Multi-camera linkage monitoring method and system

Abstract

The invention discloses a multi-camera linkage monitoring method and system, and the method comprises the steps: obtaining the first image data of a first camera, and enabling the first image data to comprise the data of a target person; performing analysis based on the character data and a pre-trained extraction model, and determining character identification data; sending the character identification data to a second camera, collecting second image data of the target character by the second camera, and performing data compression based on the character identification data to form compressed data; receiving compressed data of the second camera, performing data restoration based on the character identification data to obtain restored data, and performing monitoring analysis based on the first image data and the restored data; according to the scheme, the data volume transmitted by the camera can be reduced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and more particularly to a multi-camera linkage monitoring method and system. Background Art

[0002] In the existing camera monitoring process, all the captured images usually need to be transmitted to the analysis center for monitoring and analysis.

[0003] However, directly transmitting all the image data requires a large amount of data to be transmitted. Summary of the Invention

[0004] The present invention provides a camera linkage monitoring method and system, which can reduce the amount of data transmitted by the camera.

[0005] In order to solve the above technical problems, the present invention is implemented as follows: In a first aspect, the present application provides a multi-camera linkage monitoring method, the method comprising: obtaining first image data of a first camera, the first image data including person data of a target person; analyzing based on the person data and a pre-trained extraction model to determine person identification data; sending the person identification data to a second camera, the second camera collecting second image data of the target person, and compressing the data based on the person identification data to form compressed data; receiving the compressed data of the second camera, and restoring the data based on the person identification data to obtain restored data, so as to perform monitoring and analysis based on the first image data and the restored data.

[0006] Further, the step of the second camera obtaining the compressed data includes: obtaining the second image data, determining the target person based on the person identification data, and dividing the second image data into object data including the target person and background data; compressing the object data based on the person identification data to obtain first data; performing a second compression on the background data to obtain second data; forming compressed data based on the first data and the second data.

[0007] Further, the restoring the data based on the person identification data to obtain restored data includes: obtaining the compressed data, and extracting the first data and the second data; restoring the first data based on the person identification data to obtain first intermediate data; restoring the second data based on historical background data to obtain second intermediate data; determining the restored data based on the first intermediate data and the second intermediate data.

[0008] Furthermore, the first data includes the location data and posture data of the target person. The step of obtaining the first data includes: determining the location data of the target person based on the position of the object data in the second image data; determining multiple key points of the target person and determining the posture data of the target person based on the key points; and determining the first data based on the location data and posture data of the target person.

[0009] Furthermore, the process of restoring the first data based on the person identification data to obtain the first intermediate data includes: acquiring person-related data determined based on the first image data based on the person identification data; adjusting the person-related data based on the pose data of the target person to form the first intermediate data; and fusing the first intermediate data and the second intermediate data based on the position data of the target person.

[0010] Furthermore, the training data includes a first training image and a second training image from different cameras, wherein the first training image and the second training image contain the same person; the training steps of the extraction model include: inputting the first training image into the extraction model to determine the first extraction data; inputting the second training image into the extraction model to determine the second extraction data; and adjusting the extraction model based on the person recognition results in the first extraction data and the second training image, the second extraction data, and the person recognition results in the first training image.

[0011] Furthermore, the person identification data includes person feature information; the analysis based on person data and a pre-trained extraction model to determine person identification data includes: analyzing person data and a pre-trained extraction model to determine person feature information; determining the number of people in the second camera, and filtering the person feature information based on the number of people to form person identification data, wherein the more people there are, the more person feature information is contained in the person identification data; wherein, the person feature information includes person size information, and the person identification data includes size data adjusted according to the relative position of the first camera and the second camera; the person feature information includes person color data, and the person color data includes pixel values ​​of pixels in multiple regions of the person; the person identification data includes level data that divides pixel values ​​into corresponding levels, with pixel values ​​0-255 divided into eight levels; the second camera uploads second image data in the first time period, and uploads compressed data in the second time period after the first time period, so as to analyze the difference in the number of people and person feature information in the second camera based on the second image data to determine the person identification data corresponding to each person; the second camera periodically uploads second image data.

