Intelligent analysis equipment, intelligent analysis method and invigilation all-in-one machine

Abstract

The invention provides intelligent analysis equipment, an intelligent analysis method and an invigilation all-in-one machine. In one example of the invention, the intelligent analysis equipment designs an associated basic library for the embedded equipment, the basic library can be integrated into the analysis service equipment, and the embedded equipment can control the analysis service equipment through the basic library, so that the analysis service equipment can be controlled by the intelligent analysis equipment. And the embedded equipment with relatively low performance can realize expansion of intelligent analysis computing power by controlling the analysis service equipment with relatively high performance.

Description

Technical Field

[0001] The present application relates to the field of edge intelligent analysis technology, and in particular to an intelligent analysis device, an intelligent analysis method and an all-in-one proctor. Background Art

[0002] An embedded device is a computer system with built-in specific functions, which is usually embedded in other devices to perform specific control, monitoring or data processing tasks.

[0003] At present, in the field of intelligent analysis, embedded devices are mainly used for front-end data collection so that analysis service equipment can perform relevant intelligent analysis based on the data collected by the embedded devices.

[0004] For example, assuming that the embedded device is a smart camera (such as an IPC (network camera)), the smart camera can collect image data and send the collected image data to the analysis service device, and the analysis service device performs intelligent analysis based on the image data collected by the smart camera. Summary of the invention

[0005] In view of this, the present application provides an intelligent analysis device, an intelligent analysis method and an all-in-one proctor.

[0006] Specifically, the present application is implemented through the following technical solutions: According to a first aspect of an embodiment of the present application, there is provided an intelligent analysis device, wherein the intelligent analysis device integrates and deploys an embedded device and an analysis service device; the analysis service device integrates a basic library associated with the embedded device, the basic library includes a function file compiled and generated according to the business functions between the embedded device and the analysis service device, the function file defines a calling interface corresponding to each business function, and the business function includes a data callback function and a parameter configuration function; The analysis service device is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The embedded device is used to search for the basic library by broadcasting in a startup state, and connect to the basic library when the basic library is found; The embedded device is further used to send the task parameters corresponding to the intelligent analysis to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library when the intelligent analysis instruction is detected; The analysis service device is further used to perform intelligent analysis and processing based on the task parameters using the currently loaded algorithm model; The embedded device is also used to obtain intelligent analysis results from the analysis service device through a calling interface corresponding to a data callback function in the basic library.

[0007] According to a second aspect of an embodiment of the present application, there is provided an intelligent analysis method, which is applied to an embedded device in the intelligent analysis device provided in the first aspect, and the method includes: When the embedded device is in a startup state, searching the basic library by broadcasting, and connecting with the basic library when the basic library is found; wherein the basic library is run and initialized by the analysis service device in a startup state; When an intelligent analysis instruction is detected, the task parameters corresponding to the intelligent analysis are sent to the analysis service device through the calling interface corresponding to the algorithm operation parameter configuration function in the basic library, so that the analysis service device performs intelligent analysis processing based on the task parameters and using the currently loaded algorithm model; The intelligent analysis result is obtained from the analysis service device through the calling interface corresponding to the data callback function in the basic library.

[0008] According to a third aspect of an embodiment of the present application, an all-in-one proctoring machine is provided, wherein a network camera and a network video recorder are integrated and deployed, and the network video recorder is integrated with a basic library associated with the network camera, and the basic library includes a function file compiled and generated according to the business functions between the network camera and the network video recorder, and the function file defines a calling interface corresponding to each business function, and the business function includes a data callback function, a parameter configuration function, and a task execution result delivery function; wherein: The network video recorder is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The network camera is used to search the basic library by broadcasting in the startup state, and connect to the basic library when the basic library is found; The network camera is further used to disassemble the intelligent analysis service to obtain the third analysis task and the fourth analysis task when the intelligent analysis instruction is detected; The network camera is further used to perform the third analysis task according to the task parameters corresponding to the third analysis task; and send the task parameters corresponding to the fourth analysis task to the network video recorder through the calling interface corresponding to the parameter configuration function in the basic library; The network video recorder is further used to perform intelligent analysis and processing using the currently loaded algorithm model according to the task parameters corresponding to the fourth analysis task; The network camera is further used to send the execution result of the third analysis task to the network video recorder through the calling interface corresponding to the task execution result sending function in the basic library; The network camera is further used to integrate the execution result of the third analysis task with the execution result of the third analysis task to obtain an intelligent analysis result; The network camera is also used to obtain intelligent analysis results from the network video recorder through a calling interface corresponding to the data callback function in the basic library.

