Monitoring method, device, server and computer-readable storage medium

Abstract

The present application discloses a monitoring method, device, system, a server and a computer-readable storage medium. The relevant embodiments can be applied to various scenarios such as community security, artificial intelligence, and cloud technology. The method includes: when the server starts the current monitoring task, obtaining the feature with the highest similarity to the facial feature of the pedestrian image to be compared, and if the feature with the highest similarity meets the monitoring rules, then triggering the pedestrian monitoring alarm notification, extracting the captured vehicle information from the vehicle image to be compared, at least obtaining the real-time captured vehicle information, performing vehicle attribute recognition and vehicle trajectory analysis on the real-time captured vehicle information, and searching for vehicles by time range, location, and vehicle attributes, combining the captured vehicle information and the algorithm recognition results to perform monitoring and control, and if the vehicle attribute recognition result, vehicle trajectory analysis result, vehicle search result and monitoring and control monitoring result meet the monitoring rules, then triggering the vehicle monitoring alarm notification.

Description

[0001] This application is a divisional application of the Chinese patent application filed with the China Patent Office on March 20, 2020, with application number 2020102041251 and invention name “Monitoring method, device, server and computer-readable storage medium”. Technical Field

[0002] The present application relates to the field of artificial intelligence technology, and more specifically, to a monitoring method, device, server, and computer-readable storage medium. Background Art

[0003] With the development of artificial intelligence (AI) technology and the growing demand for security from all walks of life, traditional, simple, passive security methods are no longer sufficient for the diverse daily lives and work environments. Driven by technologies like big data and AI, security is evolving towards urbanization, integration, and proactive security, with intelligent security becoming the mainstream trend. Residential communities, with their high concentrations of people and complex demographics, are among the most vulnerable to enhanced security. Over time, many residential access control systems have experienced frequent malfunctions, and video surveillance equipment remains stuck in the traditional monitoring phase, mostly limited to post-event evidence collection. It lacks the ability to prevent or warn of suspicious or unusual behavior, and generally remains in a passive state, capable of monitoring but not controlling. This makes it difficult to meet actual security needs.

[0004] Therefore, how to improve the real-time performance of intelligent monitoring is a technical problem that those skilled in the art need to solve. Summary of the Invention

[0005] The purpose of this application is to provide a monitoring method, device, server and computer-readable storage medium to improve the real-time performance of intelligent monitoring.

[0006] To achieve the above objectives, the present application provides a first aspect of a monitoring method, comprising:

[0007] Receive the control rules of this control task and a list of cameras that need to be controlled; wherein the camera list includes multiple target cameras, and the target cameras are used to capture images of the target object;

[0008] The control task is started, and if the image to be compared of the target object meets the control rule, a control alarm notification is triggered;

[0009] The process of acquiring the image to be compared includes: acquiring trajectory data of the target object using the target camera; wherein the trajectory data includes multiple target images containing the target object; and determining the image to be compared in the trajectory data based on the quality score of the target object in the target image.

[0010] To achieve the above-mentioned purpose, the second aspect of the present application provides a monitoring device, comprising:

[0011] A determination module is configured to receive the control rules for this control task and a list of cameras to be controlled; wherein the camera list includes a plurality of target cameras, and the target cameras are configured to capture images of the target object;

[0012] a module for acquiring an image to be compared, configured to acquire trajectory data of the target object using the target camera; wherein the trajectory data includes a plurality of target images containing the target object; and determining an image to be compared in the trajectory data based on a quality score of the target object in the target image;

[0013] The trigger module is used to start the current control task and trigger a control alarm notification if the image to be compared of the target object meets the control rule.

[0014] To achieve the above objectives, the third aspect of the present application provides a server, including:

[0015] memory for storing computer programs;

[0016] A processor is used to implement the steps of the above-mentioned monitoring method when executing the computer program.

[0017] To achieve the above-mentioned purpose, the fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned monitoring method are implemented.

[0018] From the above scheme, it can be seen that a monitoring method provided by the present application includes: receiving the control rules of this control task and a list of cameras that need to be controlled; wherein, the camera list includes multiple target cameras, and the target cameras are used to capture images of the target object; starting the current control task, if the image to be compared of the target object meets the control rules, triggering a control alarm notification; wherein, the acquisition process of the image to be compared includes: using the target camera to obtain the trajectory data of the target object; wherein, the trajectory data includes multiple target images containing the target object; determining the image to be compared in the trajectory data based on the quality score of the target object in the target image.

[0019] The monitoring method provided by the present application realizes real-time capture and control of the target object. The user can deploy a list of cameras that need to be controlled and set control rules to meet flexible control needs. At the same time, the image to be compared is determined in the trajectory data based on the quality score of the target object in each target image, thereby ensuring the accuracy of the control. In addition, the present application does not limit the source of the trajectory data, and can reuse the deployed traditional security cameras or stream machines without the need for additional deployment and maintenance costs. Since the source of the video stream is not restricted, it can support both sensed access control scenarios and senseless monitoring scenarios, and meet the needs of intelligent security in an all-round and integrated manner. The present application also discloses a monitoring device, a server, and a computer-readable storage medium, which can also achieve the above-mentioned technical effects.

[0020] It should be understood that the foregoing general description and the following detailed description are merely illustrative and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. The drawings are used to provide a further understanding of the present disclosure and constitute part of the specification. Together with the following specific embodiments, they are used to explain the present disclosure, but do not constitute a limitation of the present disclosure. In the drawings:

[0022] Figure 1 An architectural diagram of a monitoring system provided in an embodiment of the present application;

[0023] Figure 2 A flow chart of a monitoring method provided in an embodiment of the present application;

[0024] Figure 3 A flowchart of a base database registration provided in an embodiment of the present application;

[0025] Figure 4 A flowchart of another monitoring method provided in an embodiment of the present application;

[0026] Figure 5 A flow chart of another monitoring method provided in an embodiment of the present application;

[0027] Figure 6 This is a schematic diagram of the interaction between the intelligent monitoring all-in-one machine and the external modules;

[0028] Figure 7 This is a structural diagram of the internal structure of the intelligent monitoring all-in-one machine;

[0029] Figure 8 A schematic diagram of a face function interaction interface;

[0030] Figure 9 A schematic diagram of a vehicle function interaction interface;

[0031] Figure 10 A schematic diagram of a vehicle function interaction interface;

[0032] Figure 11 A structural diagram of a monitoring device provided in an embodiment of the present application;

[0033] Figure 12 A structural diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0035] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0036] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0037] The facial recognition kiosk is a hardware-software integrated product that embeds facial detection and recognition algorithms into intelligent hardware. It features a built-in image acquisition camera, data transmission port, facial recognition chip, fingerprint acquisition chip, and interactive display. It is primarily used for 1:1 identity verification, facial attendance, and facial access control. Suitable applications include self-service bank account opening, bus ID verification, and areas that rely on gates, such as residential units, office buildings, construction sites, and hotels.

