Method, device and electronic device for upgrading intelligent recognition algorithm of device

Abstract

The present application relates to a method, device and electronic device for upgrading an intelligent recognition algorithm of a device. The method includes: dividing the intelligent recognition devices to be upgraded into a first type of devices with a remaining capacity value greater than a preset threshold and a second type of devices with a remaining capacity value not greater than the preset threshold; sending the intelligent recognition algorithm model to be upgraded to the first type of devices, allocating the task of extracting the feature values of the pictures in the comparison library to the first type of devices, and obtaining the feature value extraction results obtained by the first type of devices executing the feature value extraction task according to the intelligent recognition algorithm model; sending the intelligent recognition algorithm model to the second type of devices, and sending the feature value extraction results of all the pictures in the comparison library to all the intelligent recognition devices to be upgraded. Through the present application, the problem that the intelligent recognition function cannot be used normally during the upgrade process of the algorithm model of the intelligent recognition device is solved. The effect that the intelligent recognition function of the intelligent recognition device is not affected when upgrading the algorithm model of the intelligent recognition device is achieved.

Description

Technical Field

[0001] This application relates to the technical field of video surveillance, and particularly to a method, device, and electronic device for upgrading an intelligent recognition algorithm of a device. Background Art

[0002] In the field of video surveillance, a series of algorithms are currently used to detect the images that appear in the surveillance scene. At the same time, for some structured objects, such as human faces or license plates, etc., after detection, it is necessary to compare with the images in the pre-prepared comparison library, so as to be able to identify the specific information of a person or a vehicle; to improve the relevant comparison efficiency, often the images in the comparison library are first extracted with eigenvalue information to generate a binary information of a fixed length, called an eigenvalue. After the extraction is completed, the relevant eigenvalues and the corresponding images are stored one by one, and then these eigenvalues are used to compare with the eigenvalues of the newly recognized images in the actual face or vehicle comparison process, which can improve the comparison efficiency.

[0003] With the continuous application of the recognition scene, the accuracy of the algorithm needs to be continuously improved in the application scene. For this reason, people continuously improve the algorithm model to improve the accuracy of the algorithm. For special application scenes, it makes it more convenient for users to use intelligent recognition devices. However, after upgrading to a new algorithm model, there is a situation where the relevant original eigenvalues extracted in the original comparison library cannot be recognized. It is necessary to re-extract new eigenvalues according to the new algorithm, and then the new eigenvalues can be recognized by the intelligent recognition device. And the re-extraction of the new eigenvalues requires the computing power of the intelligent recognition device, and the computing power of the intelligent recognition device is limited. When performing the re-extraction of the new eigenvalues, usually the intelligent recognition function of the intelligent recognition device is turned off, and all the computing power of the intelligent recognition device is concentrated to be used for extracting the new eigenvalues, resulting in the situation that the intelligent recognition function of the intelligent recognition device cannot be used normally during the upgrade process of the algorithm model.

[0004] Regarding the technical problem that the intelligent recognition function of the intelligent recognition device cannot be used normally during the upgrade process of the algorithm model in the related technology, no effective solution has been proposed yet. Summary of the Invention

[0005] In this embodiment, a method, device, and electronic device for upgrading an intelligent recognition algorithm of a device are provided to solve the technical problem that the intelligent recognition function of the intelligent recognition device cannot be used normally during the upgrade process of the algorithm model in the related technology.

[0006] In the first aspect, in this embodiment, a method for upgrading an intelligent recognition algorithm of a device is provided, including:

[0007] According to the capability information of the intelligent recognition device to be upgraded, the intelligent recognition device to be upgraded is divided into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold; the intelligent recognition algorithm model to be upgraded is sent to the first device, a task of extracting the feature values of the pictures in the comparison library is assigned to the first device, and the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model is acquired; after the feature value extraction results of all the pictures in the comparison library are obtained, the intelligent recognition algorithm model is sent to the second device, and the feature value extraction results of all the pictures in the comparison library are sent to all the intelligent recognition devices to be upgraded.