[0009] The intelligent analysis device of the embodiment of the present application is deployed by integrating an embedded device and an analysis service device; the analysis service device is integrated with a basic library associated with the embedded device, and the basic library includes a function file compiled and generated according to the business functions between the embedded device and the analysis service device, and the function file defines the call interface corresponding to each business function, and the business function includes a data callback function and a parameter configuration function. The analysis service device runs the basic library in the startup state, initializes the basic library, and loads the algorithm model in the analysis engine; the embedded device searches for the basic library in the startup state by broadcasting, and connects to the basic library when the basic library is searched, and when the intelligent analysis instruction is detected, the task parameters corresponding to the intelligent analysis business are sent to the analysis service device through the call interface corresponding to the parameter configuration function in the basic library, and the analysis service device performs intelligent analysis processing based on the task parameters and the currently loaded algorithm model, and then the embedded device can obtain the intelligent analysis results from the analysis service device through the call interface corresponding to the data callback function in the basic library. By designing an associated basic library for embedded devices, the basic library can be integrated into the analysis service device, and the embedded device can control the analysis service device through the basic library. Therefore, relatively low-performance embedded devices can achieve the expansion of intelligent analysis computing power by controlling relatively high-performance analysis service devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of the structure of an intelligent analysis device shown as an exemplary embodiment of the present application; Figure 2 A schematic diagram of the structure of an intelligent analysis device shown as an exemplary embodiment of the present application; Figure 3 This is a schematic diagram of the structure of an all-in-one proctoring device shown as an exemplary embodiment of the present application; Figure 4 A schematic diagram of a process of sending parameters and returning data in an all-in-one proctoring machine according to an exemplary embodiment of the present application; Figure 5This is a schematic diagram of interaction in a rope skipping test scenario in a physical education test, shown as an exemplary embodiment of the present application; Figure 6 A schematic diagram of a skipping rope test process for a physical education test is shown as an exemplary embodiment of the present application; Figure 7 The figure is a flowchart of an intelligent analysis method shown as an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0011] In order to enable those skilled in the art to better understand the technical solutions provided by the embodiments of the present application, some technical terms in the embodiments of the present application are explained below.

[0012] Super Brain Engine (also called computing engine or analysis engine): Super Brain’s intelligent hardware unit, which has computing power and runs algorithm models for algorithm analysis.

[0013] All-in-one invigilation machine: A machine used in school examinations, which integrates a display screen, smart camera, super brain, and switch, and can be flexibly deployed in any place for invigilation.

[0014] Smart camera: A camera that has algorithm analysis capabilities and can control the super brain engine through a basic library.

[0015] Algorithm model: It is installed in the super brain engine and uses the computing power of the engine to analyze the video data to obtain a model package of proctoring data.

[0016] DSP (Digital Signal Processor): The intelligent control and encoding and decoding module of UltraBrain, which configures the engine and algorithm operation parameters, as well as the pre-processing and post-processing of algorithm analysis.

[0017] APP (Application): The main control module of Ultrain, which is mainly responsible for the business scheduling of Ultrain's basic functions, parameter configuration and storage, interaction with DSP, adding cameras, etc.

[0018] See also Figure 1 , is a schematic diagram of the structure of an intelligent analysis device provided in an embodiment of the present application, such as Figure 1 As shown, the intelligent analysis device integrates and deploys an embedded device and an analysis service device; the analysis service device integrates a basic library associated with the embedded device, the basic library includes a function file compiled and generated according to the business functions between the embedded device and the analysis service device, the function file defines the calling interface corresponding to each business function, the business function includes a data callback function and a parameter configuration function; wherein: The analysis service device is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The embedded device is used to search for the basic library by broadcasting in the startup state, and connect with the basic library when the basic library is found; The embedded device is also used to send the task parameters corresponding to the intelligent analysis to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library when the intelligent analysis instruction is detected; The analysis service device is also used to perform intelligent analysis and processing based on the task parameters and using the currently loaded algorithm model; The embedded device is also used to obtain intelligent analysis results from the analysis service device through the calling interface corresponding to the data callback function in the basic library.

[0019] In the embodiments of the present application, in order to break through the one-way control and resource usage of the analysis service device over the embedded device in the traditional intelligent analysis field, an associated basic library can be designed for the embedded device, and the basic library can be integrated into the analysis service device. The embedded device can control the analysis service device through the basic library. Thus, a relatively low-performance embedded device can achieve the expansion of intelligent analysis computing power by controlling a relatively high-performance analysis service device, and can also expand other functions of the super brain.

[0020] Exemplarily, the basic library may include a function file compiled based on the business functions between the embedded device and the analysis service device. The function file defines a calling interface corresponding to each business function. The embedded device can control the analysis service device to implement relevant business functions by calling the relevant calling interface.

[0021] Exemplarily, the business function may include but is not limited to a data callback function and a parameter configuration function.

