we can consider that the confidence that the target person enters the monitoring range of the camera B this time is 400/1000=40%. This method can also predict the possible position of the target person entering the next monitoring range. For example, among the 400 people, 300 people entered the surveillance range of the camera B from the south. This means that when the target person enters the monitoring range of the camera B this time, there is a 300/400=75% chance that the target person enters the monitoring range of the camera B from the south, so it can The visual range of the camera B is adjusted in advance, and after the adjustment, the camera B regards the true south of its monitoring range as the visual range. The program can also be used for statistics and forecasts in time periods.

Method 4: Predict at least one next hop camera based on the geographic location information where the target camera is located. For example: geographically, there are 3 cameras around camera A, then these 3 cameras may all be the next hop cameras. Or there is a camera A in the middle of a road, and a camera B and a camera C at both ends of the road. If the target person is currently located in the monitoring range of the camera A, the target object may enter the monitoring range of the camera B or the monitoring range of the camera C next. It can be considered that at the next moment the target person has a 50% chance of entering the monitoring range of camera B, and a 50% chance of entering the monitoring range of camera C.

You can also combine 2 or 3 or 4 of the above methods to set different weight values. For example, the weights of method 1, method 2, and method 3 are 40%, 40%, and 20% respectively. According to method 1, method 2, and method 3, the confidence that the target person enters the monitoring range of camera B is obtained respectively 30%, 50% and 40%, then based on the weights, the confidence that the target person enters the camera B is: 30%*40%+50%*40%+40%*20%=40%.

Step 24: The display platform 13 receives an instruction from the data analysis platform, the instruction carries the IDs of the K cameras, and caches the real-time video information of the k cameras. After certain video (or image) data is cached, the display platform 13 can play the video of the k cameras. This is because whether the display platform 13 obtains the video information of k cameras is triggered by the prediction result, not by the successful detection of the target person in the second monitoring range. Therefore, when the target person has not entered the second monitoring range, the data analysis platform can already instruct the K cameras to send videos for the display platform 13 to buffer. This means that when the target person has not entered (or just) entered the second area, the display platform 13 can already display the actions of the target person.

Another display method is: after the target person enters the second monitoring range, the camera C sends the video to the data analysis platform 11, and the data analysis platform 11 performs the method described in step 22 Matching, the video is not sent to the display platform 13 for caching until the matching is successful. Compared with this other display method, the display speed of this embodiment is faster, and seamless switching of surveillance videos can be achieved under ideal conditions.

It should be noted that, in this embodiment: (1) After the data analysis platform 11 detects the target person (see the matching scheme in step 22), the data analysis platform 11 instructs the display platform 13 to select from camera A or Obtain the real-time video of camera A from the multimedia data center 12; (2) When the data analysis platform 11 predicts the monitoring range of the next hop camera that the target person will enter, the data analysis platform 11 instructs the display platform 13 Get real-time video of at least one next hop camera. The two commands in (1) and (2) can be sent at the same time in the order of time, or any one of them can be in the front. In addition, in addition to instructing the display platform 13 to obtain the video from the camera, the camera may also be instructed to send the video to the display platform 13, or the multimedia data center 12 may be instructed to send the video of the camera to the display platform 13. Or instruct the display platform 13 to generate video from the camera to the display platform. The camera video displayed by the display platform may be a real-time video of the camera, or in some cases, it may also be a non-real-time video with a certain lag.

Refer to Figure 7 for a schematic diagram of a display method. As shown in the figure, the display platform 13 has four screens, namely a large screen 41, a small screen 42, a small screen 43, and a small screen 44. Before Jensen leaves the monitoring area 32, the content of camera A is played on the big screen, and the display platform 13 caches the content of the top 3 cameras (that is, camera B, camera C, and camera D) in advance; After Sen leaves the monitoring range 32, since the confidence of camera C is the highest, the large screen is switched to play the video of camera C; the small screen 41 and the small screen 42 play the video of camera B and camera D respectively (with the confidence The order is consistent). Or: Select the corresponding screen to play in the order of the degree of confidence. For example, camera C with the highest confidence selects the leftmost screen to play, and camera E with the lowest confidence uses the rightmost screen to play.

In addition, the display platform can also discriminately play the multimedia data of the current camera and the predicted multimedia data of the next hop camera. For example: the multimedia data of the current camera is played on the big screen, and the multimedia data of the next hop camera is played on the small screen.

