Algorithm cloud scheduling management method, system and readable storage medium for AI cameras

Abstract

The present application relates to an algorithm cloud scheduling management method, system and readable storage medium for an AI camera. The method includes: obtaining multiple scene algorithm packages of a registered AI camera for functional testing, and storing the first scene algorithm package that passes the functional test; judging whether to authorize the operation of the registered AI camera according to the operation request of the registered AI camera obtained; if the registered AI camera passes the operation authorization, obtaining the default algorithm package of the registered AI camera from the first scene algorithm package, and sending a first operation instruction to the registered AI camera; the registered AI camera executes the first operation instruction according to the default algorithm package, and returns the first operation result. This method is used to implement functional testing of scene algorithm packages and algorithm package configuration of different types of AI cameras, reduce the impact of algorithm package errors on AI cameras, and improve the operation stability of AI cameras.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an algorithm cloud scheduling management method, system and readable storage medium for an AI camera. Background Art

[0002] With the rapid development of AI smart cameras in the field of artificial intelligence, the market demand for AI cameras has become increasingly strong, and both the number of camera categories and the number of AI algorithm scenarios have seen explosive growth.

[0003] However, the wide variety of algorithm scenarios and device models will not only affect the user experience, but also greatly increase the maintenance cost and operation difficulty of the operator. The iterative upgrade of various AI algorithm products will also increase the R&D and maintenance manpower costs of equipment manufacturers. Even for users of a single device, the demand for camera usage will change with the change of usage scenarios. For example, merchant users want the device to realize scene functions such as customer flow statistics during store business hours, but outside of business hours, the device needs to realize security functions such as regional monitoring and fire identification.

[0004] In the prior art, although algorithm scheduling for various AI cameras can be achieved, the compatibility issues between AI cameras and algorithms are not taken into consideration, and abnormal operation of AI cameras often occurs during the scheduling process. Summary of the invention

[0005] Based on this, it is necessary to provide an algorithm cloud scheduling management method, system and readable storage medium for AI cameras that can realize the evaluation storage and version update of the scene algorithm package of AI cameras, reduce the iteration cost of algorithm packages, adaptively schedule algorithms for AI cameras, and improve the operating stability of AI cameras in response to the above-mentioned technical problems.

[0006] In a first aspect, the present application provides an algorithm cloud scheduling management method for an AI camera, the method comprising:

[0007] Obtain multiple scene algorithm packages of the registered AI camera for functional testing, and store the first scene algorithm package that passes the functional test;

[0008] Determine whether to authorize the operation of the registered AI camera according to the acquired operation request of the registered AI camera;

[0009] If the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera; the registered AI camera executes the first operation instruction according to the default algorithm package and returns a first operation result.

[0010] In one embodiment, the step of obtaining a scene algorithm package of a registered AI camera for functional testing and storing a first scene algorithm package that passes the functional test comprises:

[0011] Acquire a test data set and a test prototype of the scene algorithm package; the test prototype simulates the function of the registered AI camera;

[0012] The scenario algorithm package and the test data set are sent to the test prototype; the test prototype runs the scenario algorithm package and tests the test data set to obtain a test result set of the test data set;

[0013] The test result set is summarized and analyzed to determine the functional test results of the scenario algorithm package.

[0014] In one embodiment, the test result set includes a correct identification positive example and an incorrect identification positive example, and the summarizing and analyzing the test result set to determine the functional test result of the scenario algorithm package includes:

[0015] Calculating the recall rate and the precision rate of the test result set according to the correctly identified positive examples and the incorrectly identified positive examples;

[0016] The recall rate is compared with a preset recall threshold, and the accuracy rate is compared with a preset accuracy threshold. If the recall rate is greater than the preset recall threshold and the accuracy rate is greater than the preset accuracy threshold, the scene algorithm package is marked as the first scene algorithm package for storage.

[0017] In one embodiment, judging whether to authorize the operation of the registered AI camera according to the acquired operation request of the registered AI camera includes:

[0018] Parse the operation request to obtain device information of the registered AI camera;

[0019] The device information is verified, and if it passes, the registered AI camera is authorized to operate.