[0038] The inventors of this application have discovered through research that face recognition cameras in related technologies are incompatible with traditional security cameras, requiring additional deployment and maintenance costs. Furthermore, face recognition cameras employ a sensory monitoring method, requiring active user cooperation in most cases, resulting in significant limitations in their application scenarios. Furthermore, the data volume supported by face recognition cameras is mostly at the cell level, making it difficult to scale up, and limiting their scalability. This suggests that face recognition cameras in related technologies suffer from low monitoring efficiency and are limited in their application scenarios and scalability.

[0039] Therefore, this application utilizes facial recognition technology to provide a clustered, all-in-one AI security monitoring system for community scenarios. Its main functions include proactive real-time capture and display of objects in the monitored video, timely deployment and alarming of controlled objects, and dynamic trajectory retrieval. In this application, already deployed traditional security cameras or streamers can be reused. That is, the input video stream of the monitoring system can be video captured by traditional security cameras or video streams from streamers, without requiring additional deployment and maintenance costs. Because the monitoring system adopts a clustered design, horizontal expansion can be achieved by simply adding new service nodes, which facilitates the expansion of the monitoring data volume and can support security needs at the district, municipal administrative district, and even city levels. Furthermore, because the source of the video stream is unrestricted, it can support both active access control scenarios and passive monitoring scenarios, fully integrating to meet the needs of intelligent security. Thus, this application can provide end users such as property management and buildings with business capabilities such as facial control retrieval and vehicle recognition, directly providing users with AI all-in-one solutions for building and security scenarios.

[0040] In order to facilitate understanding of the monitoring method provided by this application, the system used is introduced below. Figure 1 , which shows an architecture diagram of a monitoring system provided by an embodiment of the present application, such as Figure 1 As shown, the system includes a target camera 100, a server 200 and a client 300. The target camera 100 and the server 200, and the server 200 and the client 300 are connected via a network.

[0041] In practice, users can deploy a list of cameras to be monitored, specifically target cameras 100. These cameras capture image and video streams and can include cameras and stream generators for various application scenarios, not specifically limited here, such as face capture cameras. Video streams require decoding using a video decoding service, allowing server 200 to identify the target objects. Expanding monitoring points requires simply connecting the target cameras at the newly added monitoring points to the system.

[0042] The server 200 is used to process the image or video input by the target camera 100 and identify the target object therein. The server 200 can provide users with services such as control and retrieval. It is understandable that in order to improve business processing capabilities, the server 200 can adopt a clustered design and use LB (LoadBalance) technology to share tasks such as network services and network traffic to multiple network node devices or multiple links in the cluster, thereby ensuring high reliability of the business. In this embodiment, each server node can be responsible for the monitoring task of a certain area, that is, it is responsible for the images collected by the target camera 100 in a certain camera list. When expanding the entire monitoring system, it is only necessary to set up corresponding server nodes for the newly added areas, which is conducive to the expansion of the monitoring data level.

[0043] The client 300 can be a mobile terminal such as a mobile phone or a fixed terminal such as a PC (Chinese full name: personal computer, English full name: personal computer), which can display the captured images collected by the target camera 100 in real time. Users can also deploy control rules, upload registered images to the control image library, enter search items, etc. through the client 300.

[0044] The embodiment of the present application discloses a monitoring method, which improves the real-time performance of intelligent monitoring.

[0045] See also Figure 2 , a flow chart of a monitoring method provided in an embodiment of the present application, such as Figure 2 As shown, including:

[0046] S101: The client sends the control rules of this control task and a list of cameras to be controlled to the server; wherein the camera list includes multiple target cameras, and the target cameras are used to capture images of the target object;

[0047] In practice, users can configure the monitoring rules and camera list for this monitoring task through the client's interactive interface. A monitoring rule might specify that the target object in the captured image must be an object in the monitoring image library. Users can create different monitoring image libraries for different types of objects and select the corresponding library for this monitoring task.

[0048] If the target object is specifically a vehicle, the control rules here can be determined based on the control items, that is, the vehicle in the captured image meets the pre-set control items. The control items here may include the capture location, capture time, vehicle information, etc.

[0049] S102: The server starts the current control task;

[0050] S103: The target camera collects trajectory data of the target object; wherein the trajectory data includes multiple target images containing the target object;

[0051] In this step, the target camera captures a target image containing the target object, and multiple target images constitute the trajectory data of the target object. The target object here can include a portrait, a vehicle, etc., which are not specifically limited here.

[0052] For portraits, a face detection algorithm can be used to determine the face frame containing the face. The face detection algorithm takes an image as input and outputs a sequence of face frame coordinates. Typically, the output face frame is an upright square, but it can also be an upright rectangle or a rotated rectangle.

[0053] For vehicles, this embodiment does not limit the specific recognition algorithm. Multiple features of pixel points in the target image can be extracted through a convolutional neural network. All pixel points in the target image are classified based on the extracted features and divided into vehicle areas and background areas, and finally the area corresponding to the target object is obtained.

[0054] S104: The target camera sends the trajectory data to the server;

[0055] S105: The server determines an image to be compared in the trajectory data based on the quality score of the target object in the target image;

[0056] In this step, the server performs a quality assessment on the target objects in each target image and determines the image to be compared based on the quality score of the target objects in each target image. In order to ensure the accuracy of the control, the quality score of the target objects in the image to be compared needs to be greater than the set first preset value. If there is no target image with a quality score greater than or equal to the first preset value in the trajectory data, the target image with the highest quality score is determined as the image to be compared.