[0008] In some embodiments, when the intelligent recognition device to be upgraded is the first device, the method further includes: sending the intelligent recognition algorithm model to be upgraded to the first device, and acquiring the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model; after the feature value extraction results of all the pictures in the comparison library are obtained, sending the feature value extraction results of all the pictures in the comparison library to the first device.

[0009] In some embodiments, when there are multiple first devices, the method further includes: respectively acquiring the remaining capability values of each first device; according to the remaining capability values, assigning the task amount of extracting the feature values of the pictures in the comparison library and the range of the pictures in the comparison library corresponding to the task amount to each first device.

[0010] In some embodiments, when all the intelligent recognition devices to be upgraded are the second devices, the method further includes: adjusting the capability allocation of at least one of the second devices so that at least one of the second devices becomes the first device with a remaining capability value greater than the preset threshold.

[0011] In some embodiments, adjusting the capability allocation of at least one of the second devices includes: turning off at least one intelligent function of at least one of the second devices.

[0012] In some of these embodiments, before classifying the intelligent recognition device to be upgraded into a first device with a remaining capacity value greater than a preset threshold and a second device with a remaining capacity value not greater than the preset threshold according to the capacity information of the intelligent recognition device to be upgraded, the method further includes: determining whether the management platform of the intelligent recognition device to be upgraded has the ability to execute the eigenvalue extraction task based on the intelligent recognition algorithm model; when the management platform has the ability to execute the eigenvalue extraction task based on the intelligent recognition algorithm model, the management platform executes at least a part of the eigenvalue extraction task for the pictures in the comparison library according to the algorithm model.

[0013] In some of these embodiments, the method further includes: determining the comparison library and the intelligent recognition device to be upgraded according to the algorithm model, where the algorithm model corresponds to the comparison library, and the comparison library corresponds to the intelligent recognition device to be upgraded.

[0014] In a second aspect, an intelligent recognition algorithm upgrade device for a device is provided in this embodiment, including: an identification module, a first distribution module, a first acquisition module, a second acquisition module, and a second distribution module; the identification module is configured to classify the intelligent recognition device to be upgraded into a first device with a remaining capacity value greater than a preset threshold and a second device with a remaining capacity value not greater than the preset threshold according to the capacity information of the intelligent recognition device to be upgraded; the first distribution module is configured to send the algorithm model to be upgraded to the first device;

[0015] The first acquisition module: is configured to acquire the eigenvalue extraction result obtained by the first device executing the eigenvalue extraction task based on the intelligent recognition algorithm model; the second acquisition module is configured to acquire the eigenvalues of all the pictures in the comparison library; the second distribution module is configured to send the algorithm model to the second device, and send the eigenvalues of all the pictures in the comparison library to the first device and the second device.

[0016] In a third aspect, an electronic device is provided in this embodiment, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the intelligent recognition algorithm upgrade method for the device according to any one of the first aspects.

[0017] In a fourth aspect, a computer-readable storage medium is provided in this embodiment, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the intelligent recognition algorithm upgrade method for the device according to any one of the first aspects are implemented.

[0018] Compared with the related art, in an intelligent recognition algorithm upgrade method of a device provided in this embodiment, according to the capability information of the intelligent recognition device to be upgraded, the intelligent recognition device to be upgraded is divided into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold; the intelligent recognition algorithm model to be upgraded is sent to the first device, a task of extracting the feature values of the pictures in the comparison library is assigned to the first device, and the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model is acquired; after the feature value extraction results of all the pictures in the comparison library are obtained, the intelligent recognition algorithm model is sent to the second device, and the feature value extraction results of all the pictures in the comparison library are sent to all the intelligent recognition devices to be upgraded. It solves the technical problem in the related art that when an intelligent recognition device upgrades an algorithm model, the intelligent recognition function of the intelligent recognition device is usually manually turned off, and all the computing power of the intelligent recognition device is concentrated for the extraction work of new feature values of the intelligent recognition device, resulting in the inability to normally use the intelligent recognition function of the intelligent recognition device during the upgrade process of the algorithm model. It realizes the technical effect that the algorithm model of the intelligent recognition device can be upgraded without turning off the intelligent recognition function of the intelligent recognition device, and the intelligent recognition function of the intelligent recognition device is not affected during the upgrade process of the algorithm model of the intelligent recognition device.