[0020] In one embodiment, the operation request also includes the latest key of the registered AI camera, and the latest key is automatically updated when the registered AI camera is powered on; if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and the first operation instruction is sent to the registered AI camera, including:

[0021] If the registered AI camera passes the operation authorization, obtaining the default algorithm package of the registered AI camera from the first scene algorithm package, and configuring the registered AI camera;

[0022] According to the latest key of the registered AI camera, the business data is padded and encrypted, and the first operation instruction is generated and sent to the registered AI camera; the registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result.

[0023] In one embodiment, the method further comprises:

[0024] Receive a first operation result returned by the registered AI camera;

[0025] According to the first operation result, determining the operation status of the registered AI camera;

[0026] If the first running result shows that the default algorithm package of the registered AI camera executes abnormally, obtain the preferred algorithm package in the first scene algorithm package and reconfigure the registered AI camera; the preferred algorithm package is the latest version of the algorithm package in the first scene algorithm package.

[0027] In one embodiment, the method further comprises:

[0028] According to the camera scheduling request sent by the third-party platform, the target AI camera is determined, and the corresponding instructions to be executed and prompt instructions are generated;

[0029] Sending a prompt instruction to the target AI camera; the target AI camera calls and executes the to-be-executed instruction according to the prompt instruction.

[0030] In one embodiment, before obtaining multiple scene algorithm packages of the registered AI camera for functional testing, the method further includes:

[0031] Obtain device information of multiple AI cameras to be registered; generate a registration account for the AI ​​cameras to be registered based on the device information and mark them as the registered AI cameras.

[0032] In a second aspect, the present application also provides an algorithm cloud scheduling management system for an AI camera, the system comprising:

[0033] A process approval module is used to obtain multiple scene algorithm packages of a registered AI camera for functional testing, and store the first scene algorithm package that passes the functional test; and determine whether to authorize the operation of the registered AI camera according to the operation request of the registered AI camera obtained;

[0034] The application center module is used to obtain the default algorithm package of the registered AI camera from the first scene algorithm package if the registered AI camera passes the operation authorization, and send a first operation instruction to the registered AI camera; the registered AI camera executes the first operation instruction according to the default algorithm package and returns a first operation result.

[0035] In a third aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps corresponding to the method described in the first aspect above are implemented.

[0036] The algorithm cloud scheduling management method, system and readable storage medium of the above-mentioned AI camera perform functional testing by obtaining multiple scene algorithm packages of the registered AI camera, and store the first scene algorithm package that passes the functional test; based on the obtained operation request of the registered AI camera, determine whether to authorize the operation of the registered AI camera; if the registered AI camera passes the operation authorization, obtain the default algorithm package of the registered AI camera from the first scene algorithm package, and send a first operation instruction to the registered AI camera, and the registered AI camera executes the first operation instruction according to the default algorithm package, and returns a first operation result, thereby realizing functional testing of the scene algorithm package and algorithm package configuration of different types of AI cameras, reducing the impact of algorithm package errors on the AI ​​camera, and improving the operation stability of the AI ​​camera. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments of the present application or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 This is an application environment diagram of an algorithm cloud scheduling management method for an AI camera in one embodiment;

[0039] Figure 2 An algorithm cloud scheduling management system for an AI camera in one embodiment;

[0040] Figure 3 It is a flowchart of an algorithm cloud scheduling management method for an AI camera in one embodiment;

[0041] Figure 4 A schematic diagram of a flow chart of functional testing of a scene algorithm package in one embodiment;

[0042] Figure 5 A schematic diagram of a process for summarizing and analyzing a test result set in one embodiment;

[0043] Figure 6 A schematic diagram of a process for determining whether to authorize the operation of a registered AI camera in one embodiment;

[0044] Figure 7 A schematic diagram of a process of issuing a first operation instruction using the latest key in an embodiment;

[0045] Figure 8 A schematic diagram of a process of analyzing a first operation result returned by a registered camera in one embodiment;

[0046] Fig. 9 A schematic diagram of a process of interacting with a third-party platform in one embodiment;

[0047] Fig.10 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantity limitation, and may indicate the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0050] The algorithm cloud scheduling management method for AI cameras provided in the embodiments of the present application can be applied to Figure 1 In the application environment shown, the AI ​​camera terminal 102 communicates with the server 104 through the network. The data storage system 106 can store the data that the server 104 needs to process. The data storage system 106 can be integrated on the server 104, or it can be placed on the cloud or other network servers.