[0022] In the embodiment of the present application, the analysis service device can run the basic library in the startup state, initialize the basic library, and load the local current algorithm model in the analysis engine.

[0023] When the embedded device is in the startup state, it can search for the basic library by broadcasting, and connect with the basic library when the basic library is found.

[0024] Relevant personnel can initiate intelligent analysis tasks by inputting intelligent analysis instructions into the embedded device.

[0025] When an embedded device detects an intelligent analysis instruction, it can send the task parameters corresponding to the intelligent analysis to the analysis service device through the basic library.

[0026] Exemplarily, the task parameters corresponding to the intelligent analysis can be determined through the interface of the embedded device or the platform interface configuration.

[0027] For example, the task parameters may include, but are not limited to, one or more of the parameters such as analysis time, sensitivity, and analysis area.

[0028] Exemplarily, the embedded device may send the task parameters corresponding to the intelligent analysis to the analysis service device by calling the interface corresponding to the parameter configuration function in the basic library.

[0029] The analysis service device can perform intelligent analysis and processing based on the task parameters sent by the embedded device and using the currently loaded algorithm model.

[0030] Embedded devices can obtain intelligent analysis results from analysis service devices through basic libraries.

[0031] Exemplarily, the embedded device can obtain the intelligent analysis result from the analysis service device by calling the interface corresponding to the data callback function in the basic library.

[0032] Exemplarily, the embedded device may output the acquired intelligent analysis results in the form of voice, image and / or text, or the embedded device may report it to a designated device.

[0033] For example, embedded devices can display intelligent analysis results through display screens.

[0034] It should be noted that, in the embodiment of the present application, when the intelligent analysis device is started for the first time or the algorithm model of the analysis service device is updated, the algorithm operation parameters of the algorithm model can also be configured.

[0035] Exemplarily, the algorithm operation parameters may include the type, name, ID, and bound engine number of the algorithm.

[0036] Exemplarily, the algorithm operation parameters may be sent from the embedded device to the analysis service device through a calling interface corresponding to the parameter configuration function in the basic library; or, may be configured by the analysis service device.

[0037] In some embodiments, the analysis service device is further used to send basic information of the embedded device to the basic library, wherein the basic information includes one or more of an IP address, a port number, and a serial number; The embedded device is specifically used to send a search message including basic information of the device by broadcasting in a startup state, and the basic library returns a response message when receiving the search message and determining that the basic information included in the search message matches the basic information of the embedded device; The embedded device is further specifically used to connect with the basic library when receiving a response message returned by the basic library.

[0038] Exemplarily, the analysis service device may obtain basic information of the embedded device, and send the obtained basic information of the embedded device to the basic library.

[0039] For example, a user may configure basic information of an embedded device that is integrated and deployed with the analysis service device in the analysis service device.

[0040] Exemplarily, the basic information of the embedded device may include, but is not limited to, one or more of an IP address, a port number, and a serial number.

[0041] Exemplarily, when the embedded device is in a startup state, it can send a search message including basic information of the device in a broadcasting manner.

[0042] For example, the embedded device may broadcast the search message in the local area network, and the basic library in the same local area network as the embedded device may receive the search message.

[0043] When the basic library receives a search message sent by an embedded device, it can obtain basic information of the embedded device included in the search message, and compare the basic information of the embedded device included in the search message with the basic information of the embedded device sent by the analysis service device.

[0044] If the two match, the basic library can return a response message to the embedded device, and the embedded device establishes a connection with the embedded device based on the response message.

[0045] If the two do not match, the basic library may not respond to the search message.

[0046] In some embodiments, the business function also includes an algorithm model update function; The embedded device is also used to send a new algorithm model to the analysis service device through the call interface corresponding to the algorithm model update function in the basic library; The analysis service device is also used to update the currently loaded algorithm model to the received new algorithm model.

[0047] Exemplarily, the algorithm model in the analysis service device may also be updated via an embedded device.

[0048] Correspondingly, the above function file can also define the calling interface corresponding to the algorithm model update function.

[0049] The embedded device can send a new algorithm model to the analysis service device by calling the calling interface corresponding to the algorithm model update function.

[0050] When the analysis service device receives a new algorithm model sent by the embedded device, the algorithm model currently loaded in the analysis engine may be updated to the received new algorithm model.

[0051] For example, the analysis service device may delete the currently loaded algorithm model and load the received new algorithm model in the analysis engine.

[0052] In some embodiments, the embedded device is further used to, when the intelligent analysis instruction is detected, disassemble the intelligent analysis service to obtain a first analysis task and a second analysis task; The embedded device is further used to execute the first analysis task according to the task parameters corresponding to the first analysis task; and send the task parameters corresponding to the second analysis task to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library; The analysis service device is specifically used to perform intelligent analysis and processing based on the task parameters corresponding to the second analysis task and using the currently loaded algorithm model.