[0019] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other feature values, objects, and advantages of the present application more concise and understandable. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments and descriptions thereof of the present application are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0021] Figure 1 is a flowchart of the intelligent recognition algorithm upgrade method of the device in this embodiment;

[0022] Figure 2 is a flowchart of the intelligent recognition algorithm upgrade method of the device in this preferred embodiment;

[0023] Figure 3 is a structural block diagram of the intelligent recognition algorithm upgrade device of the device in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] To understand the purpose, technical solution, and advantages of the present application more clearly, the present application will be described and explained below with reference to the drawings and embodiments.

[0025] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meanings understood by those with ordinary skills in the technical field to which this application belongs. In this application, words such as "a", "an", "one kind", "the", "these", etc. do not indicate a limitation in quantity, and they can be singular or plural. The terms "including", "containing", "having" and any variants thereof involved in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product or device including a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent in these processes, methods, products or devices. The terms "connected", "coupled", etc. involved in this application do not limit to physical or mechanical connections, but may include electrical connections, whether directly connected or indirectly connected. The term "a plurality of" involved in this application means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, and B exists alone. Usually, the character " / " indicates that the objects associated 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 sorting of the objects.

[0026] In this embodiment, a method for upgrading the intelligent recognition algorithm of a device is provided. Figure 1 It is a flowchart of the method for upgrading the intelligent recognition algorithm of the device in this embodiment, as Figure 1 shown, and this process includes the following steps:

[0027] Step S101: According to the capability information of the intelligent recognition device to be upgraded, the intelligent recognition device to be upgraded is divided into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold.

[0028] Step S102: Send the intelligent recognition algorithm model to be upgraded to the first device, assign the task of extracting the feature values of the pictures in the comparison library to the first device, and obtain the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model.

[0029] Step S103: After obtaining the feature values of all the pictures in the comparison library, send the algorithm model to the second device, and send the feature values of all the pictures in the comparison library to the first device and the second device.

[0030] First, determine whether the remaining capacity value of the intelligent recognition device to be upgraded is greater than a preset threshold, and classify the intelligent recognition device to be upgraded into two categories, namely, the first device with a remaining capacity value greater than the preset threshold and the second device with a remaining capacity value not greater than the preset threshold. Generally, it is the management platform that determines whether the remaining capacity value of the intelligent recognition device to be upgraded is greater than the preset threshold. The management platform can be a software management platform, such as DSS (Decision Support System), or a storage device with intelligent analysis capabilities, such as NVR (Network Video Recorder). The fact that the remaining capacity value of the first device is greater than the preset threshold means that the first device is in an idle state. Correspondingly, the fact that the remaining capacity value of the second device is not greater than the preset threshold means that the intelligent capacity of the second device is occupied. For example, the second device performs intelligent recognition functions based on the original un-upgraded algorithm model and the feature values of the pictures in the original comparison library.

[0031] Since the first device has the ability to extract the feature values of the pictures in the comparison library, send the intelligent recognition algorithm model to be upgraded to the first device. The first device executes the feature value extraction task according to the intelligent recognition algorithm model. After the first device finishes extracting the feature values of all the pictures in the comparison library, collect all the feature value extraction results. Send the intelligent recognition algorithm model to the second device and send all the feature value extraction results to the intelligent recognition device to be upgraded to complete the upgrade of the intelligent recognition device to be upgraded.

[0032] According to actual measurements, for example, using Hisilicon model 3559A as the computing power card, one computing power card can re-extract 22 feature value vectors per second based on pictures. If the number of pictures in the comparison library (such as the face library, license plate library, etc.) is 1 million, then in the actual usage scenario, it takes 45,000 seconds, that is, 12.5 hours. Due to the limited intelligent analysis ability of the device chip, when extracting the feature values of the pictures in the comparison library, a large amount of computing power is required, and it is necessary to turn off or reduce the analysis ability of the intelligent device, that is, turn off the intelligent recognition function of the intelligent recognition device and concentrate all the computing power of the chip on the extraction of feature values.