Step 25: The camera that has received the adjustment adjusts its visual range. The adjusted visual range includes the position where the target person enters the surveillance range of the camera. It should be noted that this step is an optional step and is only executed when the data analysis platform 11 sends adjustment information.

Referring to Figure 6(b), when the camera C adjusts its visual range according to this adjustment information, (75°E, 32.5°N) is included in its monitoring range, which means that after Jensen enters the monitoring range 31, it will Was monitored immediately. Therefore, it has better monitoring timeliness.

The above process is an example of video matching and caching. However, this method is not only suitable for video, in other embodiments, media data such as pictures and sounds can also be matched and cached in advance.

The following describes a flowchart of another embodiment of a monitoring method in conjunction with FIG. 8. The principle of this embodiment is the same as that of the embodiment described in FIG. 2, and the following description will focus on the differences.

Step 41: The display platform 13 sends a monitoring task to the data analysis platform 11. The monitoring task includes a monitoring scope and a monitoring target list. Regarding the contents and functions of the monitoring scope and monitoring target list, refer to step 21. Compared with step 21 of the previous embodiment, this step differs in that the trigger of the monitoring task is different.

Step 42: The data analysis platform 11 uses the target person characteristics to match the person characteristics of the person in the video shot by the camera (video person characteristics), and instructs the display platform to obtain the video of the camera that is successfully matched. Wherein, the multimedia data center 12 communicates with multiple cameras, and can obtain the IDs of the multiple cameras and the images of persons captured by the multiple cameras. The data analysis platform 11 is, for example, a cloud computing platform. This step is similar to step 22, and you can refer to step 22 for details, so it will not be repeated.

Step 43: When the matching is successful, the data analysis platform 11 predicts that when the target person leaves the monitoring range of camera A (the first monitoring range) at the current moment, the monitoring range that the target person will enter at the next moment (the second monitoring range) Range) corresponds to the camera. Send an instruction to instruct the display platform to acquire at least one of the predicted cameras. This step is similar to step 23, except that in step 23, among the cameras predicted in step 23, k cameras are screened according to their confidence, and the IDs of these k cameras are carried in the instruction and sent to the display platform 13; In the step, the predicted IDs of all cameras are sent to the display platform 13 through the instruction. In addition, the instruction further carries the confidence level corresponding to each camera.

Step 44: The display platform 13 receives an instruction from the data analysis platform, and the instruction carries the camera ID and the confidence level corresponding to each camera ID. The display platform screens out k cameras according to the confidence levels corresponding to each camera ID, and obtains the videos of these k cameras for caching and display. The screening method is for example: selecting the camera with the top k confidence level; selecting the camera with the confidence level greater than the confidence threshold.

It can be seen that, compared with the embodiment described in FIG. 2, this embodiment moves the operation of screening out k cameras from the data analysis platform 11 to the display platform 13. For the rest, please refer to step 24, which will not be repeated here.

Step 45, refer to step 25, which will not be repeated here.

Figure 9 provides an embodiment of a monitoring device. The monitoring device 9 may be a data analysis platform 11 or a program running in the data analysis platform 11. The monitoring device 9 includes: an acquisition module 51 for acquiring a monitoring task that indicates a target object; an analysis module 52, which communicates with the acquisition module 51, for determining a target based on the characteristic information of the target object The target camera is used to monitor the first area, and the target object is located in the first area at the current moment; the prediction module 53, which communicates with the analysis module 52, is used to predict the next hop camera, the next The area monitored by the one-hop camera is the predicted surveillance area that the target object will enter at the next moment; the sending module 54 communicates with the analysis module 52 and the prediction module 53, and is used to send the target camera information and the Information of the next hop camera.

Unless otherwise specified, the monitoring device 5 can execute the method in FIG. 2. Specifically, step 21 can be performed by the acquisition module 51; step 22 can be performed by the analysis module 52; step 23 can be performed by the prediction module 53; the instructions in step 24 and all of the instructions in step 23 The adjustment information can be executed by the sending module 54. Since the specific functions of each module have been described in detail in the previous method embodiments, they will not be repeated here.