[0053] Exemplarily, in order to reduce the load of the analysis service device, tasks that can be performed by an embedded device can be performed by the embedded device.

[0054] Correspondingly, when the embedded device detects the intelligent analysis instruction, it disassembles the intelligent analysis service to obtain the first analysis task and the second analysis task.

[0055] Exemplarily, the embedded device can disassemble the intelligent analysis service according to the function of the device, and disassemble the intelligent analysis task into tasks that can be processed by the device (i.e., the first analysis task mentioned above) and tasks that need to be processed by the analysis service device (i.e., the second analysis task mentioned above).

[0056] Exemplarily, the embedded device can determine the function of the device based on the algorithm model running on the device.

[0057] The embedded device may determine, among the task parameters, task parameters corresponding to the first analysis task and task parameters corresponding to the second analysis task.

[0058] On the one hand, the embedded device can execute the first analysis task according to the task parameters corresponding to the first analysis task.

[0059] On the other hand, the embedded device can send the task parameters corresponding to the second analysis task to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library.

[0060] The analysis service device can perform intelligent analysis and processing based on the task parameters corresponding to the second analysis task and using the currently loaded algorithm model, that is, execute the second analysis task.

[0061] In one example, the above business function may also include a task execution result delivery function; The embedded device is further used to send the execution result of the first analysis task to the analysis service device through a calling interface corresponding to the task execution result sending function in the basic library; The analysis service device is also used to integrate the execution result of the first analysis task with the execution result of the second analysis task to obtain an intelligent analysis result.

[0062] Exemplarily, when the embedded device executes the first analysis task in the above manner and obtains the execution result of the first analysis task, the execution result of the first analysis task can be sent to the analysis service device through the calling interface corresponding to the task execution result sending function in the basic library.

[0063] When the analysis service device receives the execution result of the first analysis task and obtains the execution result of the second analysis task, the execution result of the first analysis task and the execution result of the second analysis task may be integrated to obtain an intelligent analysis result.

[0064] In some embodiments, the embedded device is an intelligent image acquisition device; the first analysis task is a face comparison task; The embedded device is further used to perform face detection on the collected image data, and compare the detected face information with the pre-recorded reference face information, and send the face comparison result to the analysis service device through the calling interface corresponding to the face information comparison result sending function in the basic library; wherein the face comparison result includes that the detected face information is consistent with the pre-recorded reference face information; or, the detected face information is inconsistent with the pre-recorded reference face information; The analysis service device is also used to bind the face comparison result with the execution result of the second analysis task to obtain an intelligent analysis result.

[0065] For example, when the intelligent analysis task is to collect personnel assessment statistics, face comparison is required during the assessment and statistics process for any personnel to avoid replacement assessment.

[0066] Accordingly, the embedded device can perform face detection on the collected image data, and compare the detected face information with the pre-recorded reference face information.

[0067] Exemplarily, the benchmark facial information is the facial information of the person currently taking the assessment.

[0068] Exemplarily, the benchmark facial information may be pre-imported into the embedded device, or may be collected by the embedded device before the assessment begins.

[0069] Exemplarily, the face comparison result may include that the detected face information is consistent with the pre-recorded reference face information, or that the detected face information is inconsistent with the pre-recorded reference face information.

[0070] When the analysis service device receives the face comparison result sent by the embedded device, it can bind the face comparison result with the execution result of the second analysis task to obtain an intelligent analysis result.

[0071] Embedded devices can obtain intelligent analysis results bound to face comparison results from analysis service devices through the data callback function of the basic library.

[0072] Exemplarily, for the face comparison result sent to the analysis service device, the embedded device may also determine a corresponding target ID based on the reference face information, and send the target ID to the analysis service device.

[0073] Exemplarily, a target ID uniquely identifies a reference face.

[0074] In some embodiments, the embedded device is an intelligent image acquisition device; The analysis service device is also used to obtain image data collected by the embedded device; The embedded device is further used to send the analysis channel identifier to the analysis server through the calling interface corresponding to the algorithm operation parameter configuration function in the basic library when the intelligent analysis instruction is detected; The analysis service device is specifically used to perform intelligent analysis and processing on the target image data based on the algorithm operation parameters and using the currently loaded algorithm model; wherein the target image data is the image data of the target channel of the embedded device; and the target channel is the channel corresponding to the analysis channel identifier.

[0075] For example, take the embedded device as an intelligent image acquisition device (such as IPC) as an example.

[0076] In the case where the embedded device is an intelligent image acquisition device, the analysis service device can obtain image data (capture data or video data) acquired by the embedded device, and perform intelligent analysis on the acquired image data acquired by the embedded device.