[0012] Secondly, this application provides a multi-camera linkage monitoring system, the system comprising: a first data acquisition module for acquiring first image data from a first camera, the first image data including person data of a target person; a person identification acquisition module for analyzing the person data and a pre-trained extraction model to determine person identification data; a compressed data acquisition module for sending the person identification data to a second camera, the second camera acquiring second image data of the target person, and compressing the data based on the person identification data to form compressed data; and a compressed data restoration module for receiving the compressed data from the second camera and restoring the data based on the person identification data to obtain restored data, for monitoring and analysis based on the first image data and the restored data.

[0013] Thirdly, this application provides an electronic device, including: a memory and at least one processor; the memory is used to store computer execution instructions; the at least one processor is used to execute the computer execution instructions stored in the memory, causing the at least one processor to perform the method as described in the first aspect.

[0014] Fourthly, this application provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0015] This application embodiment can be applied to security monitoring scenarios, utilizing multiple cameras for joint monitoring and analysis. Furthermore, this solution can compress images of the same person captured by multiple cameras, thereby reducing the amount of data that needs to be transmitted. Specifically, in this solution, the first camera captures data of the target person. Based on the target person's movement trajectory, the corresponding second camera is determined. This solution can analyze the person data and a pre-trained extraction model to determine person identification data; this identification data is then sent to the second camera. After the second camera identifies the target person based on the identification data, it can segment the captured second image data into object data containing the target person and background data, thereby performing corresponding data compression and transmitting the compressed data. Data restoration is then performed based on the person identification data to obtain restored data. After obtaining the data captured by the first camera and the restored data, monitoring and analysis can be performed based on the first image data and the restored data. This solution can compress images related to the person and the background of the person based on images of the same person captured by two cameras, thereby reducing the amount of data transmitted. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 It is a schematic flow chart of a multi-camera linkage monitoring method according to an embodiment of the present application; Figure 2 It is a schematic structural diagram of a multi-camera linkage monitoring system according to an embodiment of the present application. Specific Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] The embodiments of the present application can be applied to the security monitoring scenario. Multiple cameras can be used for joint monitoring and analysis. Moreover, this solution can use the same person captured by multiple cameras to perform image compression corresponding to the person, thereby reducing the amount of data that needs to be transmitted by the cameras. Specifically, in this solution, the first camera can capture person data including the target person, determine the corresponding second camera according to the movement trajectory of the target person. This solution can analyze based on the person data and a pre-trained extraction model to determine person identification data; and send the person identification data to the second camera. After the second camera identifies the target person according to the person identification data, it can divide the captured second image data into object data and background data including the target person, thereby performing corresponding data compression and transmitting the compressed data to restore the data based on the person identification data to obtain restored data. After obtaining the data captured by the first camera and the restored data, monitoring and analysis can be performed based on the first image data and the restored data. In this solution, the images related to the person can be compressed based on the same person captured by two cameras, and the background of the person can also be compressed, thereby reducing the amount of data transmitted.

[0019] Specifically, the embodiments of the present application provide a multi-camera linkage monitoring method, as Figure 1 shown, the method includes: Step 102, obtain the first image data of the first camera, and the first image data includes person data of the target person.

[0020] Step 104, analyze based on the person data and a pre-trained extraction model to determine person identification data.

[0021] Step 106, send the person identification data to the second camera, and the second camera collects the second image data of the target person and performs data compression based on the person identification data to form compressed data.

[0022] Step 108: Receive compressed data from the second camera and restore the data based on the person identification data to obtain restored data, so as to perform monitoring and analysis based on the first image data and the restored data.

[0023] This application embodiment can be applied to security monitoring scenarios, utilizing multiple cameras for joint monitoring and analysis. Furthermore, this solution can compress images of the same person captured by multiple cameras, thereby reducing the amount of data that needs to be transmitted. Specifically, in this solution, the first camera captures data of the target person. Based on the target person's movement trajectory, the corresponding second camera is determined. This solution can analyze the person data and a pre-trained extraction model to determine person identification data; this identification data is then sent to the second camera. After the second camera identifies the target person based on the identification data, it can segment the captured second image data into object data containing the target person and background data, thereby performing corresponding data compression and transmitting the compressed data. Data restoration is then performed based on the person identification data to obtain restored data. After obtaining the data captured by the first camera and the restored data, monitoring and analysis can be performed based on the first image data and the restored data. This solution can compress images related to the person and the background of the person based on images of the same person captured by two cameras, thereby reducing the amount of data transmitted.