[0051] On the server 104, multiple scene algorithm packages of the registered AI camera terminal 102 are obtained for functional testing, and the first scene algorithm package that passes the functional test is stored; according to the operation request of the registered AI camera terminal 102 obtained, it is determined whether to authorize the operation of the registered AI camera terminal 102; if the registered AI camera terminal 102 passes the operation authorization, the default algorithm package of the registered AI camera terminal 102 is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera terminal 102. The registered AI camera terminal 102 executes the first operation instruction according to the default algorithm package and returns a first operation result.

[0052] Among them, the AI ​​camera terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices and portable wearable devices. IoT devices can be smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0053] In an exemplary embodiment, Figure 2 As shown, an algorithm cloud scheduling management system for AI cameras is provided, and the system is built on Figure 1 The system includes: a metadata management module 201, a process approval module 202, an application center module 203, a license management module 204, a troubleshooting support module 205 and a partner center module 206. Each module is connected to each other in communication.

[0054] The metadata management module 201 is used to register the device information of the AI ​​camera, realize the information registration and storage of the AI ​​camera, the capability configuration of the scene algorithm package, etc. The device information includes the device manufacturer, chip model, device model, scene algorithm package, etc.

[0055] The process approval module 202 is used for the listing application, functional testing, test process status query, test record query, etc. of the scene algorithm package. For example, multiple scene algorithm packages of the registered AI camera are obtained for functional testing, and the first scene algorithm package that passes the functional test is stored; according to the operation request of the registered AI camera obtained, it is determined whether to authorize the operation of the registered AI camera.

[0056] The application center module 203: is used to configure the device model and function of the AI ​​camera, manage the default algorithm packages supported by different AI cameras, and realize unified management and control of the default AI intelligent scene algorithm of the device from the device model level. And the query of the scene algorithm package is provided within the module, and the scene algorithm package within the control range of the registered account of the current AI camera can be fully queried. For example, if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera; the registered AI camera executes the first operation instruction according to the default algorithm package, and returns a first operation result.

[0057] The license management module 204 is used to manage various certification licenses such as the authorization of designated AI camera devices, registration of third-party platforms, and platform authentication. Only registered devices and third-party platforms can interact with this system; the designated AI camera device authorization is to authorize the AI ​​camera to perform specific AI scene algorithm authorization from the device model dimension to achieve the customized AI scene algorithm operation requirements of specific models.

[0058] The troubleshooting support module 205 is used for the current AI camera running algorithm query, algorithm scheduling log query, AI camera device binding log query, AI camera device problem reporting query, etc. Configure corresponding instructions for various problems to restore the AI ​​camera to normal operation.

[0059] The partner center module 206 is used to manage the registered accounts of various users, divide the registered accounts into roles and assign different operation permissions.

[0060] In an exemplary embodiment, Figure 3 As shown, an algorithm cloud scheduling management method for an AI camera is provided, and the method is applied to Figure 1The system on the server 104 in the example is used for explanation, including the following steps 301 to 303.

[0061] Step 301, obtain multiple scene algorithm packages of the registered AI camera for functional testing, and store the first scene algorithm package that passes the functional test.

[0062] The registered AI camera is an AI camera that has been successfully registered on the system.

[0063] Specifically, the device information of multiple AI cameras to be registered is obtained, and according to the device information, a registration account is generated for the AI ​​camera to be registered and marked as the registered AI camera, and role permissions are assigned to the registered account. Multiple scene algorithm packages of the registered AI camera are obtained for functional testing, and the first scene algorithm package that passes the functional test is stored.

[0064] Step 302: Determine whether to authorize the operation of the registered AI camera according to the acquired operation request of the registered AI camera.

[0065] Specifically, although the registered AI camera has registered an account on the system, each time the registered AI camera is powered on again, it is necessary to obtain an operation request of the registered AI camera, and the system determines whether to authorize the operation of the registered AI camera.

[0066] Step 303: If the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera.

[0067] Among them, the registered AI camera executes the first operation instruction according to the default algorithm package and returns a first operation result.