[0057] It should be noted that this embodiment does not limit the specific calculation method of the quality score. For example, it can be calculated based on the conditional parameter value of the target object and a preset quality determination function. The conditional parameter value is the value of the conditional parameter, and the conditional parameter may include the blurriness of the face image, the three-dimensional deflection angle of the face, the brightness of the face image, or the area of the face image. Among them, the three-dimensional deflection angle of the face may refer to the pitch angle (pitch), yaw angle (yaw) and roll angle (roll) of the face, and the conditional parameter value may include the blurriness value, the three-dimensional deflection angle value, the brightness value or the area value, etc. The quality determination function is determined based on the face recognition accuracy distribution and / or confidence distribution corresponding to the conditional parameters of the face image. In a specific implementation, the face recognition accuracy distribution or confidence distribution corresponding to the conditional parameters can be fitted with a data curve to obtain a fitting function, and then the quality determination function is determined based on the fitting function. The recognition accuracy distribution corresponding to the conditional parameters may refer to the distribution of recognition accuracy with respect to blur, the distribution of recognition accuracy with respect to three-dimensional deflection angle, the distribution of recognition accuracy with respect to brightness, or the distribution of recognition accuracy with respect to the area of the target area, etc.; the confidence distribution corresponding to the conditional parameters may refer to the distribution of confidence with respect to blur, the distribution of confidence with respect to three-dimensional deflection angle, the distribution of confidence with respect to brightness, or the distribution of confidence with respect to the area of the target area, etc.

[0058] In a specific implementation, the quality scores of the target objects in all target images in the trajectory data can be calculated, and the target image with the highest quality score can be selected as the image to be compared. Of course, to improve efficiency, the quality scores of the target objects in each target image can be calculated accordingly. As long as there is a target image with a quality score greater than a set first preset value, it will be used as the image to be compared, without calculating subsequent target images. In other words, this step may include: determining the current target image in the trajectory data, calculating the quality score of the target object in the current target image; if the quality score is greater than or equal to the first preset value, determining the current target image as the image to be compared; if the quality score is less than the first preset value, re-entering the step of determining the current target image in the trajectory data; if there is no target image in the trajectory data with a quality score greater than or equal to the first preset value, determining the target image with the highest quality score as the image to be compared. It is understood that the first preset value is not limited here, and the user can make timely adjustments based on actual monitoring conditions, such as monitoring accuracy.

[0059] S106: If the image to be compared of the target object meets the control rule, the server triggers a control alarm notification to the client;

[0060] In this step, the server determines whether the image to be compared meets the control rules during the execution of the control task. If so, it triggers a control alarm notification to the client.

[0061] Exemplarily, this step may include: comparing the image to be compared of the target object with the images in the control image library, and if the comparison result meets the preset conditions, triggering a control alarm notification. In a specific implementation, if the target object is a portrait, the face feature extraction algorithm can convert a face image into a string of fixed-length values. This string of values is called a face feature, which has the ability to characterize the characteristics of the face. The face feature extraction algorithm will align the face to a predetermined pattern based on the coordinates of the key points of the facial features, and then calculate the features. The Face Recognition algorithm is used to identify the identity corresponding to the input face image. Its input is a face feature, which is compared one by one with the features corresponding to N images in the control image library to find the feature with the highest similarity to the input feature. This highest similarity value is compared with a preset threshold. If it is greater than the threshold, the image corresponding to the feature is returned. The user can set the threshold for this control task in the control rules.

[0062] If the target object is a vehicle, fuzzy license plate number recognition and control are supported, and the user can set the license plate number recognition threshold. For example, for an image input from the control image library, if the license plate number recognition threshold is set to 90%, the system will recognize the license plate number "Liao A2438E" with a 92.5% probability and the license plate number "Liao A2488E" with a 95.5% probability. In this case, vehicles with the license plate numbers "Liao A2438E" and "Liao A2488E" will be controlled simultaneously.

[0063] It is understood that the deployment control image library includes multiple images, and users can upload images to the deployment control image library. That is, this embodiment may also include: obtaining a registered image and calculating a quality score of the registered image; if the quality score of the registered image is greater than or equal to a first preset value, calculating the similarity between the registered image and all images in the deployment control image library; and when all the similarities are less than a second preset value, saving the registered image to the deployment control image library.

[0064] In specific implementations, after receiving the registration image, the server needs to judge the image quality. This can filter out images of too low quality and prevent interference with the deployment control. For images that pass the quality score screening, similarity calculations must be performed with the features of existing images in the deployment control image library to filter out images with too high similarity and prevent duplicate registration. By controlling the registration process, the high quality of the deployment control image library is guaranteed, which greatly helps improve the accuracy of deployment control. At the same time, the above-mentioned first preset value and similarity filter value can be dynamically adjusted to meet the business needs of recall-focused scenarios.

[0065] This embodiment can be applied to the deployment and alarm of smart security. It uses facial recognition technology to directly capture facial photos from the monitoring screen, analyze facial features in real time, quickly complete the comparison and recognition of photos with facial blacklists or whitelists, calculate the similarity between the current face and the facial template in the face database, and can issue alarm prompts to realize an intelligent, socialized, and large-scale security system. On the technical level, the deployment and alarm includes two main processes: face registration and dynamic retrieval. Registered face photos include collected daily photos, ID photos, or target pictures extracted from captured images, and the retrieval request consists of faces captured in real-time monitoring. Face deployment and alarm tasks support the setting of users, communities, and cameras, and also support the setting of task validity periods.

[0066] like Figure 3 As shown, after receiving a registration image, the quality of the registered image is assessed using the quality score interface in the facial feature microservice. If the quality score falls below a threshold, registration is terminated. If the quality score is greater than the threshold, the feature interface is called to extract the image features of the registered image. The retrieval interface in the face retrieval microservice is then used to perform a 1:N feature search among the registered images. This involves calculating the similarity between the image features of the registered image and those of N registered images. If the maximum similarity (top 1) falls below the threshold, the registration interface is called to proceed; otherwise, registration is terminated. By controlling the registration process, the high quality of the registered images is ensured, significantly improving the accuracy of surveillance deployment.