[0024] This solution can identify people in images captured by a camera and form a positioning box. The image within the positioning box can be used as object data of the person, and the data outside the positioning box can be used as the background. Specifically, as an optional embodiment, the step of the second camera acquiring compressed data includes: acquiring second image data; determining the target person based on person identification data; and dividing the second image data into object data containing the target person and background data; compressing the object data based on the person identification data to obtain first data; performing a second compression based on the background data to obtain second data; and forming compressed data based on the first data and the second data. Since the background similarity is relatively large, a compression scheme with a higher compression ratio can be adopted. For the background data part, this solution can compare the background with the background of the previous few frames. If the difference is small, this part of the data can be directly omitted from transmission.

[0025] In the analysis, the data can be restored using corresponding methods based on the compressed first and second data. Specifically, as an optional embodiment, the data restoration based on the person identification data to obtain the restored data includes: acquiring compressed data and extracting the first and second data; restoring the first data based on the person identification data to obtain the first intermediate data; restoring the second data based on historical background data to obtain the second intermediate data; and determining the restored data based on the first and second intermediate data.

[0026] For compressing object data of people, this solution can identify and send the position and posture of the person in the camera section without transmitting the image of the person, thus greatly reducing the amount of data transmitted. Specifically, as an optional embodiment, the first data includes the position data and posture data of the target person. The step of obtaining the first data includes: determining the position data of the target person based on the position of the object data in the second image data; determining multiple key points of the target person and determining the posture data of the target person based on the key points; and determining the first data based on the position data and posture data of the target person. After receiving the position and posture, the analyzer can use previously stored data to restore the data, such as using previously stored person images to restore the data. Specifically, as an optional embodiment, the restoration of the first data based on person identification data to obtain the first intermediate data includes: obtaining person-related data determined based on the first image data based on the person identification data; adjusting the person-related data based on the posture data of the target person to form the first intermediate data; and fusing the first intermediate data and the second intermediate data based on the position data of the target person. The person-related data includes pixel data and morphological data corresponding to the person in the first image data, etc. Morphological data may include, but is not limited to, a person’s height, weight, chest circumference, limb condition, etc., but may include any morphological data known to those skilled in the art.

[0027] In this invention, feature fusion includes bit-by-bit addition, mathematically expressed as: Existing feature vector To fuse these two feature vectors, we directly add their corresponding elements. This operation assumes that the two vectors have the same dimension. If they have different dimensions, a linear transformation can be used to convert them into vectors of the same dimension. .

[0028] This scheme's extraction model can extract human features and form human identification data. Therefore, this scheme can adopt an unsupervised approach for recognition. For example, this scheme can acquire multiple images of the same person taken by different cameras, extract features from each other, and perform identification and matching, thereby achieving an unsupervised learning process. Specifically, as an optional embodiment, the training data includes a first training image and a second training image from different cameras, both of which contain the same human figure. The training steps of the extraction model include: inputting the first training image into the extraction model to determine the first extraction data; inputting the second training image into the extraction model to determine the second extraction data; and adjusting the extraction model based on the human recognition results in the first extraction data and the second training image, the second extraction data, and the human recognition results in the first training image.