Figure 10 provides another embodiment of a monitoring device. The monitoring device 6 may be a display platform 13 or a program running on the display platform 13. The monitoring device 6 includes: a task module 61 for triggering a monitoring task, the monitoring task indicating a target object; a processing module 62 for receiving information about the target camera and information about the next hop camera, wherein the target camera It is used to monitor the first area, the target object is located in the first area at the current moment, and the area monitored by the at least one next-hop camera is the predicted surveillance area that the target object will enter at the next moment.

Optionally, the monitoring device 6 further includes a confidence level acquiring module 62: configured to acquire the confidence level of each next hop camera, and the confidence level is used to characterize that each next hop camera captures the camera at the next moment. The likelihood of the target object.

Optionally, the monitoring device 6 further includes a multimedia acquisition module 63, configured to obtain the multimedia data captured by the target camera according to the information of the target camera and the information of the at least one next hop camera. And the multimedia data captured by the at least one next hop camera; the playback module 64 is configured to respectively play the multimedia data captured by the target camera and the multimedia data captured by the at least one next hop camera on the screen.

Unless otherwise specified, the monitoring device 6 may execute the method executed by the display platform 13 described in FIG. 8. Specifically, step 41 can be performed by the task module 61; in step 44, the confidence level can be obtained by the confidence level obtaining module 62, the multimedia data can be obtained by the multimedia obtaining module 63, and the playback operation can be performed by the playing module 64.

Referring to Fig. 11, the present invention also provides an embodiment of a monitoring device. The monitoring device 7 includes a processor 71 and an interface 72 for executing the method in steps 21-25. The processor 71 is configured to execute: obtain a monitoring task, the monitoring task indicates a target object; according to the characteristic information of the target object, determine a target camera, the target camera is used to monitor the first area, the target object is The current moment is located in the first area; the next-hop camera is predicted, and the area monitored by the next-hop camera is the predicted surveillance area that the target object will enter at the next moment. The interface is used for executing: sending the information of the target camera and the information of the next hop camera. The monitoring device may be a data analysis platform 11. For the specific execution process, refer to step 21 to step 25 and FIG. 2. Optionally, the monitoring device 7 further includes a memory (for example, a memory), the memory is used to store a computer program, and the processor 71 is used to execute steps 21-25 by running the computer program in the memory. method.

The present invention also provides an embodiment of a computer-readable medium, the computer-readable storage medium stores instructions, which are used to execute the method of steps 21-25 when the computing processor executes the instructions.

The present invention also provides a computer program product, characterized in that the computer program product contains instructions, and when the computing processor executes the instructions, it is used to perform the method of the steps 21-25 of the claims.

The invention also provides an embodiment of the display device. The display device (for example, the display platform 13) includes a processor and an interface, and the processor is used to execute the methods in steps 41-45. Optionally, the display device further includes a memory (for example, a memory), the memory is used to store a computer program, and the processor for displaying the video is used to execute steps 41-45 by running the computer program in the memory. method.

The present invention also provides an embodiment of a computer-readable medium, the computer-readable storage medium stores instructions, which are used to execute the method of steps 41-45 when the computing processor executes the instructions.

The present invention also provides a computer program product, characterized in that the computer program product contains instructions, and when the computing processor executes the instructions, it is used to execute the method of the steps 41-45 of claims.

As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions recorded in the embodiments are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

A monitoring method, characterized in that it comprises:

Acquiring a monitoring task, where the monitoring task indicates a target object;

Determine a target camera according to the characteristic information of the target object, where the target camera is used to monitor a first area, and the target object is located in the first area at the current moment;

Predicting a next-hop camera, where the area monitored by the next-hop camera is the predicted monitoring area that the target object will enter at the next moment;

Send the information of the target camera and the information of the next hop camera.
The monitoring method according to claim 1, wherein the predicting the next hop camera comprises:

Predict the next hop camera based on the information of the target camera and/or the information of the target object.
The monitoring method according to claim 2, wherein the predicting the next hop camera according to the information of the target camera and/or the information of the target object specifically comprises:

Predict the geographic location of the next hop area according to the information of the target camera and/or the information of the target object;

According to the geographic location of the next hop area, output the list of next hop cameras.
The monitoring method according to claim 2 or 3, wherein the prediction of the next hop camera based on the information of the target camera and/or the information of the target object specifically includes any one of the following: Several combinations:

The first method: Calculate the historical motion trajectory of the object shot by the target camera, and predict the next-hop camera based on the statistical historical motion trajectory of the object;