[0077] Considering that the image data of different channels of the intelligent image acquisition device may be different, the embedded device can also control the analysis service device to perform intelligent analysis and processing on the image data of a specific channel.

[0078] When an intelligent analysis instruction is detected, the embedded device can also send the analysis channel identifier to the analysis service device through the calling interface corresponding to the algorithm running parameter configuration function in the basic library, so that the analysis service device can determine the channel (which can be called the target channel) corresponding to the image data to be analyzed based on the analysis channel identifier.

[0079] Accordingly, the analysis service device can obtain the image data of the target channel of the embedded device based on the target channel information, and perform intelligent analysis and processing on the target image data based on the algorithm operation parameters issued by the embedded device and using the currently loaded algorithm model.

[0080] In order to enable those skilled in the art to better understand the technical solutions provided in the embodiments of the present application, the technical solutions provided in the embodiments of the present application are described below with reference to specific examples.

[0081] like Figure 2 As shown, the intelligent analysis device may include an intelligent image acquisition unit (taking an intelligent camera as an example) and an analysis service unit (also known as a super brain). The intelligent camera includes an intelligent analysis module, a network interaction module, and an image acquisition and processing module; the analysis service unit includes a scheduling module, a parameter management module, and an engine module (i.e., an analysis engine). The network interaction module of the smart camera can establish a connection with the basic library in the engine module in the super brain.

[0082] Through the basic library, you can call and control the engine module of Ultrain to perform data transmission, parameter configuration, etc.

[0083] The smart camera controls the resources of the super brain engine and expands the algorithm analysis performance by calling the basic library.

[0084] The scheduling module of Ultrain can obtain the code stream from the smart camera and send the obtained code stream to the engine module.

[0085] The parameter management module of Ultrain can configure the parameters of the algorithm model running in the engine module based on the algorithm operating parameters sent by the smart camera.

[0086] The engine module of UltraBrain can run basic libraries and algorithm models, and use the algorithm models to perform intelligent analysis on the image data of smart cameras to obtain analysis results, such as intelligently counting the number of rope skipping.

[0087] For example, the basic library is an interface library provided by smart camera developers (i.e., embedded device developers) that needs to be integrated into Ultrain and can be called by the smart camera. The interface library includes the calling interface of the business functions between the smart camera and Ultrain.

[0088] Exemplarily, the business functions may include but are not limited to: data callback function, algorithm operation parameter configuration function, algorithm model update function and face comparison result delivery function.

[0089] For example, the basic library can be a .so file that is coded and compiled by the smart camera R&D personnel according to the business functions. The basic library can be integrated into the upgrade package of the Super Brain. After the Super Brain device is started, the startup command of the basic library will be called to run the basic library in the engine module of the Super Brain device. After the smart camera establishes a connection with the Super Brain, the interface of the basic library can be called to implement the functional business.

[0090] It should be noted that the basic library can also be expanded according to application scenarios and super brain functions.

[0091] For example, all functions and hardware capabilities of UltraBrain can be controlled by embedded devices through the basic library, and the basic library content can be expanded according to actual usage scenarios to flexibly add required functions.

[0092] The technical solution provided in the embodiment of the present application is described below in conjunction with specific application examples.

[0093] Taking the smart sports examination scenario as an example, the intelligent analysis equipment can be an all-in-one invigilator.

[0094] like Figure 3 As shown in the figure, the invigilator is integrated with a smart camera ( Figure 3 Taking IPC as an example), Super Brain ( Figure 3 In this example, NVR (Network Video Recorder) is used as an example), and display screens (such as Figure 3 The display screen is used to display the configuration menu. Relevant personnel can configure the parameters of the smart camera or super brain through the display screen and send the configuration parameters to the smart camera.

[0095] It should be noted that, in actual applications, the monitoring machine may also include a GPS (Global Positioning System) for locating the monitoring machine; a speaker for audio output; a power board and a fan for powering and cooling the monitoring machine; and an interface board, etc., and its specific functional implementation will not be repeated here.

[0096] The smart camera sends the operation parameters to the basic library through the local area network in the proctor, and the basic library sends them to the engine module of the super brain. The engine module of the super brain receives the operation parameters and can use the algorithm model to perform intelligent analysis on the code stream of the smart camera based on the operation parameters, and return the results to the smart camera, which reports to the display screen / platform. The schematic diagram can be shown as follows: Figure 4shown.

[0097] For example, Figure 3 As shown, the smart camera and super brain are integrated and deployed in the all-in-one proctor. The smart camera controls the engine resources of the super brain to expand the analysis performance of the smart camera and realize multi-algorithm high-performance analysis.

[0098] For example, the display screen of the all-in-one invigilation machine can open external network ports and internal (camera + super brain) LAN connections, adopt the Android system, use a touch screen, and integrate a full set of physical examination business functions and camera maintenance functions.