[0068] Specifically, if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera. If the registered AI camera fails to pass the operation authorization, an authorization error message is returned to the registered AI camera terminal.

[0069] In the algorithm cloud scheduling management method of the above-mentioned AI camera, functional testing is performed by obtaining multiple scene algorithm packages of the registered AI camera, and the first scene algorithm package that passes the functional test is stored; according to the obtained operation request of the registered AI camera, it is determined whether to authorize the operation of the registered AI camera; if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera. The registered AI camera executes the first operation instruction according to the default algorithm package and returns a first operation result, thereby realizing functional testing of the scene algorithm package and algorithm package configuration of different types of AI cameras, reducing the impact of algorithm package errors on the AI ​​camera, and improving the operation stability of the AI ​​camera.

[0070] In an exemplary embodiment, Figure 4 As shown, step 301 obtains the scene algorithm package of the registered AI camera for functional testing, and stores the first scene algorithm package that passes the functional test, specifically including the following steps 401 to 403. Among them:

[0071] Step 401, obtaining a test data set and a test prototype of the scene algorithm package.

[0072] Wherein, the test prototype simulates the function of the registered AI camera.

[0073] Step 402: Send the scenario algorithm package and the test data set to the test prototype.

[0074] The test prototype runs the scenario algorithm package and tests the test data set to obtain a test result set of the test data set.

[0075] The test result set includes correctly identified positive examples and incorrectly identified positive examples.

[0076] Step 403, summarize and analyze the test result set to determine the functional test results of the scenario algorithm package.

[0077] Specifically, obtain a test data set of the scene algorithm package, such as an image set or a video set. Send the scene algorithm package and the test data set to a test prototype, issue a test instruction to the test prototype, run the scene algorithm package, and obtain a test result set of the test data set according to the test case. Summarize and analyze the correct recognition positive examples and the incorrect recognition positive examples in the test result set to determine the functional test results of the scene algorithm package.

[0078] In this embodiment, by testing the scene algorithm package using a test data set on a test prototype, the function detection of the scene algorithm package is implemented, thereby avoiding errors in the AI ​​camera due to problems with the algorithm package itself when the algorithm package is subsequently scheduled for use.

[0079] In one embodiment, if Figure 5 As shown, step 403 summarizes and analyzes the test result set to determine the functional test results of the scenario algorithm package, which specifically includes the following steps 501 to 502.

[0080] Step 501, calculating the recall rate and the precision rate of the test result set according to the correctly identified positive examples and the incorrectly identified positive examples.

[0081] The recall rate is the number of correctly identified positive examples divided by the number of all test cases. The accuracy rate is the number of correctly identified positive examples divided by the total number of samples in the test result set.

[0082] Step 502, compare the recall rate with the preset recall threshold, and compare the accuracy rate with the preset accuracy threshold. If the recall rate is greater than the preset recall threshold and the accuracy rate is greater than the preset accuracy threshold, mark the scene algorithm package as the first scene algorithm package for storage.

[0083] Specifically, if the recall rate is greater than the preset recall threshold, and the accuracy rate is greater than the preset accuracy threshold, the scene algorithm package is marked as the first scene algorithm package for storage and enters the shelf process. Otherwise, the scene algorithm package fails the functional test and cannot enter the shelf process.

[0084] In this embodiment, the recall rate and accuracy of the test result set of the scenario algorithm package are calculated, and the recall rate and accuracy are compared with the thresholds respectively to determine whether the scenario algorithm package passes the functional test. Only the scenario algorithm package with a recall rate greater than the preset recall threshold and an accuracy rate greater than the preset accuracy threshold can be marked as the first scenario algorithm package, thereby improving the functional testing accuracy of the scenario algorithm package.

[0085] In an exemplary embodiment, Figure 6 As shown, step 302 determines whether to authorize the operation of the registered AI camera according to the operation request of the registered AI camera obtained, and specifically includes the following steps 601 to 602.

[0086] Step 601: parse the operation request to obtain the device information of the registered AI camera.

[0087] Step 602: Verify the device information. If the device information passes, authorize the registered AI camera to run.

