Equipment control method and device, equipment and storage medium

Abstract

The invention discloses an equipment control method and device, equipment and a storage medium, and the method comprises the steps: obtaining a plurality of target objects and the motion information of each target object from a plurality of frames of images detected by target equipment, predicting the number of frames of each target object in a detection region corresponding to the target equipment according to the motion information, obtaining complete process execution time consumed by the target device for processing a single frame of image, calculating predicted time consumed by the target device for processing each target object according to the existing frame number and the complete process execution time consumed, and calculating average time consumed by the target device for processing a single target object according to the predicted time consumed, a power consumption adjustment coefficient is calculated according to the average time consumption and the target time consumption, the target equipment is controlled according to the power consumption adjustment coefficient, and the target time consumption is the time consumption meeting the user requirement corresponding to the target equipment, so that the equipment power consumption can be adjusted on the premise of ensuring that the equipment performance can meet the user requirement. And the balance between equipment performance and equipment power consumption is realized.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a device control method, apparatus, device and storage medium. Background Art

[0002] At present, for low-power processing devices, the faster the task is processed, the better. Instead, the power consumption of the device should be reduced as much as possible while meeting the needs of users. However, the control method of related devices usually takes improving the processing performance of the processing device as the control target when controlling the processing device, resulting in high energy consumption of the device and imbalance between device performance and device power consumption. Summary of the invention

[0003] The main purpose of this application is to provide a device control method, apparatus, device and storage medium, aiming to solve the technical problem that the related device control method takes improving the processing performance of the processing device as the control goal, resulting in high device energy consumption and imbalance between device performance and device power consumption.

[0004] To achieve the above object, the present application provides a device control method, the device control method comprising:

[0005] Acquire multiple target objects and motion information of each target object from multiple frames of images detected by the target device;

[0006] Predicting the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information;

[0007] Obtaining the execution time of the complete process required for the target device to process a single frame of image;

[0008] Calculate the predicted time taken by the target device to process each target object according to the number of existing frames and the execution time taken by the complete process, and calculate the average time taken by the target device to process a single target object according to the predicted time taken;

[0009] A power consumption adjustment coefficient is calculated according to the average time consumption and the target time consumption, and the target device is controlled according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption required to meet the user demand corresponding to the target device.

[0010] Optionally, controlling the target device according to the power consumption adjustment coefficient includes:

[0011] Acquiring hardware parameters of each computing unit and / or execution strategy of each functional component, wherein the target device includes multiple computing units and multiple functional components;

[0012] The hardware parameters of each computing unit and / or the execution strategy of each functional component are adjusted according to the power consumption adjustment coefficient.

[0013] Optionally, the execution strategy of each functional component includes the calculation frequency of each functional component;

[0014] The step of adjusting the execution strategy of each functional component according to the power consumption adjustment coefficient includes:

[0015] When the power consumption adjustment coefficient is greater than a preset threshold, selecting a secondary functional component from a plurality of functional components according to the importance of each functional component to the target device, and reducing the calculation frequency of the secondary functional component;

[0016] When the power consumption adjustment coefficient is equal to a preset threshold, controlling each functional component to maintain the current calculation frequency;

[0017] When the power consumption adjustment coefficient is less than a preset threshold, a target functional component whose calculation frequency is lower than a preset normal value is selected from each functional component, and the calculation frequency of the target functional component is increased.

[0018] Optionally, the execution strategy of each functional component includes a working mode of each functional component;

[0019] The step of adjusting the execution strategy of each functional component according to the power consumption adjustment coefficient includes:

[0020] When the power consumption adjustment coefficient is greater than a preset threshold, the working mode of at least one functional component is switched to a high-performance mode, and the functional component switched to the high-performance mode is migrated to a first computing unit for execution, wherein the first computing unit is the computing unit with the strongest performance among the multiple computing units;

[0021] When the power consumption adjustment coefficient is equal to a preset threshold, controlling each functional component to maintain a current working mode;

[0022] When the power consumption adjustment coefficient is less than a preset threshold, the working mode of at least one functional component is switched to a low energy consumption mode, and the functional component switched to the low energy consumption mode is migrated to a second computing unit for execution, wherein the second computing unit is a computing unit with the lowest energy consumption among multiple computing units.