[0067] To improve the accuracy and recall of face detection tasks, two retrieval algorithm processes were designed:

[0068] A. The hit status S of each track is initialized to false. For each received image, if S is false and the image quality score is greater than threshold T1, the search interface is called. If the top-1 hit similarity is greater than T2, the hit status S is set to true and the search result is returned. By controlling the quality score of the retrieved image, the accuracy of the face detection task is guaranteed.

[0069] B. If the trajectory ends and the highest quality score Q of the images in the trajectory is less than T1, a search is performed using the image with the highest quality score Q in the trajectory. If the Top-1 score is higher than T2, the hit status S is set to true and the search result is returned. If the quality scores of the images in the trajectory are generally low, the image with the highest quality score is used as the search image to ensure the recall rate of the face detection task.

[0070] The monitoring method provided in the embodiment of the present application realizes real-time capture and control of the target object. The user can deploy a list of cameras that need to be controlled and set control rules to meet flexible control needs. At the same time, the image to be compared is determined in the trajectory data based on the quality score of the target object in each target image, thereby ensuring the accuracy of the control. In addition, the embodiment of the present application does not limit the source of the trajectory data, and can reuse the deployed traditional security cameras or stream machines without the need for additional deployment and maintenance costs. Since the source of the video stream is not restricted, it can support both sensed access control scenarios and senseless monitoring scenarios, and meet the needs of intelligent security in an all-round and integrated manner.

[0071] This embodiment will introduce the real-time snapshot function of the monitoring system in detail, specifically:

[0072] See also Figure 4 , Figure 4 A flow chart of another monitoring method provided in an embodiment of the present application is shown as follows: Figure 4 As shown, including:

[0073] S201: Acquire trajectory data of a target object; wherein the trajectory data includes a plurality of target images, each of which is an image obtained after image acquisition of the target object;

[0074] The execution entity of this embodiment is Figure 1 The server 200 is used to capture the target object in real time.

[0075] S202: Determine the target image with the highest quality score in the trajectory data as the image to be extracted, and obtain snapshot information of the image to be extracted; wherein the snapshot information at least includes a snapshot location and a snapshot time;

[0076] In this step, the target image with the highest quality score is selected as the image to be extracted. The captured information of the target object is then extracted from the image to be extracted. This information can include the capture location and time. If the target object in the image to be extracted is a vehicle, the captured information can also include vehicle information. This vehicle information includes the license plate thumbnail, license plate number, vehicle body color, vehicle type, vehicle brand, and vehicle systems. The license plate thumbnail specifically refers to the area in the image to be extracted that corresponds to the license plate number.

[0077] In a specific implementation, a convolutional neural network can be used to extract features from the vehicle area, dividing the vehicle area into a portion with a license plate and a portion without a license plate, ultimately generating a license plate thumbnail. Another deep convolutional neural network is then selected to process the features of the vehicle area. This convolutional neural network extracts license plate type features by comprehensively extracting relevant information such as the color, shape, pattern, and text distribution within the vehicle, obtaining multiple features of the pixels within the determination area. Based on these extracted features, the license plate number, vehicle body color, vehicle type, brand, and vehicle system can be identified.

[0078] Because this step extracts the target object's snapshot information, any snapshot information can be used for control during the deployment. This means that control items are pre-set, and an alarm is triggered when the snapshot information meets the control items. For example, if the control item is red vehicles, during the real-time capture process, if the vehicle body color in the snapshot information is red, an alarm is triggered.

[0079] S203: extracting the area corresponding to the target object from the image to be extracted as a captured image, and saving the captured image and the captured information corresponding to each captured image into a captured database.

[0080] In this step, the area corresponding to the target object in the image to be extracted is used as a captured image and saved, along with the captured information extracted in the previous step, to a captured image database, thus implementing a real-time capture function. As a preferred implementation, this embodiment further includes displaying all captured images in the captured image database and the captured information corresponding to each captured image. In a specific implementation, the captured images and captured information can be displayed in real time on the client, facilitating manual monitoring.

[0081] It is understandable that the snapshot database in this embodiment and the control image library in the previous embodiment support basic management operations, and users can manage them through the client's interactive interface, such as adding images, deleting images, refreshing the database, classifying images, uploading or modifying image information, etc. That is, this embodiment also includes: when an operation command is received, determining the operation object corresponding to the operation command; wherein the operation object includes the control image library or the snapshot database; performing the management operation corresponding to the operation command on the operation object; wherein the management operation includes any one of adding an image to the operation object, deleting an image from the operation object, refreshing the operation object, classifying all images in the operation object, and uploading or modifying image information of images in the operation object.

[0082] It can be seen that this embodiment realizes real-time capture and display of the target object, and establishes a capture database for the captured images and captured information, which is beneficial to subsequent retrieval and manual monitoring.

[0083] Based on the above embodiment, this embodiment will introduce the retrieval function of the monitoring system in detail, specifically:

[0084] See also Figure 5 , Figure 5 A flow chart of another monitoring method provided in an embodiment of the present application is as follows: Figure 5 As shown, including:

[0085] S301: When a search command is received, determining a search item corresponding to the search command; wherein the search item includes any one or a combination of the image to be searched and the captured information;

[0086] The execution entity of this embodiment is Figure 1 The server 200 in the embodiment of the present invention is used to search the snapshot database. In this step, the user can set the search items through the interactive interface of the client, which may include images to be searched, snapshot information, etc.

[0087] S302: Determine the search result corresponding to the search item in the snapshot database.

[0088] The search results of this step are the captured images in the snapshot database that meet the search terms and their corresponding snapshot information. In a specific implementation, if the search terms include the image to be retrieved, this step may include: determining the object to be retrieved in the image to be retrieved; comparing the area corresponding to the object to be retrieved in the image to be retrieved with the image in the snapshot database to obtain the search results. For example, a user can use a certain face photo to search for a person in the snapshot database. The server compares the facial features of the input image to be retrieved with the facial features in the snapshot database, and returns the top-N captured images with a similarity greater than a threshold and their corresponding snapshot information and similarity information.