[0033] According to actual measurements, 1000 - 5000 feature values can be sent per second, and it takes about 200 - 1000 seconds to complete the sending of a 1-million comparison library.

[0034] Through the above steps, it is possible to first distinguish the first device and the second device, and let the first device in the idle state complete the extraction of the feature values of the comparison library. After the first device completes the extraction of the feature values, the algorithm model to be upgraded is sent to the second device, and then the feature values of all the pictures in the comparison library that are extracted are sent to all the intelligent recognition devices to be upgraded. Compared with the prior art, the upgrade task of the intelligent recognition device can be completed without turning off the intelligent recognition function of the intelligent recognition device, and the intelligent recognition function can be used normally during the upgrade process of the algorithm model of the intelligent recognition device. This solves the technical problem in the related art that when the intelligent recognition device upgrades the algorithm model, usually the intelligent recognition function of the intelligent recognition device is manually turned off, and all the computing power of the intelligent recognition device is concentrated to be used for the extraction of new feature values of the intelligent recognition device, resulting in the inability to use the intelligent recognition function normally during the upgrade process of the algorithm model of the intelligent recognition device. It realizes the technical effect that the algorithm model of the intelligent recognition device can be upgraded without turning off the intelligent recognition function of the intelligent recognition device, and the intelligent recognition function of the intelligent recognition device is not affected during the upgrade process of the algorithm model of the intelligent recognition device.

[0035] In some of these embodiments, when the intelligent recognition device to be upgraded is the first device, the method for upgrading the intelligent recognition algorithm of the device further includes: sending the intelligent recognition algorithm model to be upgraded to the first device, and obtaining the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model; after obtaining the feature value extraction results of all the pictures in the comparison library, sending the feature value extraction results of all the pictures in the comparison library to the first device.

[0036] If it is determined that all the intelligent recognition devices to be upgraded are the first device, only need to send the algorithm model to be upgraded to the first device. When the first device finishes extracting all the feature values of the pictures in the comparison library according to the algorithm model, send the feature values of all the pictures in the comparison library to the first device, and the upgrade work of the intelligent recognition device to be upgraded is completed. Compared with the prior art in this process, the upgrade operation of the intelligent recognition device to be upgraded can be completed without turning off the recognition function of the intelligent recognition device to be upgraded, effectively avoiding the situation that the intelligent recognition device cannot be used during the upgrade period due to upgrading the intelligent recognition device.

[0037] In some of these embodiments, in the case where there are multiple first devices, the method for upgrading the intelligent recognition algorithm of the device further includes: respectively obtaining the remaining capacity values of each first device; and allocating the feature value extraction task amount of the pictures in the comparison library and the range of the pictures in the comparison library corresponding to the task amount to each first device according to the remaining capacity values.

[0038] If the ability values of each first device to extract feature values are the same, the task of extracting the feature values of the pictures in the comparison library can be evenly distributed to each first device, which can shorten the extraction time of the feature values of the pictures in the comparison library.

[0039] However, if the ability values of each first device to extract feature values are different, the task of extracting the feature values of the pictures in the comparison library needs to be allocated according to the size of the ability value of the first device. The greater the ability value of the first device, the greater the task of extracting the feature values of the pictures in the comparison library allocated to the first device. For example, there are two first devices, E and F, and the ratio of the ability values of device E and device F to extract the feature values of the pictures in the comparison library is 6:4. For instance, if the total number of pictures in the comparison library is 1 million, the pictures in the comparison library are divided into two parts: the first 600,000 pictures and the last 400,000 pictures. The task of extracting the feature values of the first 600,000 pictures in the comparison library is allocated to device E, and the task of extracting the feature values of the last 400,000 pictures in the comparison library is allocated to device F. Compared with the related art, the extraction time of the feature values can be greatly shortened.

[0040] In some of these embodiments, when all the intelligent recognition devices to be upgraded are second devices, the method for upgrading the intelligent recognition algorithm of the devices further includes: adjusting the ability allocation of at least one second device so that at least one second device becomes a first device with a remaining ability value greater than a preset threshold. Adjusting the ability allocation of at least one second device includes: turning off at least one intelligent function of at least one second device.