[0023] Optionally, adjusting the hardware parameters of each computing unit according to the power consumption adjustment coefficient includes:

[0024] Acquire current hardware parameters of each computing unit, and calculate parameter adjustment values ​​according to the power consumption adjustment coefficient and the current hardware parameters;

[0025] The current hardware parameters of each computing unit are adjusted according to the parameter adjustment value, wherein the current hardware parameters include a current computing frequency and / or a current voltage.

[0026] Optionally, predicting the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information includes:

[0027] Predicting a set of motion position points of each target object according to the motion information;

[0028] Predicting the existence time of each target object in the detection area corresponding to the target device according to the set of motion position points;

[0029] The number of frames in which each target object exists in the detection area corresponding to the target device is calculated based on the existence time and the number of image frames processed per second by the target device.

[0030] Optionally, the time consumption of executing the complete process required for obtaining the target device to process a single frame of image includes:

[0031] Obtaining the execution time consumed by each computing unit when executing each functional component to process a single frame of image, wherein the target device includes a plurality of computing units for heterogeneous computing;

[0032] The execution times are summed to obtain the complete process execution time required for the target device to process a single frame of image.

[0033] Optionally, calculating the power consumption adjustment coefficient according to the average time consumption and the target time consumption includes:

[0034] The power consumption adjustment coefficient is calculated according to the average time consumption and the target time consumption by using the adjustment coefficient calculation formula; the adjustment coefficient calculation formula is:

[0035]

[0036] Where k is the power consumption adjustment coefficient, T ave is the average time, T ue Time consuming for the goal.

[0037] In addition, to achieve the above-mentioned purpose, the present application also proposes a device control device, the device control device comprising:

[0038] An information acquisition module, used to acquire multiple target objects and motion information of each target object from multiple frames of images detected by the target device;

[0039] A frame number prediction module, used to predict the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information;

[0040] A time-consuming acquisition module is used to acquire the time-consuming execution of the complete process required for the target device to process a single frame of image;

[0041] A time calculation module, used to calculate the predicted time required for the target device to process each target object according to the number of existing frames and the execution time of the complete process, and calculate the average time required for the target device to process a single target object according to the predicted time;

[0042] The device control module is used to calculate the power consumption adjustment coefficient according to the average time consumption and the target time consumption, and control the target device according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption that meets the user demand corresponding to the target device.

[0043] In addition, to achieve the above-mentioned purpose, the present application also proposes a device control device, which includes a memory, a processor, and a device control program stored in the memory and executable on the processor, and the device control program is configured to implement the device control method as described above.

[0044] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, on which a device control program is stored, and when the device control program is executed by a processor, the device control method as described above is implemented.

[0045] One or more technical solutions proposed in this application have at least the following technical effects:

[0046] In the present application, a method is disclosed for obtaining multiple target objects and motion information of each target object from multiple frames of images detected by a target device, predicting the number of frames in which each target object exists in a detection area corresponding to the target device based on the motion information, obtaining the execution time of the complete process required for the target device to process a single frame of image, calculating the predicted time for the target device to process each target object based on the number of existing frames and the execution time of the complete process, and calculating the average time required for the target device to process a single target object based on the predicted time, calculating the power consumption adjustment coefficient based on the average time and the target time, and controlling the target device based on the power consumption adjustment coefficient, wherein the target time is the time required to meet the user requirements corresponding to the target device; since the present application analyzes multiple frames of images detected by the target device to predict the average time required for the target device to process a single target object, and controls the power consumption of the target device based on the average time and real-time requirements, it is possible to adjust the power consumption of the device while ensuring that the performance of the device can meet the user requirements, thereby achieving a balance between the performance of the device and the power consumption of the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0049] Figure 1 This is a flow chart of the first embodiment of the device control method of the present application;