[0089] If the search item is specifically the license plate number of the target vehicle, this embodiment further includes: displaying the target vehicle's driving trajectory on a map based on the captured location in the search results. In a specific implementation, because the search results include the captured information of each captured image, the vehicle trajectory can be presented in the form of a map based on the captured location in the captured information, allowing intuitive viewing of the vehicle's passing location information.

[0090] It can be seen that this embodiment realizes the retrieval function of the monitoring system based on the snapshot database. Users can deploy retrieval rules by setting retrieval items to support the image search business function.

[0091] For ease of understanding, this application is introduced in conjunction with an application scenario. The intelligent monitoring all-in-one machine provides intelligent security services for the community, realizing face monitoring and vehicle monitoring functions. Figure 6 and Figure 7 , Figure 6 This is a diagram of the interaction between the intelligent monitoring all-in-one machine and external modules. Figure 7 This is a structural diagram of the internal structure of the intelligent monitoring all-in-one machine.

[0092] The client-side system is used for real-time face and vehicle capture, supporting both image and video streams. For image streams, the intelligent capture device leverages its built-in face and vehicle algorithms to detect and report images containing faces and vehicles from real-time surveillance video. For video streams, the video decoding service deployed on the appliance node decodes the input video stream and performs face detection and monitoring optimization according to the face algorithm process. Both systems are fully aligned in terms of data flow and algorithm logic, and the algorithm results are output to the same application layer module, greatly improving system consistency and maintainability.

[0093] The appliance access layer provides request forwarding, converges appliance ports, and unifies external services. The face service, an application layer related to facial services, serves as the logical control core of the entire appliance node. It receives facial calculation results pushed by the end-side. Its main functions include facial data collection, trajectory capture and display, real-time facial surveillance and alarming, and dynamic registration and retrieval of pedestrian databases. It also provides surveillance task management, pedestrian and facial database retrieval, and a facial database proxy interface. The face service interacts with the face SDK service, person retrieval service, and the TDSQL storage layer. The face SDK service accepts facial image input and provides face detection, facial key point location, face quality score calculation, and facial feature extraction interfaces. Based on the input facial image, facial features, the face database ID to be retrieved, and a similarity threshold, the face retrieval service returns the top-N face IDs in the face database with a similarity greater than the threshold, along with the corresponding similarity scores.

[0094] The Vehicle Capture Service receives a stream of vehicle images pushed from the client and calculates vehicle attributes. It interacts with the Vehicle Attribute Service and reports the calculation results. The Vehicle Attribute Service receives vehicle images as input and provides interfaces for vehicle detection, vehicle attribute calculation, and license plate recognition.

[0095] The heartbeat reporting service reports heartbeats to the master central node, notifying the local node of its liveness and load capacity. It also pulls the assigned capture / streaming device from the central node. When the all-in-one appliance is deployed in a cluster, only a single interface can be exposed to the user, hiding the internal logic between the appliance nodes. The Prometheus module is responsible for subscribing to monitoring metadata for the access, compute, and storage layers. The Grafana module, based on the configured dashboard and specified rules, obtains metadata from Prometheus for calculation, displays it on the front end, and pushes alerts based on the corresponding alert rules.

[0096] The whole system has good scalability. Figure 6 The heartbeat reporting service in the cluster reports the liveness and load of each integrated node to the cluster's master module. Based on this information, the master automatically assigns the corresponding node to each connected camera or video stream, thus hiding the internal details of the cluster from the user. Theoretically, this system supports unlimited horizontal scalability, enabling regional and even city-level security operations.

[0097] The facial function mainly includes the intelligent capture machine and the facial AI module, and is combined with the web backend and front end to provide real-time capture reporting, facial control, capture library retrieval and other functions. Figure 8 The interactive interface for the face function.

[0098] The all-in-one machine evaluates the quality of the facial images captured in real time, selects the facial image with the highest facial quality score in the trajectory for storage, and outputs the coordinate position and image of the face in the overall picture, thereby realizing the real-time facial capture function.

[0099] Users can create and manage databases for facial images of different categories, supporting general management operations such as creating, deleting, and updating databases, as well as batch and single facial image storage. Users can also import information from key population databases into the photo library management platform, upload key personnel information including name, address, and facial photos, perform face detection and facial feature extraction on portrait photos, and ultimately store relevant features in a unified database.

[0100] The user selects the facial database to be monitored, chooses the camera list to be monitored, enters the monitoring task name, and sets the monitoring comparison threshold. By starting the monitoring task, the captured image is compared with the facial images in the selected monitoring database. When the similarity exceeds the monitoring comparison threshold, a monitoring alarm notification is issued, thus implementing the facial monitoring function.

[0101] Users can use a face photo to search for people by image in the snapshot database. After entering the image to be compared and setting the comparison similarity threshold, the system will return the snapshot image and similarity information whose comparison results exceed the threshold, thereby realizing the image search business function through the AI all-in-one machine.

[0102] Vehicle functions include real-time vehicle capture, vehicle attribute recognition, license plate recognition, vehicle control management, vehicle retrieval, and vehicle trajectory analysis. Figure 9 It is the interactive interface for vehicle functions.

[0103] The all-in-one machine analyzes the images uploaded by the vehicle capture machine through algorithms, forms structured information and presents it. Compared with simple captured images, it is easier to browse information and can include license plate thumbnail, license plate number, body color, vehicle type, vehicle brand, vehicle series, checkpoint name, checkpoint passing time, etc. The license plate thumbnail shows the captured license plate photo, making it easy to check the license plate number. Users can click to view details to view the captured panoramic view, such as Figure 10 As shown, the captured panorama shows a large panoramic view of the vehicle passing by, making it easy to view the entire vehicle photo. It can automatically identify vehicle color, vehicle brand, vehicle series, and vehicle type, and can search by time range, location, and vehicle attributes.

[0104] Users can also input license plate images to identify vehicle numbers (including province, city, and number), and provide fuzzy license plate search capabilities. It also provides the ability to combine captured data with GIS (Geographic Information System) to present vehicle trajectories on a map, allowing users to intuitively view vehicle locations.