[0041] If all the intelligent recognition devices to be upgraded are second devices, it is necessary to manually turn off the intelligent functions of at least one second device, so that the second device after turning off the intelligent functions has the ability to extract the feature values of the pictures in the comparison library, and then the operation of extracting the feature values of the pictures in the comparison library can be completed. Compared with the prior art, only by turning off some of the intelligent recognition devices to be upgraded, the operation of extracting the feature values of the pictures in the comparison library can be completed, effectively avoiding the problem that the intelligent recognition devices cannot be used for a long time during the upgrade of the intelligent recognition devices.

[0042] In some of these embodiments, according to the ability information of the intelligent recognition devices to be upgraded, before dividing the intelligent recognition devices to be upgraded into first devices with a remaining ability value greater than a preset threshold and second devices with a remaining ability value not greater than a preset threshold, the method further includes: determining whether the management platform of the intelligent recognition devices to be upgraded has the ability to execute the feature value extraction task according to the intelligent recognition algorithm model; when the management platform has the ability to execute the feature value extraction task according to the intelligent recognition algorithm model, the management platform executes at least a part of the task of extracting the feature values of the pictures in the comparison library according to the algorithm model.

[0043] When the management platform is a software management platform, such as DSS, first determine whether the DSS software management platform has the ability to extract the feature values of the pictures in the comparison library. When the DSS software management platform has the ability to extract the feature values of the pictures in the comparison library, the task of extracting the feature values of the pictures in the comparison library is completed on the software management platform, and the algorithm model to be upgraded and all the feature values of the pictures in the comparison library are sent to the intelligent recognition device to be upgraded, completing the upgrade task of the intelligent recognition device to be upgraded. Compared with the related technology, completing the task of extracting the feature values of the pictures in the comparison library on the management platform can greatly shorten the upgrade time of the intelligent recognition device to be upgraded, and can effectively avoid the situation that the intelligent recognition function of the intelligent recognition device is turned off due to the upgrade of the intelligent recognition device.

[0044] In some of these embodiments, the method for upgrading the intelligent recognition algorithm of the device further includes: determining a comparison library and an intelligent recognition device to be upgraded according to the algorithm model, where the algorithm model corresponds to the comparison library, and the comparison library corresponds to the intelligent recognition device to be upgraded.

[0045] The following describes and illustrates this embodiment through preferred embodiments. In this embodiment, the management platform takes DSS as an example. The management platform DSS is connected to 3 face recognition turnstile devices, which are respectively named device A, device B, and device C.

[0046] Figure 2 is the flowchart of the method for upgrading the intelligent recognition algorithm of the device in this preferred embodiment, as Figure 2 shown, and this process includes the following steps:

[0047] Step S201: The management platform obtains the algorithm model to be upgraded and proceeds to step S202.

[0048] Step S202: The management platform searches for an intelligent recognition device that is consistent with the algorithm model of the management platform and also consistent with the picture content in the comparison library of the management platform, and proceeds to step S203.

[0049] First, search for the records of the original algorithm model and the original feature values sent by the management platform before the upgrade on the management platform, and find the corresponding intelligent recognition device. In this embodiment, device A, device B, and device C are taken as examples. Because in a set of intelligent recognition systems, the relevant algorithm models and the pictures in the comparison library must be consistent in order to transmit the same set of feature values between devices, improve the efficiency of picture comparison, and at the same time, for a complete scenario, it is often necessary to deploy the same comparison library to summarize the relevant recognition data.

[0050] Step S203: Determine whether the management platform has the ability to extract the feature values of the pictures in the comparison library. When the management platform has the ability to extract the feature values of the pictures in the comparison library, proceed to Step S204; when the management platform does not have the ability to extract the feature values of the pictures in the comparison library, proceed to Step S206.

[0051] Step S204: The management platform completes the operation of extracting the feature values of the pictures in the comparison library according to the algorithm model to be upgraded on the management platform, and then proceeds to Step S205.