[0088] Specifically, the operation request is parsed to obtain the device information of the registered AI camera, and the device information is verified. If it matches the device information stored when registering on the system, the registered AI camera is authorized to operate. Otherwise, an authorization error message is returned to the AI ​​camera terminal.

[0089] In one embodiment, the operation request also includes the latest key of the registered AI camera, and the latest key is automatically updated when the registered AI camera is powered on. Figure 7 As shown, in step 303, if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and a first operation instruction is sent to the registered AI camera, specifically including the following steps 701 to 702.

[0090] Step 701: If the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and the registered AI camera is configured.

[0091] Step 702: fill and encrypt the business data according to the latest key of the registered AI camera, generate the first operation instruction and send it to the registered AI camera; the registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result.

[0092] Specifically, after the registered AI camera passes the operation authorization, the system selects the default algorithm package from the first scene algorithm package according to the functions configured by the device model of the AI ​​camera and sends it to the registered AI camera for configuration. Since the operation request of the registered AI camera contains the latest key generated after this power-on, the business data to be sent is padded and encrypted according to the ECB mode of the AES algorithm combined with PKCS5Padding according to the latest key to generate the first operation instruction. The registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result to the system.

[0093] In this embodiment, by automatically updating the latest key of the registered AI camera each time it is powered on, not only is "one machine, one key" achieved between the system and each registered AI camera, but also the encrypted issuance and decrypted execution of instructions are achieved, thereby improving the security of the operation of the registered AI camera.

[0094] In an exemplary embodiment, after step 303, as Figure 8 As shown, the method further includes the following steps 304 to 306.

[0095] Step 304: Receive the first operation result returned by the registered AI camera.

[0096] Step 305: Determine the operating status of the registered AI camera according to the first operating result.

[0097] Step 306: If the first running result shows that the default algorithm package of the registered AI camera executes abnormally, obtain the preferred algorithm package in the first scene algorithm package and reconfigure the registered AI camera.

[0098] Among them, the preferred algorithm package is the latest version algorithm package in the first scenario algorithm package.

[0099] Specifically, the first operation result includes the result after the first execution instruction is successfully executed or the result returned when the first execution instruction fails to execute due to an exception in the default algorithm package. The operation status of the registered AI camera is judged according to the first operation result. If the first operation result shows that the default algorithm package of the registered AI camera executes abnormally, the preferred algorithm package in the first scene algorithm package, that is, the latest version of the algorithm package, is obtained, and the registered AI camera is reconfigured. The reconfigured registered AI camera receives the first execution instruction re-issued by the system and continues to run. If the first operation result shows that the first execution instruction is successfully executed, the first operation result is stored.

[0100] Optionally, when the first running result shows that the default algorithm package of the registered AI camera executes abnormally, in addition to obtaining the latest version of the algorithm package for configuration, the default algorithm package can be obtained again for reconfiguration.

[0101] In this embodiment, by analyzing the first running result, the running status of the algorithm package on the currently registered AI camera, the camera is scheduled for algorithm scheduling, thereby improving the timeliness and accuracy of the AI ​​camera algorithm scheduling.

[0102] In an exemplary embodiment, the system also has the ability to interact with a third-party platform, and can receive scheduling from the third-party platform to implement the direct scheduling function of the AI ​​camera. Fig. 9 As shown, the method further includes the following steps 801 to 802.

[0103] Step 801, determine the target AI camera according to the camera scheduling request sent by the third-party platform, and generate corresponding instructions to be executed and prompt instructions.

[0104] Specifically, before obtaining the camera scheduling instruction sent by the third-party platform, the third-party platform first sends a device configuration acquisition request to the system, and the system returns the configuration information of the specified AI camera to the third-party platform according to the device configuration acquisition request. The third-party platform sends a camera call request to the system according to the configuration information. The system determines the target AI camera according to the camera scheduling request, and generates the to-be-executed instructions and prompt instructions corresponding to the target AI camera.

[0105] Step 802: Send a prompt instruction to the target AI camera. The target AI camera calls and executes the to-be-executed instruction according to the prompt instruction.