[0105] Users can also set vehicle control rules, creating control tasks based on time intervals, license plate numbers, and locations. Once a control task is activated, the system combines captured data with algorithm recognition results to monitor the control. If a control rule is met, an alarm will be generated immediately.

[0106] A monitoring device provided in an embodiment of the present application is introduced below. The monitoring device described below and the monitoring method described above can be referenced to each other.

[0107] See also Figure 11 , a structural diagram of a monitoring device provided in an embodiment of the present application, such as Figure 11 As shown, including:

[0108] The determination module 201 is configured to receive the control rules of the current control task and a list of cameras to be controlled; wherein the list of cameras includes a plurality of target cameras, and the target cameras are configured to capture images of the target object;

[0109] The image acquisition module 202 is configured to acquire trajectory data of the target object using the target camera; wherein the trajectory data includes a plurality of target images containing the target object; and determine the image to be compared in the trajectory data based on the quality score of the target object in the target image;

[0110] The trigger module 203 is used to start the current monitoring task and trigger a monitoring alarm notification if the to-be-compared image of the target object meets the monitoring rule.

[0111] The monitoring device provided in the embodiment of the present application realizes real-time capture and control of the target object. The user can deploy a list of cameras that need to be controlled and set control rules to meet flexible control needs. At the same time, the image to be compared is determined in the trajectory data based on the quality score of the target object in each target image, thereby ensuring the accuracy of the control. In addition, the embodiment of the present application does not limit the source of the trajectory data, and can reuse the deployed traditional security cameras or stream machines without the need for additional deployment and maintenance costs. Since the source of the video stream is not restricted, it can support both sensed access control scenarios and senseless monitoring scenarios, and meet the needs of intelligent security in an all-round and integrated manner.

[0112] On the basis of the above embodiment, as a preferred implementation mode, the trigger module 203 specifically starts the current control task, compares the image to be compared of the target object with the image in the control image library, and if the comparison result meets the preset conditions, triggers the control alarm notification module.

[0113] Based on the above embodiment, as a preferred implementation, it further includes:

[0114] A first calculation module is used to obtain a registered image and calculate a quality score of the registered image;

[0115] a second calculation module, configured to calculate the similarity between the registered image and all images in the control image library if the quality score of the registered image is greater than or equal to a first preset value;

[0116] A saving module is used to save the registered image to the control image library when all the similarities are less than a second preset value.

[0117] Based on the above embodiment, as a preferred implementation manner, the first determining module includes:

[0118] a calculation unit, configured to determine a current target image in the trajectory data and calculate a quality score of the target object in the current target image;

[0119] a first determining unit, configured to determine the current target image as an image to be compared if the quality score is greater than or equal to a first preset value;

[0120] a second determining unit, configured to re-enter the step of determining a current target image in the trajectory data if the quality score is less than the first preset value;

[0121] The third determining unit is configured to determine the target image with the highest quality score as the image to be compared if there is no target image with a quality score greater than or equal to the first preset value in the trajectory data.

[0122] Based on the above embodiment, as a preferred implementation, it further includes:

[0123] A second acquisition module is configured to determine the target image with the highest quality score in the trajectory data as the image to be extracted, and obtain snapshot information of the image to be extracted; wherein the snapshot information includes at least a snapshot location and a snapshot time;

[0124] The extraction module is used to extract the area corresponding to the target object in the image to be extracted as a captured image, and save the captured image and the captured information corresponding to each captured image into a captured database.

[0125] Based on the above embodiment, as a preferred implementation, it further includes:

[0126] The display module is used to display all captured images in the captured database and the captured information corresponding to each captured image.

[0127] Based on the above embodiment, as a preferred implementation manner, the second acquisition module includes:

[0128] a fourth determining unit, configured to determine a target image with the highest quality score in the trajectory data as an image to be extracted;

[0129] An acquisition unit is used to acquire the capture position, capture time and vehicle information of the image to be extracted if the target object in the image to be extracted includes a vehicle, so as to obtain the capture information of the image to be extracted; wherein, the vehicle information includes any one or a combination of several items including a license plate thumbnail, a license plate number, a vehicle body color, a vehicle type, a vehicle brand and a vehicle system, and the license plate thumbnail is specifically the area corresponding to the license plate number in the image to be extracted.

[0130] Based on the above embodiment, as a preferred implementation, the first execution module includes:

[0131] a fifth determining unit, configured to receive a control item of the current control task and determine the control rule based on the control item, wherein the control item includes any one or a combination of any several items of the captured information;

[0132] An execution unit is configured to execute the current control task by determining whether the image to be compared satisfies the control rule.

[0133] Based on the above embodiment, as a preferred implementation, it further includes:

[0134] A second determining module is configured to, upon receiving a search command, determine a search item corresponding to the search command; wherein the search item includes any one or a combination of the image to be searched and the captured information;

[0135] The third determining module is used to determine the search result corresponding to the search item in the snapshot database.

[0136] Based on the above embodiment, as a preferred implementation manner, the search item includes an image to be retrieved, and the third determination module includes:

[0137] a sixth determining unit, configured to determine an object to be retrieved in the image to be retrieved;

[0138] The comparison unit is used to compare the area corresponding to the object to be retrieved in the image to be retrieved with the image in the snapshot database to obtain a retrieval result.

[0139] Based on the above embodiment, as a preferred implementation, if the search item is specifically the license plate number of the target vehicle, the third determination module further includes:

[0140] A display unit is used to display the driving track of the target vehicle on a map based on the captured position in the retrieval result.

[0141] Based on the above embodiment, as a preferred implementation, it further includes:

[0142] a fourth determining module, configured to, upon receiving an operation command, determine an operation object corresponding to the operation command; wherein the operation object includes the control image library or the snapshot database;

[0143] The second execution module is used to perform the management operation corresponding to the operation command on the operation object; wherein the management operation includes any one of adding an image to the operation object, deleting an image from the operation object, refreshing the operation object, classifying all images in the operation object, uploading or modifying image information of the image in the operation object.

[0144] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0145] This application also provides a server, see Figure 12 , a structural diagram of a server 200 provided in an embodiment of the present application, such as Figure 12 As shown, it may include a processor 21 and a memory 22.