[0052] Step S205: The management platform distributes the algorithm model to Devices A, B, and C, and then distributes the feature values extracted in Step S204 by the platform to Devices A, B, and C to complete the upgrade operation of the algorithm models of Devices A, B, and C, and then proceeds to Step S206.

[0053] In this embodiment, the upgrade operation of the algorithm model of the intelligent recognition device can be completed without closing the recognition function of the intelligent recognition device. When upgrading the algorithm model, the use of the intelligent recognition device is not affected.

[0054] Step S206: The management platform obtains the ability of Devices A, B, and C to extract the feature values of the pictures in the comparison library, and proceeds to Step S207.

[0055] Step S207: Determine whether Devices A, B, and C all have the ability to extract the feature values of the pictures in the comparison library. When Devices A, B, and C all have the ability to extract the feature values of the pictures in the comparison library, proceed to Step S208; when Devices A, B, and C do not all have the ability to extract the feature values of the pictures in the comparison library, proceed to Step S211.

[0056] Step S208: The management platform distributes the algorithm model to be upgraded to Devices A, B, and C, and proceeds to Step S209.

[0057] Step S209: The management platform obtains the ability values of Devices A, B, and C to extract feature values, allocates the task volume of extracting the feature values of the pictures in the comparison library according to the ability values of extracting feature values, and allocates the corresponding range of pictures in the comparison library according to the task volume, and then proceeds to Step S210.

[0058] For example, if the ratio of the ability values of Devices A, B, and C to extract the feature values of the pictures in the comparison library is 4:4:2, and the size of the comparison library that needs to be re-extracted is 1 million, then the platform distributes the serial number intervals of Devices A, B, and C to 1-400000, 400001-800000, and 800001-1000000 respectively. Devices A, B, and C extract feature values according to the pictures in the comparison library and the corresponding serial number intervals.

[0059] Step S210: The management platform obtains all the feature values of the pictures in the comparison library extracted by Device A, Device B, and Device C, and proceeds to Step S217.

[0060] Step S211: Determine whether any of Device A, Device B, and Device C has the ability to extract the feature values of the pictures in the comparison library. When any of Device A, Device B, and Device C has the ability to extract the feature values of the pictures in the comparison library, proceed to Step S213; when none of Device A, Device B, and Device C has the ability to extract the feature values of the pictures in the comparison library, proceed to Step S212.

[0061] Step S212: Adjust the ability allocation of at least one of Device A, Device B, and Device C so that at least one of Device A, Device B, and Device C becomes a device with the ability to extract the feature values of the pictures in the comparison library, and proceed to Step S213.

[0062] Step S213: The management platform identifies the devices among Device A, Device B, and Device C that have the ability to extract the feature values of the pictures in the comparison library, and the management platform sends the algorithm model to be upgraded to the devices that have the ability to extract the feature values of the pictures in the comparison library.

[0063] In this embodiment, taking Device A and Device B having the ability to extract the feature values of the pictures in the comparison library and Device C not having the ability to extract the feature values of the pictures in the comparison library as an example, the management platform sends the algorithm model to be upgraded to Device A and Device B.

[0064] Step S214: The management platform obtains the ability values of Device A and Device B to extract feature values, allocates the task volume of extracting the feature values of the pictures in the comparison library according to the ability values of extracting feature values, and allocates the corresponding range of pictures in the comparison library according to the task volume, and proceeds to Step S215.

[0065] For example, if the ratio of the ability values of Device A and Device B to extract the feature values of the pictures in the comparison library is 8:2 and the size of the comparison library that needs to be re-extracted is 1 million, the platform sends the serial number ranges of Device A and Device B as 1 - 800000 and 800001 - 1000000 respectively. Device A and Device B extract feature values according to the pictures in the comparison library and the corresponding serial number ranges.

[0066] Step S215: The management platform obtains all the feature values of the pictures in the comparison library extracted by Device A and Device B, and proceeds to Step S216.

[0067] Step S216: The management platform sends the algorithm model to be upgraded to Device C.