[0050] Figure 2 This is a flow chart of the second embodiment of the device control method of the present application;

[0051] Figure 3 This is a schematic diagram of a specific flow chart of an embodiment of the device control method of the present application;

[0052] Figure 4 This is a flow chart of the third embodiment of the device control method of the present application;

[0053] Figure 5 A schematic diagram of a scenario of an embodiment of a device control method of the present application;

[0054] Figure 6 This is a schematic diagram of the module structure of the device control device of the embodiment of the present application;

[0055] Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the device control method in the embodiment of the present application.

[0056] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0057] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0058] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0059] In the field of intelligent analysis, post-processing functional components such as target detection, filtering, scoring, and recognition are often included. Among them, post-processing functional components such as scoring and recognition are highly correlated with the number of targets, especially AI intelligence. The complexity of different scenarios varies greatly, and the number of targets varies greatly. Therefore, different scenarios have different requirements for computing power, and the time consumption of different scenarios is also different. However, for users, when the performance reaches a certain value, the performance is improved, and the user experience is the same, and the actual value is not great. For example, in the attendance scenario based on feature recognition, the entire process can basically achieve a good experience effect in 200ms, and the experience of 100ms and 200ms is not much different. However, when controlling the processing equipment, the relevant equipment control method usually takes improving the processing performance of the processing equipment as the control target, resulting in high equipment energy consumption and imbalance between equipment performance and equipment power consumption.

[0060] Therefore, in order to overcome the above-mentioned defects, the present application provides a solution, which includes: obtaining multiple target objects and motion information of each target object from multiple frame images detected by the target device, predicting the number of frames in which each target object exists in the detection area corresponding to the target device based on the motion information, obtaining the execution time of the complete process required for the target device to process a single frame image, calculating the predicted time for the target device to process each target object based on the number of existing frames and the execution time of the complete process, and calculating the average time required for the target device to process a single target object based on the predicted time, calculating the power consumption adjustment coefficient based on the average time and the target time, and controlling the target device according to the power consumption adjustment coefficient, wherein the target time is the time required to meet the user needs corresponding to the target device.

[0061] Since the present application analyzes multiple frames of images detected by the target device to predict the average time required for the target device to calculate a single target object, and calculates the power consumption adjustment coefficient based on the average time and the target time, and controls the target device based on the power consumption adjustment coefficient, it is possible to adjust the device power consumption while ensuring that the device performance can meet user needs, thereby achieving a balance between device performance and device power consumption.

[0062] It should be noted that the execution subject of the embodiment of the present application can be a device control device with data processing, network communication and program running functions, such as a controller of a target device, or other electronic devices that can achieve the same or similar functions, and this embodiment does not impose any restrictions on this.

[0063] Based on this, the present application embodiment provides a device control method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the device control method of the present application.

[0064] In a first embodiment, the device control method includes:

[0065] Step S10: acquiring multiple target objects and motion information of each target object from multiple frames of images detected by the target device.

[0066] It should be understood that the application scenario of the embodiment of the present application can be to control the target device according to the power consumption adjustment coefficient in the intelligent analysis scenario. In a specific implementation, the intelligent analysis scenario can be an image analysis scenario (such as an attendance scenario based on feature recognition, etc.), and the target device can be an image acquisition device (such as an attendance machine with a camera, etc.), and the image acquisition device can continuously capture images to obtain a series of multi-frame images that are adjacent in time.