[0106] Specifically, the prompt instruction is forwarded to the target AI camera by an external unified signaling platform. The target AI camera actively queries the corresponding instructions to be executed from the system according to the prompt instruction. After the query is successful, the system will respond to the third-party platform that the current scheduling operation is successful. When the target AI camera completes the execution of the instructions to be executed according to the configuration of the default algorithm package, the operation result will be returned to the system. If the operation result shows that there is a problem in the instruction execution process, such as algorithm response timeout, abnormal algorithm exit, download failure, etc., the system will re-acquire the latest information of the specified algorithm package for secondary scheduling and issuance, which can effectively reduce the frequency of manual troubleshooting and keep the equipment running stably. The third-party platform can confirm the instruction scheduling result of the target AI camera by checking the current running status of the algorithm package.

[0107] In a preferred embodiment, an algorithm cloud scheduling management method for an AI camera is provided, which is applied to Figure 2 The algorithm cloud scheduling management system of the AI ​​camera shown in the figure includes the following contents:

[0108] S1, obtaining device information of multiple AI cameras to be registered, generating a registration account for the AI ​​camera to be registered according to the device information and marking it as the registered AI camera, and assigning role permissions to the registration account.

[0109] S2, obtain multiple scene algorithm packages of the registered AI camera, test data sets of each scene algorithm package, and test prototypes for functional testing. If the test prototype runs the scene algorithm package to test the test data set, and the obtained test result set shows that the recall rate of the scene algorithm package is greater than the preset recall threshold and the accuracy rate is greater than the preset accuracy threshold, then the scene algorithm package is marked as the first scene algorithm package for storage.

[0110] S3, according to the operation request of the registered AI camera obtained, the operation request is parsed to obtain the device information of the registered AI camera. The device information is verified to determine whether the operation of the registered AI camera is authorized. If the device information matches the device information stored when the system is registered, the operation of the registered AI camera is authorized, otherwise, the authorization error information is returned to the AI ​​camera terminal.

[0111] S4, if the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and the registered AI camera is configured. If the registered AI camera fails to pass the operation authorization, an authorization error message is returned to the registered AI camera terminal.

[0112] S5, according to the latest key of the registered AI camera, the business data is padded and encrypted in ECB mode of AES algorithm combined with PKCS5Padding, a first operation instruction is generated, and sent to the registered AI camera. The registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result.

[0113] S6, receiving the first operation result returned by the registered AI camera. If the first operation result shows that the default algorithm package of the registered AI camera is executed abnormally, the latest version of the algorithm package in the first scene algorithm package is obtained, and the registered AI camera is reconfigured. The reconfigured registered AI camera accepts the first execution instruction re-issued by the system and continues to run. If the first operation result shows that the first execution instruction is successfully executed, the first operation result is stored.

[0114] In this preferred embodiment, by using the accuracy and recall of the algorithm package test results to evaluate the performance of the algorithm package, it is not only possible to standardize the algorithm listing process, but also to control the performance level of all listed algorithms to a certain extent, and improve the stability of the algorithm package running on the AI ​​camera. At the same time, in the authorization and instruction scheduling process of the AI ​​camera, combined with the latest key generated after each power-on of the AI ​​camera, not only "one machine, one secret" is realized, but also the information security of the interaction process between the AI ​​camera and the system is ensured. By supporting the operation and scheduling of AI algorithm packages compatible with multiple usage scenarios, the user experience of a single AI scene algorithm is improved, and the running algorithm package can be freely combined to support customization, maximizing the rich user experience of the AI ​​camera, and recording the algorithm operation status and scheduling records at the same time, realizing quantitative management for the use process of the AI ​​camera. In addition, when an error occurs during operation, it is reported, and by configuring solutions for different problem types, the algorithm is automatically scheduled to solve the problem, which reduces the frequency of participation of maintenance personnel and improves the operation stability of the AI ​​camera to a certain extent.

[0115] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0116] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.10 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store AI camera data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an algorithm cloud scheduling management method for an AI camera is implemented.

[0117] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0118] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps corresponding to the methods described in the above embodiments are implemented.

[0119] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps corresponding to the methods described in the above embodiments are implemented.