[0068] Step S217: The management platform sends all the eigenvalue of the pictures in the comparison library to Device A, Device B, and Device C.

[0069] In this embodiment, when upgrading the components of the intelligent recognition device, it is not necessary to turn off or completely turn off the intelligent recognition function of the intelligent recognition device to be upgraded, and the upgrade operation of the intelligent recognition device to be upgraded can be completed, effectively avoiding the situation that the intelligent recognition function of the intelligent recognition device cannot be used during the upgrade of the intelligent recognition device.

[0070] In this embodiment, there is also provided an intelligent recognition algorithm upgrade device for a device, which is used to implement the above-mentioned embodiments and preferred implementation manners, and those that have been described will not be repeated here. The following terms such as "module", "unit", "sub-unit", etc. can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0071] Figure 3 is the structural block diagram of the intelligent recognition algorithm upgrade device for the device in this embodiment, as Figure 3 shown, the device includes: an identification module 10, a first distribution module 20, a first acquisition module 30, a second acquisition module 40, and a second distribution module 50.

[0072] The identification module 10 is configured to classify the intelligent recognition device to be upgraded into a first device with a remaining capacity value greater than a preset threshold and a second device with a remaining capacity value not greater than the preset threshold according to the capacity information of the intelligent recognition device to be upgraded.

[0073] The first distribution module 20 is configured to send the algorithm model to be upgraded to the first device.

[0074] The first acquisition module 30: is configured to acquire the eigenvalue extraction result obtained by the first device performing the eigenvalue extraction task according to the intelligent recognition algorithm model.

[0075] The second acquisition module 40 is configured to acquire the eigenvalues of all the pictures in the comparison library.

[0076] The second distribution module 50 is configured to send the algorithm model to the second device, and send the eigenvalues of all the pictures in the comparison library to the first device and the second device.

[0077] It should be noted that the above-mentioned each module can be a functional module or a program module, and can be implemented either by software or by hardware. For the modules implemented by hardware, the above-mentioned each module can be located in the same processor; or the above-mentioned each module can also be located in different processors in any combination form.

[0078] In this embodiment, an algorithm model upgrade system is further provided, including: a terminal device, a transmission device, and a server device; wherein, the terminal device is connected to the server device through the transmission device;

[0079] The terminal device is used to configure a combined service file for algorithm model upgrade;

[0080] The transmission device is used to transmit the combined service file for configuring algorithm model upgrade;

[0081] The server device is used to execute the steps in any of the above method embodiments.

[0082] In this embodiment, an electronic device is further provided, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any of the above method embodiments.

[0083] Optionally, the above electronic device may further include a transmission device and input / output devices, wherein the transmission device is connected to the above processor, and the input / output devices are connected to the above processor.

[0084] Optionally, in this embodiment, the above processor may be configured to execute the following steps through a computer program:

[0085] S1. According to the capability information of the intelligent recognition device to be upgraded, divide the intelligent recognition device to be upgraded into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold;

[0086] S2. Send the intelligent recognition algorithm model to be upgraded to the first device, assign the task of extracting the feature values of the pictures in the comparison library to the first device, and obtain the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model;

[0087] S3. After obtaining the feature value extraction results of all the pictures in the comparison library, send the intelligent recognition algorithm model to the second device, and send the feature value extraction results of all the pictures in the comparison library to all the intelligent recognition devices to be upgraded.

[0088] It should be noted that specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation manners, and will not be elaborated in this embodiment.

[0089] In addition, in combination with the method for upgrading the intelligent recognition algorithm of the device provided in the above embodiments, a storage medium can also be provided in this embodiment to implement it. A computer program is stored on the storage medium; when the computer program is executed by a processor, the method for upgrading the intelligent recognition algorithm of any one of the devices in the above embodiments is implemented.

[0090] It should be understood that the specific embodiments described here are only used to explain this application, rather than to limit it. According to the embodiments provided in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the protection scope of the present application.

[0091] Obviously, the drawings are only some examples or embodiments of the present application. For those of ordinary skill in the art, the present application can also be applied to other similar situations according to these drawings without creative work. In addition, it can be understood that although the work done during the development process may be complex and time-consuming, for those of ordinary skill in the art, certain design, manufacturing, or production changes based on the technical content disclosed in the present application are only conventional technical means and should not be regarded as insufficient disclosure of the present application.