[0099] End users can operate the display screen, and the analysis services of the camera and super brain can be initiated by the display screen.

[0100] The display screen sends commands such as algorithm analysis rule parameters, task start / stop commands, and device / algorithm package upgrades to the smart camera.

[0101] Smart cameras decompose services. For smart analysis services implemented by smart cameras themselves, they complete smart analysis and report the analysis results. For intelligent analysis services that require super brain to implement, the smart camera will send task parameters and commands to the super brain, triggering the super brain to complete the analysis. The super brain will report the analysis results to the smart camera, which will then report them to the display screen or superimpose them in the code stream, and the display screen will display them or report them to an external platform.

[0102] When the task is completed, the end command triggered by the display screen is sent to the smart camera in turn, and the smart camera can send the end command to the super brain through the basic library.

[0103] For example, taking the rope skipping test, its business can be broken down into face matching and rope skipping counting; among them, face matching can be processed by a smart camera, and rope skipping counting can be processed by a super brain.

[0104] See also Figure 5 and Figure 6 , the following is explained using the skipping rope test in the physical education exam as an example.

[0105] 1. The smart camera is added to the Ultra Brain in advance. The Ultra Brain can obtain the stream from the smart camera and obtain the basic information of the smart camera, such as IP address, port number, serial number, channel number, etc.

[0106] 2. The R&D personnel of smart cameras provide the basic library, and Ultrain's DSP integrates the basic library and supports the operation and calling of the basic library.

[0107] 3. When the Ultra Brain is started, the DSP can run the basic library in the engine module, load the algorithm model, and pass the basic information of the camera obtained through the APP to the basic library (for subsequent connection with the smart camera).

[0108] Exemplarily, the algorithm model runs in the engine module and can be used for rope skipping test analysis.

[0109] 4. When the algorithm model in the Super Brain needs to be updated, the smart camera can send the new algorithm model to the Super Brain by calling the basic library, and the Super Brain will replace the original algorithm model package with the new algorithm model.

[0110] Below is the interaction logic when starting the exam.

[0111] 5. When the examinee (such as a student) is ready, the smart camera sends a command to start the rope skipping test to Super Brain through the basic library, and sends the operating parameters (or rule parameters) of the rope skipping test algorithm model to Super Brain.

[0112] Exemplarily, when the designated information entered by the examinee is detected and the examinee is detected in the designated examination area, it can be determined that the examinee is ready.

[0113] 6. When the super brain receives the command to start the test and the operating parameters, it configures the algorithm model according to the operating parameters and starts counting the number of skipping ropes.

[0114] 7. The smart camera will perform facial comparison on the examinee at the beginning and during the examination, and send the examinee's comparison results to the Super Brain through the basic library. The Super Brain parses the facial comparison information sent by the smart camera, binds the rope skipping count information with the facial comparison information, and assembles the message and returns it to the smart camera. After receiving the test results, target frame, area and other information, the smart camera will superimpose the rope skipping count and target frame into the code stream for display, and return the overall information to the upper-level device for display.

[0115] See also Figure 7 , is a flow chart of an intelligent analysis method provided in an embodiment of the present application, wherein the intelligent analysis method can be applied to an embedded device in the intelligent analysis device described in the above embodiment, such as Figure 7 As shown, the intelligent analysis method may include the following steps: Step S700: When the embedded device is in the startup state, search for the basic library by broadcasting, and connect to the basic library when the basic library is found; wherein the basic library is run and initialized by the analysis service device in the startup state.

[0116] Step S710: When an intelligent analysis instruction is detected, the task parameters corresponding to the intelligent analysis are sent to the analysis service device through the calling interface corresponding to the algorithm operation parameter configuration function in the basic library, so that the analysis service device performs intelligent analysis processing based on the task parameters and using the currently loaded algorithm model.

[0117] Step S720: Obtain intelligent analysis results from the analysis service device through the calling interface corresponding to the data callback function in the basic library.

[0118] In the embodiments of the present application, in order to break through the one-way control and resource usage of the analysis service device over the embedded device in the traditional intelligent analysis field, an associated basic library can be designed for the embedded device, and the basic library can be integrated into the analysis service device. The embedded device can control the analysis service device through the basic library. Thus, a relatively low-performance embedded device can achieve the expansion of intelligent analysis examples by controlling a relatively high-performance analysis service device, and can also expand other functions of the super brain.

[0119] Exemplarily, the basic library may include a function file compiled based on the business functions between the embedded device and the analysis service device. The function file defines a calling interface corresponding to each business function. The embedded device can control the analysis service device to implement relevant business functions by calling the relevant calling interface.

[0120] Exemplarily, the business function may include but is not limited to a data callback function and a parameter configuration function.