[0029] Furthermore, this solution can also identify the number of people in the images captured by the camera. The fewer the number of people, the fewer the human features can be included in the human identification data, thereby reducing the amount of interactive data. Specifically, as an optional embodiment, the human identification data includes human feature information; the step of analyzing the human data and a pre-trained extraction model to determine the human identification data includes: analyzing the human data and the pre-trained extraction model to determine human feature information; determining the number of people in the second camera, and filtering the human feature information based on the number of people to form human identification data, wherein the more people, the more human feature information is included in the human identification data. The information includes: person feature information (person size information), person identification data (person size data adjusted based on the relative positions of the first and second cameras, determined after adjustment based on the person size information); person feature information also includes person color data (person color data including pixel values ​​of pixels in multiple areas of the person); person identification data includes level data that divides pixel values ​​into corresponding levels, with pixel values ​​from 0 to 255 divided into eight levels); the second camera uploads second image data in the first time period and compressed data in the second time period after the first time period, to analyze the number of people and the differences in person feature information in the second camera based on the second image data, in order to determine the person identification data corresponding to each person; the second camera periodically uploads second image data, and the data after the second image data upload can be compressed based on the second image data to form compressed data for upload.

[0030] In this invention, feature selection can achieve the following technical effects: improve model performance: reduce noise and redundant features, prevent overfitting, and enhance the model's generalization ability; accelerate the training process: reduce the number of features, directly reduce computational complexity and training time; and improve model interpretability: using fewer features makes the model's decision-making process clearer and easier to understand.

[0031] In this invention, feature selection employs an embedding method: a method where the user decides which features to use, meaning feature selection and algorithm training occur simultaneously. The embedding method includes: training with machine learning algorithms and models to obtain weight coefficients for each feature; and selecting features based on these weight coefficients, from largest to smallest. Weight coefficients often represent a feature's contribution to the model or its importance. For example, the `feature_importances_` attribute in decision trees and ensemble tree models can list the contribution of each feature to tree construction. Based on this contribution evaluation, the most useful features for model construction can be identified. Therefore, compared to filtering methods, the embedding method yields results that are more precise in terms of model utility, resulting in better improvement in model effectiveness. Furthermore, because it considers the contribution of features to the model, irrelevant features (features requiring relevance filtering) and features without discriminative power (features requiring variance filtering) are removed due to their lack of contribution to the model, making it an evolution of the filtering method.

[0032] Based on the above embodiments, this application also provides a multi-camera linkage monitoring system, such as... Figure 2 As shown, the system includes: The first data acquisition module 202 is used to acquire first image data from the first camera, the first image data including the person data of the target person.

[0033] The character identifier acquisition module 204 is used to analyze character data and a pre-trained extraction model to determine character identifier data.

[0034] The compressed data acquisition module 206 is used to send person identification data to the second camera. The second camera acquires the second image data of the target person and compresses the data based on the person identification data to form compressed data.

[0035] The compressed data restoration module 208 is used to receive compressed data from the second camera and restore the data based on the person identification data to obtain restored data, which is then used for monitoring and analysis based on the first image data and the restored data.

[0036] The implementation methods of this application are similar to those of the above method embodiments. For specific implementation methods, please refer to the specific implementation methods of the above method embodiments, which will not be repeated here.

[0037] This application embodiment can be applied to security monitoring scenarios, utilizing multiple cameras for joint monitoring and analysis. Furthermore, this solution can compress images of the same person captured by multiple cameras, thereby reducing the amount of data that needs to be transmitted. Specifically, in this solution, the first camera captures data of the target person. Based on the target person's movement trajectory, the corresponding second camera is determined. This solution can analyze the person data and a pre-trained extraction model to determine person identification data; this identification data is then sent to the second camera. After the second camera identifies the target person based on the identification data, it can segment the captured second image data into object data containing the target person and background data, thereby performing corresponding data compression and transmitting the compressed data. Data restoration is then performed based on the person identification data to obtain restored data. After obtaining the data captured by the first camera and the restored data, monitoring and analysis can be performed based on the first image data and the restored data. This solution can compress images related to the person and the background of the person based on images of the same person captured by two cameras, thereby reducing the amount of data transmitted.

[0038] Based on the above embodiments, this application also provides an electronic device, including: a memory and at least one processor; the memory is used to store computer execution instructions; the at least one processor is used to execute the computer execution instructions stored in the memory, causing the at least one processor to perform the method as described in the above embodiments.