[0067] It is understandable that the target object can be a detection target of the target device. For example, in a feature recognition-based attendance scenario, the target object can be an employee. Of course, in other scenarios, the target object can also be set to other objects according to actual needs, such as vehicles, animals, specific items, etc., and this embodiment does not limit this. Acquiring multiple target objects from multiple frames of images detected by the target device can be to obtain multiple target objects from multiple frames detected by the target device through a target detection algorithm, wherein the target detection algorithm can be pre-set. In a specific implementation, the target detection algorithm can be based on a traditional computer vision technology algorithm (such as edge detection, template matching, etc.), or it can be based on a deep learning algorithm (such as CNN (Convolutional Neural Network, Convolutional Neural Network), YOLO, Faster R-CNN (FasterRegion-based Convolutional Neural Network, Fast Regional Convolutional Neural Network)), etc.), and this embodiment does not limit this.

[0068] It should be understood that the motion information of each target object can be a motion vector of each target object, which is used to represent the motion speed and motion direction of each target object. Obtaining the motion information of each target object from the multiple frames of images detected by the target device can be by analyzing the position change of each target object between the multiple frames of images to obtain the motion information such as the calculated speed and direction of each target object.

[0069] Furthermore, in order to reduce interference information and improve the accuracy of the target object, in this embodiment, obtaining multiple target objects from multiple frames of images detected by the target device may be obtaining multiple initial objects from the multiple frames of images detected by the target device, filtering out short-time and repeatedly appearing objects from the multiple initial objects, and obtaining the target object, wherein the short-time appearing object may refer to an object whose number of appearance frames is less than a preset frame number threshold, wherein the preset frame number threshold may be set in advance, and this embodiment does not impose any restrictions on this.

[0070] Step S20: predicting the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information.

[0071] It should be noted that the detection area corresponding to the target device may be an area set as the area to be detected in the image collected by the target device. For example, in an attendance scene based on feature recognition, the detection area may be an area that can be captured by the attendance machine.

[0072] It should be understood that predicting the number of frames in which each target object exists in the detection area corresponding to the target device based on motion information may be predicting the number of frames in which each target object exists in the detection area corresponding to the target device based on operation information through a preset prediction model, wherein the preset prediction model may be a prediction model pre-trained based on the extracted motion information samples using a machine learning or deep learning algorithm, and the goal of the preset prediction model may be to predict the probability of each target object being in the detection area in future frames based on the motion information of each target object, and calculate and predict the number of frames in which each target object exists in the detection area corresponding to the target device based on the probability.

[0073] Step S30: Obtaining the execution time of the complete process required for the target device to process a single frame of image.

[0074] It can be understood that the execution time of the complete process can be the execution time required for the target device to fully execute multiple functional components to process a single frame image, wherein, in the attendance scene based on feature recognition, multiple functional components may include a target detection component, an object filtering component, a target recognition component, an attribute recognition component, and a trigger event component, etc., wherein the target detection component is used to obtain multiple initial objects from the multiple frames of images detected by the target device; the object filtering component is used to filter the short-time and repeated objects in the multiple initial objects, obtain the target object, and obtain the operation information of the target object from the multiple frames of images; the target recognition component is used to identify the identity information of the target object to complete the attendance opening; the attribute recognition component is used to identify the attributes of the target object (such as age, gender, etc.); the trigger event component is used to perform specific operations when a specific event is triggered, for example, when it is detected that the target object is not wearing work clothes, an alarm is issued. In a specific implementation, the execution time of the complete process required for the target device to process a single frame image can be directly obtained from the target device, and this embodiment does not limit this.

[0075] Step S40: Calculate the predicted time required for the target device to process each target object based on the number of existing frames and the execution time required for the complete process, and calculate the average time required for the target device to process a single target object based on the predicted time.

[0076] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the number of frames in which the mth target object exists in the detection area corresponding to the target device is OTK. m , the time required for the target device to process a single frame of image is TE, then the average time required for the target device to process a single target object can be calculated by the following formula:

[0077]

[0078] Where, T ave Indicates the average time required for the target device to process a single target object. N is the total number of target objects. OTK m is the number of frames in which the mth target object exists in the detection area corresponding to the target device, and TE is the execution time of the complete process required for the target device to process a single frame of image. m *TE represents the estimated time consumed by the target device to process the mth target object.