[0121] When the analysis service device is in the startup state, it can run the basic library, initialize the basic library, and record the local current algorithm model in the calculation engine.

[0122] When the embedded device is in the startup state, it can search for the basic library by broadcasting, and connect with the basic library when the basic library is found.

[0123] Relevant personnel can initiate intelligent analysis tasks by inputting intelligent analysis instructions into the embedded device.

[0124] When an embedded device detects an intelligent analysis instruction, it can send the task parameters corresponding to the intelligent analysis to the analysis service device through the basic library.

[0125] Exemplarily, the task parameters corresponding to the intelligent analysis can be determined through the interface of the embedded device or the platform interface configuration.

[0126] For example, the task parameters may include, but are not limited to, one or more of the parameters such as analysis time, sensitivity, and analysis area.

[0127] Exemplarily, the embedded device may send the task parameters corresponding to the intelligent analysis to the analysis service device by calling the interface corresponding to the parameter configuration function in the basic library.

[0128] The analysis service device can perform intelligent analysis and processing based on the task parameters sent by the embedded device and using the currently loaded algorithm model.

[0129] Embedded devices can obtain intelligent analysis results from analysis service devices through basic libraries.

[0130] Exemplarily, the embedded device can obtain the intelligent analysis result from the analysis service device by calling the interface corresponding to the data callback function in the basic library.

[0131] Exemplarily, the embedded device may output the acquired intelligent analysis results in the form of voice, image and / or text, or the embedded device may report it to a designated device.

[0132] For example, embedded devices can display intelligent analysis results through display screens.

[0133] For example, the specific processing flow of the embedded device implementing intelligent analysis can refer to the relevant description in the above embodiments.

[0134] The embodiment of the present application also provides an all-in-one proctoring machine, which integrates and deploys a network camera and a network video recorder, wherein the network video recorder integrates a basic library associated with the network camera, wherein the basic library includes a function file compiled and generated according to the business functions between the network camera and the network video recorder, wherein the function file defines a calling interface corresponding to each business function, wherein the business functions include a data callback function, a parameter configuration function, and a task execution result delivery function; wherein: The network video recorder is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The network camera is used to search the basic library by broadcasting in the startup state, and connect to the basic library when the basic library is found; The network camera is further used to disassemble the intelligent analysis service to obtain the third analysis task and the fourth analysis task when the intelligent analysis instruction is detected; The network camera is further used to perform the third analysis task according to the task parameters corresponding to the third analysis task; and send the task parameters corresponding to the fourth analysis task to the network video recorder through the calling interface corresponding to the parameter configuration function in the basic library; The network video recorder is further used to perform intelligent analysis and processing using the currently loaded algorithm model according to the task parameters corresponding to the fourth analysis task; The network camera is further used to send the execution result of the third analysis task to the network video recorder through the calling interface corresponding to the task execution result sending function in the basic library; The network camera is further used to integrate the execution result of the third analysis task with the execution result of the third analysis task to obtain an intelligent analysis result; The network camera is also used to obtain intelligent analysis results from the network video recorder through a calling interface corresponding to the data callback function in the basic library.

[0135] For example, the specific processing flow of the network camera (IPC) and the network video recorder (NVR) to implement intelligent analysis and processing can refer to the relevant description in the above embodiments.

[0136] In one example, the third analysis task is a face matching task; and the fourth analysis task is a rope skipping counting task.

Claims

1. An intelligent analysis device, It is characterized in that The intelligent analysis device integrates and deploys an embedded device and an analysis service device; the analysis service device integrates a basic library associated with the embedded device, the basic library includes a function file compiled and generated according to the business functions between the embedded device and the analysis service device, the function file defines a calling interface corresponding to each business function, and the business function includes a data callback function and a parameter configuration function; wherein: The analysis service device is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The embedded device is used to search for the basic library by broadcasting in a startup state, and connect to the basic library when the basic library is found; The embedded device is further used to send the task parameters corresponding to the intelligent analysis service to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library when the intelligent analysis instruction is detected; The analysis service device is further used to perform intelligent analysis and processing based on the task parameters using the currently loaded algorithm model; The embedded device is also used to obtain intelligent analysis results from the analysis service device through a calling interface corresponding to a data callback function in the basic library.

2. The intelligent analysis device according to claim 1, It is characterized in that The analysis service device is further used to send basic information of the embedded device to the basic library, wherein the basic information includes one or more of an IP address, a port number, and a serial number; The embedded device is specifically configured to send a search message including basic information of the device by broadcasting in a startup state, and the basic library returns a response message when receiving the search message and determining that the basic information included in the search message matches the basic information of the embedded device; The embedded device is also specifically used to connect to the basic library when receiving a response message returned by the basic library.