[0039] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements the various processes of the above-described data processing method embodiments and achieves the same technical effects. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0040] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0041] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0042] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0043] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0044] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0045] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0046] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0047] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0048] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0049] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A multi-camera linkage monitoring method, characterized in that The method includes: Acquire first image data from the first camera, the first image data including the person data of the target person; Based on the analysis of the person data and the pre-trained extraction model, the person identification data is determined; The person identification data is sent to the second camera, which then captures the second image data of the target person and compresses the data based on the person identification data to form compressed data. The system receives compressed data from the second camera and performs data reconstruction based on the person identification data to obtain the reconstructed data, which is then used for monitoring and analysis based on the first image data and the reconstructed data.

2. The method according to claim 1, characterized in that, The steps for the second camera to acquire compressed data include: Acquire the second image data, determine the target person based on the person identification data, and divide the second image data into object data containing the target person and background data; The object data is compressed based on the person identification data to obtain the first data; the background data is compressed a second time to obtain the second data. Compressed data is formed based on the first and second data.

3. The method according to claim 2, characterized in that, The data restoration based on the person identification data, to obtain the restored data, includes: Obtain the compressed data and extract the first and second data. The first intermediate data is obtained by restoring the first data based on the character identification data; The second intermediate data is obtained by restoring the second data based on historical background data; Based on the first and second intermediate data, the data to be restored is determined.

4. The method according to claim 3, characterized in that, The first data includes the target person's location data and the target person's posture data. The steps for obtaining the first data include: Based on the position of the object data in the second image data, determine the location data of the target person; Identify multiple key points of the target person and determine the target person's posture data based on these key points; Based on the target person's location and posture data, the first data is determined.

5. The method according to claim 4, characterized in that, The process of restoring the first data based on the person identification data to obtain the first intermediate data includes: Obtain relevant data about the person based on the person identification data, which is determined from the first image data; Based on the target person's posture data, the relevant data of the person are adjusted to form the first intermediate data. The first intermediate data and the second intermediate data are then fused based on the target person's position data.

6. The method according to claim 1, characterized in that, The training data includes first and second training images from different cameras, both containing the same person; the training steps of the extraction model include: The first training image is input into the extraction model to determine the first extraction data; The second training image is input into the extraction model to determine the second extraction data; The extraction model is adjusted based on the character recognition results in the first extracted data and the second training image, as well as the character recognition results in the second extracted data and the first training image.

7. The method according to claim 1, characterized in that, The person identification data includes person feature information; the analysis based on the person data and a pre-trained extraction model to determine the person identification data includes: Based on the analysis of the person's data and the pre-trained extraction model, the characteristic information of the person is determined; The number of people in the second camera is determined, and based on the number of people, the feature information of the people is filtered to form person identification data. The more people there are, the more feature information the person identification data contains. The information includes: person feature information (person size information, person identification data including size data adjusted according to the relative positions of the first and second cameras); person feature information includes: person color data, which includes pixel values ​​of pixels in multiple areas of the person; person identification data includes: level data that divides pixel values ​​into corresponding levels, with pixel values ​​from 0 to 255 divided into eight levels; the second camera uploads second image data in the first time period and compressed data in the second time period after the first time period, in order to analyze the number of people and the differences in person feature information in the second camera based on the second image data, so as to determine the person identification data corresponding to each person; the second camera periodically uploads second image data.

8. A multi-camera linkage monitoring system, characterized in that, The system includes: The first data acquisition module is used to acquire first image data from the first camera, wherein the first image data includes the person data of the target person; The character identification acquisition module is used to analyze character data and a pre-trained extraction model to determine character identification data. The compressed data acquisition module is used to send person identification data to the second camera. The second camera acquires the second image data of the target person and compresses the data based on the person identification data to form compressed data. The compressed data restoration module is used to receive compressed data from the second camera and restore the data based on the person identification data to obtain restored data, which is then used for monitoring and analysis based on the first image data and the restored data.

9. An electronic device, characterized in that, include: Memory and at least one processor; The memory is used to store computer-executed instructions; The at least one processor is configured to execute computer execution instructions stored in the memory, such that the at least one processor performs the method as described in any one of claims 1-7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-7.

Images