[0092] The term "embodiment" in this application means that the specific characteristic values, structures, or characteristics described in combination with the embodiments may be included in at least one embodiment of the present application. The phrase appears in various positions in the specification does not necessarily mean the same embodiment, nor does it mean being independent or alternative to other embodiments and mutually exclusive. Those of ordinary skill in the art can clearly or implicitly understand that the embodiments described in this application can be combined with other embodiments without conflict.

[0093] The above embodiments only represent several implementation manners of the present application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of patent protection. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for upgrading the intelligent recognition algorithm of a device, characterized in that, Including: According to the capability information of the intelligent recognition device to be upgraded, the intelligent recognition device to be upgraded is divided into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold, and there are multiple first devices; Send the intelligent recognition algorithm model to be upgraded to the first device; Obtain the remaining capability value of each first device respectively, allocate the task volume of extracting the feature values of the pictures in the comparison library and the range of the pictures in the comparison library corresponding to the task volume according to the remaining capability value, and obtain the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model; After obtaining the feature value extraction results of all the pictures in the comparison library, send the intelligent recognition algorithm model to the second device, and send the feature value extraction results of all the pictures in the comparison library to all the intelligent recognition devices to be upgraded.

2. The method for upgrading the intelligent recognition algorithm of the device according to claim 1, characterized in that When all the intelligent recognition devices to be upgraded are the second devices, the method further includes: Adjust the capability allocation of at least one of the second devices so that at least one of the second devices becomes the first device with a remaining capability value greater than the preset threshold.

3. The method for upgrading the intelligent recognition algorithm of the device according to claim 2, characterized in that, Adjusting the capability allocation of at least one of the second devices includes: Turn off at least one intelligent function of at least one of the second devices.

4. The method for upgrading the intelligent recognition algorithm of the device according to claim 1, characterized in that, According to the capability information of the intelligent recognition device to be upgraded, before dividing the intelligent recognition device to be upgraded into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold, the method further includes: Judge whether the management platform of the intelligent recognition device to be upgraded has the ability to execute the feature value extraction task according to the intelligent recognition algorithm model; When the management platform has the ability to execute the feature value extraction task according to the intelligent recognition algorithm model, the management platform executes at least a part of the feature value extraction task of the pictures in the comparison library according to the algorithm model.

5. The intelligent recognition algorithm upgrade method for the device according to claim 1, characterized in that The method further includes: Determine the comparison library and the intelligent recognition device to be upgraded according to the algorithm model, wherein the algorithm model corresponds to the comparison library, and the comparison library corresponds to the intelligent recognition device to be upgraded.

6. An intelligent recognition algorithm upgrade device for a device, characterized in that Including: An identification module, a first distribution module, a first acquisition module, a second acquisition module and a second distribution module; The identification module is used to divide the intelligent recognition device to be upgraded into a first device with a remaining capability value greater than a preset threshold and a second device with a remaining capability value not greater than the preset threshold according to the capability information of the intelligent recognition device to be upgraded, and there are multiple first devices; The first distribution module is used to send the algorithm model to be upgraded to the first device; The first acquisition module: is used to obtain the remaining capability value of each first device respectively, allocate the task volume of extracting the feature values of the pictures in the comparison library and the range of the pictures in the comparison library corresponding to the task volume according to the remaining capability value, and obtain the feature value extraction result obtained by the first device executing the feature value extraction task according to the intelligent recognition algorithm model; The second acquisition module is configured to acquire the feature values of all images in the comparison library; The second distribution module is configured to send the algorithm model to the second device, and send the feature values of all images in the comparison library to the first device and the second device.

7. An electronic device, comprising a memory and a processor, characterized in that, A computer program is stored in the memory, and the processor is configured to run the computer program to execute the intelligent recognition algorithm upgrade method of the device according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the intelligent recognition algorithm upgrade method of the device according to any one of claims 1 to 5 are implemented.

Images