The second type: predict the next-hop camera according to the position and movement direction of the target object when it leaves the first area this time;

The third type: statistics the historical movement track of the target object leaving the first area, and predict the next-hop camera based on the statistical historical movement track of the target object;

The fourth type: predict the next hop camera based on the geographic location information where the target camera is located.
The monitoring method according to any one of claims 1 to 4, wherein the number of the next hop camera includes at least two; and the predicted next hop camera includes:

The predicted confidence of each next hop camera and each next hop camera is output, and the confidence is used to characterize the probability that each next hop camera captures the target object.
The monitoring method according to any one of claims 1-5, wherein the method further comprises:

Predict the possible position of the target object after entering the area captured by the next-hop camera;

Send a position adjustment signal to at least one of the next hop cameras, where the position adjustment signal is used to instruct the next hop camera that has received the adjustment signal to adjust the possible position to a visible position.
The monitoring method according to any one of claims 1-5, wherein the method further comprises:

According to the information of the target camera and the information of the next hop camera, instruct to play the video of the target camera and the video of the next hop camera respectively.
A monitoring method, characterized in that it comprises:

Trigger a monitoring task, the monitoring task indicates the target object;

Receive information about a target camera and information about a next-hop camera, where the target camera is used to monitor a first area, the target object is located in the first area at the current moment, and the at least one next-hop camera monitors The area of is the predicted monitoring area that the target object will enter at the next moment.
The monitoring method according to claim 8, further comprising:

Obtain the confidence of each next-hop camera, where the confidence is used to characterize the possibility that each next-hop camera will capture the target object at the next moment.
The monitoring method according to claim 8 or 9, further comprising:

Acquiring the multimedia data captured by the target camera and the multimedia data captured by the at least one next hop camera, respectively, according to the information of the target camera and the information of the at least one next hop camera;

The multimedia data captured by the target camera and the multimedia data captured by the at least one next-hop camera are respectively played on the screen.
The monitoring method according to claim 10, further comprising:

According to the confidence of each next hop camera, the display mode of the multimedia data captured by the at least the next hop camera is selected.
The monitoring method according to claim 11, wherein, according to the confidence level of each next hop camera, selecting the display mode of the multimedia data shot by the at least the next hop camera specifically includes:

According to the confidence of each next hop camera, use a large screen to play the multimedia data captured by the next hop camera with high confidence; or,

According to the confidence of each next hop camera, a small screen is used to play the multimedia data captured by the next hop camera with low confidence.
A monitoring device, characterized in that it comprises:

An acquisition module, for acquiring a monitoring task, the monitoring task indicating a target object;

An analysis module, configured to determine a target camera according to the characteristic information of the target object, the target camera is used to monitor a first area, and the target object is located in the first area at the current moment;

The prediction module is used to predict the next hop camera, and the area monitored by the next hop camera is the predicted monitoring area that the target object will enter at the next moment;

The sending module is used to send the information of the target camera and the information of the next hop camera.
The monitoring device according to claim 13, wherein the prediction module is specifically configured to:

Predict the next hop camera based on the information of the target camera and/or the information of the target object.
The monitoring device according to claim 14, wherein the prediction module is specifically configured to:

Predict the geographic location of the next hop area according to the information of the target camera and/or the information of the target object;

According to the geographic location of the next hop area, output the list of next hop cameras.
The monitoring device according to claim 14 or 15, wherein the prediction module is specifically configured to perform any one or a combination of the following:

The first method: Calculate the historical motion trajectory of the object shot by the target camera, and predict the next-hop camera based on the statistical historical motion trajectory of the object;

The second type: predict the next-hop camera according to the position and movement direction of the target object when it leaves the first area this time;

The third type: Calculate the historical movement track of the target object leaving the first area, and predict the next-hop camera based on the statistical historical movement track of the target object;

The fourth type: predict the next hop camera based on the geographic location information where the target camera is located.
The monitoring device according to any one of claims 13-16, wherein the number of the next-hop cameras is at least two; and the prediction module is further configured to:

The predicted confidence of each next hop camera and each next hop camera is output, and the confidence is used to characterize the probability that each next hop camera captures the target object.
The monitoring device according to any one of claims 13-17, wherein the prediction module is further configured to:

Predict the possible position of the target object after entering the area captured by the next-hop camera;