[0146] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 21 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0147] The memory 22 may include one or more computer-readable storage media, which may be non-transitory. The memory 22 may also include high-speed random access memory, and non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In this embodiment, the memory 22 is at least used to store the following computer program 221, wherein, after the computer program is loaded and executed by the processor 21, it can implement the relevant steps in the monitoring method performed by the server side disclosed in any of the aforementioned embodiments. In addition, the resources stored in the memory 22 may also include an operating system 222 and data 223, etc., and the storage method may be temporary storage or permanent storage. Among them, the operating system 222 may include Windows, Unix, Linux, etc.

[0148] In some embodiments, the server 200 may further include a display screen 23 , an input / output interface 24 , a communication interface 25 , a sensor 26 , a power supply 27 , and a communication bus 28 .

[0149] certainly, Figure 10 The server structure shown does not constitute a limitation on the server in the embodiment of the present application. In actual applications, the server may include Figure 10 More or fewer components than shown, or combinations of certain components.

[0150] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When the program instructions are executed by a processor, the steps of the monitoring method executed by the server in any of the above embodiments are implemented.

[0151] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method section. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to this application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

[0152] It should also be noted that, in this specification, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

Claims

1. A monitoring method based on artificial intelligence, characterized in that: The method is applied to an artificial intelligence-based monitoring system, which includes a server and a client. The monitoring method includes: The server receives the control rules of the current control task and the camera list that needs to be controlled sent by the client, wherein the camera list includes multiple target cameras for capturing images of the target object; When the server starts the current monitoring task, it receives the trajectory data of the target object sent by the target camera, wherein the trajectory data includes trajectory data corresponding to pedestrians and trajectory data corresponding to vehicles; The server determines a quality score corresponding to the target object based on conditional parameter values of the target object and a preset quality determination function, wherein the quality scores include a quality score of a pedestrian and a quality score of the vehicle, the conditional parameter values include blurriness of a facial image, a three-dimensional deflection angle of the face, and brightness of the facial image, and the quality determination function is determined by performing data curve fitting based on a facial recognition accuracy distribution and / or a confidence distribution corresponding to the conditional parameters of the facial image to obtain a fitting function; The server determines, based on the quality score of the pedestrian, a pedestrian image to be compared in the pedestrian trajectory data; The server obtains, based on the pedestrian image to be compared, a feature having the highest similarity to the facial feature of the pedestrian image to be compared, and triggers a pedestrian control alarm notification if the feature having the highest similarity satisfies the control rule; The server determines a vehicle image to be compared in the vehicle trajectory data based on the vehicle quality score, and extracts captured information of the vehicle from the vehicle image to be compared, thereby obtaining at least real-time captured vehicle information; the captured information includes the capture time and capture location, and the vehicle information includes a license plate thumbnail, license plate number, vehicle body color, vehicle type, vehicle brand, and vehicle system; The server performs vehicle attribute recognition and vehicle trajectory analysis on the real-time captured vehicle information to obtain vehicle attribute recognition results and vehicle trajectory analysis results, and searches for vehicles by time range, location, and vehicle attributes to obtain vehicle search results. The server combines the captured vehicle information and the algorithm recognition results to perform control and monitoring to obtain control and monitoring results. If the vehicle attribute recognition result, the vehicle trajectory analysis result, the vehicle search result, and the control monitoring result meet the control rules, a vehicle control alarm notification is triggered; The server determining, based on the quality score of the pedestrian, a pedestrian image to be compared in the pedestrian trajectory data, comprising: Determining a current target image in the pedestrian trajectory data, and calculating a quality score of the target object in the current target image; If the quality score is greater than or equal to a first preset value, determining the current target image as the image to be compared; If the quality score is less than the first preset value, re-entering the step of determining a current target image in the pedestrian trajectory data; If there is no target image with a quality score greater than or equal to the first preset value in the pedestrian trajectory data, determining the target image with the highest quality score as the pedestrian image to be compared; The server determining, based on the quality score of the vehicle, a vehicle image to be compared in the trajectory data of the vehicle includes: Determining a current target image in the trajectory data of the vehicle, and calculating a quality score of the target object in the current target image; If the quality score is greater than or equal to a first preset value, determining the current target image as the image to be compared; If the quality score is less than the first preset value, re-entering the step of determining a current target image in the trajectory data of the vehicle; If there is no target image with a quality score greater than or equal to the first preset value in the trajectory data of the vehicle, the target image with the highest quality score is determined as the vehicle image to be compared.

2. The monitoring method according to claim 1, characterized in that: Also includes: The server obtains a registered image and calculates a quality score of the registered image; If the quality score of the registered image is greater than or equal to a first preset value, calculating the similarity between the registered image and all images in the control image library; When all the similarities are less than a second preset value, the server saves the registered image to the control image library.

3. The monitoring method according to any one of claims 1 to 2, characterized in that: After receiving the trajectory data corresponding to the pedestrian and the trajectory data corresponding to the vehicle sent by the target camera, the method further includes: The server determines the target image with the highest quality score in the pedestrian trajectory data and the vehicle trajectory data as the image to be extracted, and obtains snapshot information of the image to be extracted; wherein the snapshot information includes at least a snapshot location and a snapshot time; The region corresponding to the target object is extracted from the image to be extracted as a captured image, and the captured image and the captured information corresponding to each captured image are saved in a captured database.

4. The monitoring method according to claim 3, characterized in that: Also includes: All captured images in the captured image database and the captured information corresponding to each captured image are displayed.

5. The monitoring method according to claim 3, characterized in that: The obtaining of the snapshot information of the image to be extracted includes: The server obtains the capture location and capture time of the image to be extracted and the vehicle information in the image to be extracted to obtain the captured information of the vehicle; The vehicle information includes any one or a combination of license plate thumbnail, license plate number, vehicle body color, vehicle type, vehicle brand, and vehicle system. The license plate thumbnail is specifically the area corresponding to the license plate number in the image to be extracted.

6. The monitoring method according to claim 5, characterized in that: Also includes: When the server receives a search command, it determines a search item corresponding to the search command; wherein the search item includes any one or a combination of the image to be searched and the captured information; The server determines a search result corresponding to the search item in the snapshot database.