3. The intelligent analysis device according to claim 1, It is characterized in that The business functions also include algorithm model update functions; The embedded device is further used to send a new algorithm model to the analysis service device through a calling interface corresponding to the algorithm model update function in the basic library; The analysis service device is also used to update the currently loaded algorithm model to the received new algorithm model.

4. The intelligent analysis device according to claim 1, It is characterized in that The embedded device is further used to disassemble the intelligent analysis service to obtain a first analysis task and a second analysis task when an intelligent analysis instruction is detected; The embedded device is further used to execute the first analysis task according to the task parameters corresponding to the first analysis task; and send the task parameters corresponding to the second analysis task to the analysis service device through the calling interface corresponding to the parameter configuration function in the basic library; The analysis service device is specifically used to perform intelligent analysis and processing based on the task parameters corresponding to the second analysis task and using the currently loaded algorithm model.

5. The intelligent analysis device according to claim 4, It is characterized in that The business functions also include a task execution result delivery function; The embedded device is further used to send the execution result of the first analysis task to the analysis service device through the calling interface corresponding to the task execution result sending function in the basic library; The analysis service device is further used to integrate the execution result of the first analysis task with the execution result of the second analysis task to obtain the intelligent analysis result.

6. The intelligent analysis device according to claim 5, It is characterized in that The embedded device is an intelligent image acquisition device; the first analysis task is a face comparison task; The embedded device is specifically used to perform face detection on the collected image data, and compare the detected face information with the pre-recorded reference face information, and send the face comparison result to the analysis service device through the calling interface corresponding to the face comparison result sending function in the basic library; wherein the face comparison result includes that the detected face information is consistent with the pre-recorded reference face information; or that the detected face information is inconsistent with the pre-recorded reference face information; The analysis service device is specifically used to bind the face comparison result with the execution result of the second analysis task to obtain an intelligent analysis result.

7. The intelligent analysis device according to claim 1, It is characterized in that The embedded device is an intelligent image acquisition device; The analysis service device is further used to obtain image data collected by the embedded device; The embedded device is further configured to send the analysis channel identifier to the analysis service device through a calling interface corresponding to a parameter configuration function in the basic library when an intelligent analysis instruction is detected; The analysis service device is specifically used to perform intelligent analysis and processing on the target image data based on the task parameters and using the currently loaded algorithm model; wherein the target image data is the image data of the target channel of the embedded device; and the target channel is the channel corresponding to the analysis channel identifier.

8. An intelligent analysis method, It is characterized in that The embedded device used in the intelligent analysis device according to any one of claims 1 to 7, the method comprising: When the embedded device is in a startup state, searching the basic library by broadcasting, and connecting with the basic library when the basic library is found; wherein the basic library is run and initialized by the analysis service device in a startup state; When an intelligent analysis instruction is detected, the task parameters corresponding to the intelligent analysis are sent to the analysis service device through the calling interface corresponding to the algorithm operation parameter configuration function in the basic library, so that the analysis service device performs intelligent analysis processing based on the task parameters and using the currently loaded algorithm model; The intelligent analysis result is obtained from the analysis service device through the calling interface corresponding to the data callback function in the basic library.

9. An all-in-one proctoring machine, It is characterized in that The invigilation machine integrates and deploys a network camera and a network video recorder, and the network video recorder integrates a basic library associated with the network camera, and the basic library includes a function file compiled and generated according to the business functions between the network camera and the network video recorder, and the function file defines a calling interface corresponding to each business function, and the business function includes a data callback function, a parameter configuration function, and a task execution result delivery function; wherein: The network video recorder is used to run the basic library in the startup state, initialize the basic library, and load the algorithm model in the analysis engine; The network camera is used to search the basic library by broadcasting in the startup state, and connect to the basic library when the basic library is found; The network camera is further used to disassemble the intelligent analysis service to obtain the third analysis task and the fourth analysis task when the intelligent analysis instruction is detected; The network camera is further used to perform the third analysis task according to the task parameters corresponding to the third analysis task; and send the task parameters corresponding to the fourth analysis task to the network video recorder through the calling interface corresponding to the parameter configuration function in the basic library; The network video recorder is further used to perform intelligent analysis and processing using the currently loaded algorithm model according to the task parameters corresponding to the fourth analysis task; The network camera is further used to send the execution result of the third analysis task to the network video recorder through the calling interface corresponding to the task execution result sending function in the basic library; The network camera is further used to integrate the execution result of the third analysis task with the execution result of the third analysis task to obtain an intelligent analysis result; The network camera is also used to obtain intelligent analysis results from the network video recorder through a calling interface corresponding to the data callback function in the basic library.

10. The invigilator according to claim 9, It is characterized in that The third analysis task is a face matching task; and the fourth analysis task is a rope skipping counting task.

Images