[0120] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0121] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0122] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An algorithm cloud scheduling management method for an AI camera, characterized in that: The method comprises: Obtain multiple scene algorithm packages of the registered AI camera for functional testing, and store the first scene algorithm package that passes the functional test; According to the acquired operation request of the registered AI camera, determine whether to authorize the operation of the registered AI camera; the operation request also includes the latest key of the registered AI camera, and the latest key is automatically updated when the registered AI camera is powered on; If the registered AI camera passes the operation authorization, the default algorithm package of the registered AI camera is obtained from the first scene algorithm package, and the registered AI camera is configured; according to the latest key of the registered AI camera, the business data is filled and encrypted, and a first operation instruction is generated and sent to the registered AI camera; the registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result; Receive a first operation result returned by the registered AI camera; According to the first operation result, determining the operation status of the registered AI camera; If the first running result shows that the default algorithm package of the registered AI camera executes abnormally, obtain the preferred algorithm package in the first scene algorithm package and reconfigure the registered AI camera; the preferred algorithm package is the latest version of the algorithm package in the first scene algorithm package.

2. The algorithm cloud scheduling management method for AI cameras according to claim 1, characterized in that: The obtaining of a scene algorithm package of a registered AI camera for function testing and storing a first scene algorithm package that passes the function test comprises: Acquire a test data set and a test prototype of the scene algorithm package; the test prototype simulates the function of the registered AI camera; The scenario algorithm package and the test data set are sent to the test prototype; the test prototype runs the scenario algorithm package and tests the test data set to obtain a test result set of the test data set; The test result set is summarized and analyzed to determine the functional test results of the scenario algorithm package.

3. The algorithm cloud scheduling management method for an AI camera according to claim 2, characterized in that: The test result set includes a correct identification positive example and an incorrect identification positive example, and the summary analysis of the test result set to determine the functional test result of the scenario algorithm package includes: Calculating the recall rate and the precision rate of the test result set according to the correctly identified positive examples and the incorrectly identified positive examples; The recall rate is compared with a preset recall threshold, and the accuracy rate is compared with a preset accuracy threshold. If the recall rate is greater than the preset recall threshold and the accuracy rate is greater than the preset accuracy threshold, the scene algorithm package is marked as the first scene algorithm package for storage.

4. The algorithm cloud scheduling management method for an AI camera according to claim 1, characterized in that: The determining whether to authorize the operation of the registered AI camera according to the acquired operation request of the registered AI camera includes: Parse the operation request to obtain device information of the registered AI camera; The device information is verified, and if it passes, the registered AI camera is authorized to operate.

5. The algorithm cloud scheduling management method for AI cameras according to claim 1, characterized in that: The method further comprises: According to the camera scheduling request sent by the third-party platform, the target AI camera is determined, and the corresponding instructions to be executed and prompt instructions are generated; Sending a prompt instruction to the target AI camera; the target AI camera calls and executes the to-be-executed instruction according to the prompt instruction.

6. The algorithm cloud scheduling management method for AI cameras according to claim 1, characterized in that: Before obtaining multiple scene algorithm packages of the registered AI camera for functional testing, the method further includes: Obtain device information of multiple AI cameras to be registered; generate a registration account for the AI ​​cameras to be registered based on the device information and mark them as the registered AI cameras.

7. An algorithm cloud scheduling management system for AI cameras, characterized in that: The system comprises: A process approval module is used to obtain multiple scene algorithm packages of a registered AI camera for functional testing, and store the first scene algorithm package that passes the functional test; according to the obtained operation request of the registered AI camera, determine whether to authorize the operation of the registered AI camera; the operation request also includes the latest key of the registered AI camera, and the latest key is automatically updated when the registered AI camera is powered on; The application center module is used for obtaining the default algorithm package of the registered AI camera from the first scene algorithm package and configuring the registered AI camera if the registered AI camera passes the operation authorization; filling and encrypting the business data according to the latest key of the registered AI camera, generating a first operation instruction and sending it to the registered AI camera; the registered AI camera decrypts and executes the first operation instruction according to the latest key, and returns the first operation result; Receive a first operation result returned by the registered AI camera; According to the first operation result, determining the operation status of the registered AI camera; If the first running result shows that the default algorithm package of the registered AI camera executes abnormally, obtain the preferred algorithm package in the first scene algorithm package and reconfigure the registered AI camera; the preferred algorithm package is the latest version of the algorithm package in the first scene algorithm package.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images