Send a position adjustment signal to at least one of the next hop cameras, where the position adjustment signal is used to instruct the next hop camera that has received the adjustment signal to adjust the possible position to a visible position.
The monitoring device according to any one of claims 13-17, wherein the device further comprises:

According to the information of the target camera and the information of the next hop camera, instruct to play the video of the target camera and the video of the next hop camera respectively.
A monitoring device, characterized in that it comprises:

The task module is used to trigger a monitoring task, and the monitoring task indicates a target object;

The processing module is configured to receive information of a target camera and information of a next-hop camera, wherein the target camera is used to monitor a first area, the target object is located in the first area at the current moment, and the at least one next-hop camera The area monitored by the one-hop camera is the predicted monitoring area that the target object will enter at the next moment.
The monitoring device according to claim 20, wherein the monitoring device further comprises a confidence degree obtaining module:

It is used to obtain the confidence of each next hop camera, and the confidence is used to characterize the possibility that each next hop camera will capture the target object at the next moment.
The monitoring device according to claim 20 or 21, further comprising:

The multimedia acquisition module is configured to acquire the multimedia data captured by the target camera and the multimedia data captured by the at least one next hop camera, respectively, according to the information of the target camera and the information of the at least one next hop camera ；

The playing module is configured to respectively play the multimedia data shot by the target camera and the multimedia data shot by the at least one next-hop camera on the screen.
The monitoring device according to claim 22, wherein the playing module is specifically configured to:

According to the confidence of each next hop camera, the display mode of the multimedia data captured by the at least the next hop camera is selected.
The monitoring device according to claim 23, wherein the playing module is specifically configured to:

According to the confidence of each next hop camera, use a large screen to play the multimedia data captured by the next hop camera with high confidence; or,

According to the confidence of each next hop camera, a small screen is used to play the multimedia data captured by the next hop camera with low confidence.
A monitoring device includes a processor and an interface, and the processor is used to execute:

Acquiring a monitoring task, where the monitoring task indicates a target object;

Determine a target camera according to the characteristic information of the target object, where the target camera is used to monitor a first area, and the target object is located in the first area at the current moment;

Predicting a next-hop camera, where the area monitored by the next-hop camera is the predicted monitoring area that the target object will enter at the next moment;

The interface is used to execute:

Send the information of the target camera and the information of the next hop camera.
The monitoring device according to claim 25, wherein the processor is configured to:

Predict the next hop camera based on the information of the target camera and/or the information of the target object.
The monitoring device according to claim 26, wherein the processor is configured to:

Predict the geographic location of the next hop area according to the information of the target camera and/or the information of the target object;

According to the geographic location of the next hop area, output the list of next hop cameras.
The monitoring device according to claim 26 or 27, wherein the processor predicts the next hop camera according to the information of the target camera and/or the information of the target object, specifically including any one of the following Kind or combination of several:

The first method: Calculate the historical motion trajectory of the object shot by the target camera, and predict the next-hop camera based on the statistical historical motion trajectory of the object;

The second type: predict the next-hop camera according to the position and movement direction of the target object when it leaves the first area this time;

The third type: statistics the historical movement track of the target object leaving the first area, and predict the next-hop camera based on the statistical historical movement track of the target object;

The fourth type: predict the next hop camera based on the geographic location information where the target camera is located.
The monitoring device according to any one of claims 20-28, wherein the number of the next-hop cameras is at least two; and the processor predicts that the next-hop cameras include:

The predicted confidence of each next hop camera and each next hop camera is output, and the confidence is used to characterize the probability that each next hop camera captures the target object.
The monitoring device according to any one of claims 20-29, wherein the processor is further configured to:

Predict the possible position of the target object after entering the area captured by the next-hop camera;

Send a position adjustment signal to at least one of the next hop cameras, where the position adjustment signal is used to instruct the next hop camera that has received the adjustment signal to adjust the possible position to a visible position.
The monitoring device according to any one of claims 20-29, wherein the processor is further configured to:

According to the information of the target camera and the information of the next hop camera, instruct to play the video of the target camera and the video of the next hop camera respectively.
A computer-readable medium, wherein the computer-readable storage medium stores instructions, which are used to execute the method of claims 1-7 when the computing processor executes the instructions.
A computer program product, characterized in that the computer product contains instructions, which are used to execute the method of claims 1-7 when the computing processor executes the instructions.