7. The monitoring method according to claim 6, characterized in that: The search item includes an image to be searched, and determining a search result corresponding to the search item in the snapshot database includes: The server determines the object to be retrieved in the image to be retrieved; The server compares the area corresponding to the object to be retrieved in the image to be retrieved with the image in the snapshot database to obtain a retrieval result.

8. The monitoring method according to claim 6, characterized in that: When the search item is specifically the license plate number of the target vehicle, after determining the search result corresponding to the search item in the snapshot database, the method further includes: The server displays the driving track of the target vehicle on a map based on the captured location in the retrieval result.

9. The monitoring method according to claim 3, characterized in that: Also includes: When the server receives the operation command, it determines the operation object corresponding to the operation command; wherein the operation object includes the control image library or the snapshot database; The server performs a management operation corresponding to the operation command on the operation object; wherein the management operation includes any one of adding an image to the operation object, deleting an image from the operation object, refreshing the operation object, classifying all images in the operation object, uploading or modifying image information of an image in the operation object.

10. A monitoring device based on artificial intelligence, characterized in that: The monitoring device includes a target camera, a server and a client; The server receives the control rules of the current control task and the camera list that needs to be controlled sent by the client, wherein the camera list includes multiple target cameras for capturing images of the target object; When the server starts the current monitoring task, it receives the trajectory data of the target object sent by the target camera, wherein the trajectory data includes the trajectory data of pedestrians and the trajectory data corresponding to vehicles; The server determines a quality score corresponding to the target object based on conditional parameter values of the target object and a preset quality determination function, wherein the quality scores include a quality score of a pedestrian and a quality score of the vehicle, the conditional parameter values include blurriness of a facial image, a three-dimensional deflection angle of the face, and brightness of the facial image, and the quality determination function is determined by performing data curve fitting based on a facial recognition accuracy distribution and / or a confidence distribution corresponding to the conditional parameters of the facial image to obtain a fitting function; The server determines, based on the quality score of the pedestrian, a pedestrian image to be compared in the pedestrian trajectory data; The server obtains, based on the pedestrian image to be compared, a feature having the highest similarity to the facial feature of the pedestrian image to be compared, and triggers a pedestrian control alarm notification if the feature having the highest similarity satisfies the control rule; The server determines a vehicle image to be compared in the vehicle trajectory data based on the vehicle quality score, and extracts captured information of the vehicle from the vehicle image to be compared, thereby obtaining at least real-time captured vehicle information; the captured information includes the capture time and capture location, and the vehicle information includes a license plate thumbnail, license plate number, vehicle body color, vehicle type, vehicle brand, and vehicle system; The server performs vehicle attribute recognition and vehicle trajectory analysis on the real-time captured vehicle information to obtain vehicle attribute recognition results and vehicle trajectory analysis results, and searches for vehicles by time range, location, and vehicle attributes to obtain vehicle search results. The server combines the captured vehicle information and the algorithm recognition results to perform control and monitoring to obtain control and monitoring results. If the vehicle attribute recognition result, the vehicle trajectory analysis result, the vehicle search result, and the control monitoring result meet the control rules, a vehicle control alarm notification is triggered; The server determining, based on the quality score of the pedestrian, a pedestrian image to be compared in the pedestrian trajectory data, comprising: Determining a current target image in the pedestrian trajectory data, and calculating a quality score of the target object in the current target image; If the quality score is greater than or equal to a first preset value, determining the current target image as the image to be compared; If the quality score is less than the first preset value, re-entering the step of determining a current target image in the pedestrian trajectory data; If there is no target image with a quality score greater than or equal to the first preset value in the pedestrian trajectory data, determining the target image with the highest quality score as the pedestrian image to be compared; The server determining, based on the quality score of the vehicle, a vehicle image to be compared in the trajectory data of the vehicle includes: Determining a current target image in the trajectory data of the vehicle, and calculating a quality score of the target object in the current target image; If the quality score is greater than or equal to a first preset value, determining the current target image as the image to be compared; If the quality score is less than the first preset value, re-entering the step of determining a current target image in the trajectory data of the vehicle; If there is no target image with a quality score greater than or equal to the first preset value in the trajectory data of the vehicle, the target image with the highest quality score is determined as the vehicle image to be compared.

11. The device according to claim 10, characterized in that The first determination module includes: a calculation unit, configured to determine a current target image in the pedestrian trajectory data and calculate a quality score of the target object in the current target image; a first determining unit, configured to determine the current target image as an image to be compared if the quality score is greater than or equal to a first preset value; a second determining unit, configured to re-enter the step of determining a current target image in the trajectory data if the quality score is less than the first preset value; The third determining unit is configured to determine the target image with the highest quality score as the image to be compared if there is no target image with a quality score greater than or equal to the first preset value in the pedestrian trajectory data.

12. The device according to any one of claims 10 to 11, characterized in that Also includes: A second acquisition module is configured to determine the target image with the highest quality score in the pedestrian trajectory data as the image to be extracted, and obtain snapshot information of the image to be extracted; wherein the snapshot information includes at least a snapshot location and a snapshot time; The extraction module is used to extract the area corresponding to the target object in the image to be extracted as a captured image, and save the captured image and the captured information corresponding to each captured image into a captured database.

13. The device according to claim 12, characterized in that Also includes: The display module is used to display all captured images in the captured database and the captured information corresponding to each captured image.

14. The device according to claim 12, characterized in that Also includes: a fourth determining module, configured to, upon receiving an operation command, determine an operation object corresponding to the operation command; wherein the operation object includes a control image library or the snapshot database; The second execution module is used to perform the management operation corresponding to the operation command on the operation object; wherein the management operation includes any one of adding an image to the operation object, deleting an image from the operation object, refreshing the operation object, classifying all images in the operation object, uploading or modifying image information of the image in the operation object.

15. A server, characterized in that: include: memory for storing computer programs; A processor, configured to implement the steps of the monitoring method according to any one of claims 1 to 9 when executing the computer program.

16. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the monitoring method according to any one of claims 1 to 9.

Images