[0079] Step S50: Calculate a power consumption adjustment coefficient according to the average time consumption and the target time consumption, and control the target device according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption required to meet the user demand corresponding to the target device.

[0080] It should be noted that the target time consumption can be the time consumption that meets the user's real-time requirements. The target time consumption can be preset by the user or obtained by analyzing the user's real-time requirements. This embodiment does not limit this. For example, in an attendance scenario based on feature recognition, if the entire process takes 200ms, it will not affect the user experience. However, if the entire process takes 300ms, some users may feel that the recognition is slow. Therefore, in this scenario, 200ms can be set as the target time consumption.

[0081] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the average time required for a target device to process a single target object is T ave , the target time is T ue , the power consumption adjustment coefficient can be calculated by the adjustment coefficient calculation formula, and the adjustment coefficient calculation formula is:

[0082]

[0083] In the formula, k represents the power consumption adjustment coefficient, T ave It represents the average time required for the target device to process a single target object, T ue Indicates the target time.

[0084] It is understandable that, according to the power consumption adjustment coefficient, the target device can be controlled by comparing the power consumption adjustment coefficient with the preset threshold value, and the target device can be controlled according to the comparison result. In a specific implementation, the preset threshold value can be set to 1. When k>1, T ave >T ue , the target device's performance is insufficient and needs to be improved; when k<1, T ave <T ue , there is redundancy in the device performance of the target device, and the device power consumption of the target device needs to be reduced; when k=1, the device performance and device power consumption are balanced, and there is no need to adjust the target device.

[0085] This embodiment analyzes multiple frames of images detected by the target device to predict the average time required for the target device to process a single target object, and controls the power consumption of the target device according to the average time and real-time requirements. This allows the device power consumption to be adjusted while ensuring that the device performance can meet user needs, thereby achieving a balance between device performance and device power consumption.

[0086] Reference Figure 2 , Figure 2 This is a flow chart of the second embodiment of the device control method of this application, based on the above Figure 1 The first embodiment shown is a second embodiment of the device control method of the present application.

[0087] In the second embodiment, the step S50 includes:

[0088] Step S501: Calculate the power consumption adjustment coefficient according to the average time consumption and the target time consumption.

[0089] It should be understood that in order to improve the device control effect, in this embodiment, the target device can be controlled from the hardware and / or software dimensions, that is, the hardware parameters of each computing unit and / or the execution strategy of each functional component are adjusted.

[0090] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the average time required for a target device to process a single target object is T ave , the target time is T ue , the power consumption adjustment coefficient can be calculated by the following formula:

[0091]

[0092] In the formula, k represents the power consumption adjustment coefficient, T ave It represents the average time required for the target device to process a single target object, T ue Indicates the target time.

[0093] Step S502: Acquire hardware parameters of each computing unit and / or execution strategy of each functional component, wherein the target device includes multiple computing units and multiple functional components.

[0094] It should be noted that the hardware parameters may be parameters that affect the device performance of the target device, such as frequency, voltage, etc., which are not limited in this embodiment; the execution strategies of each functional component may be execution strategies that affect the device performance of the target device, such as the calculation frequency and working mode of each functional component, which are not limited in this embodiment.

[0095] Step S503: adjusting the hardware parameters of each computing unit and / or the execution strategy of each functional component according to the power consumption adjustment coefficient.

[0096] It can be understood that the hardware parameters of each computing unit and / or the execution strategy of each functional component can be adjusted according to the power consumption adjustment coefficient by comparing the power consumption adjustment coefficient with a preset threshold, and adjusting the hardware parameters of each computing unit and / or the execution strategy of each functional component according to the comparison result.

[0097] In a specific implementation, the preset threshold may be set to 1. When k=1, the device performance and the device power consumption are balanced, and there is no need to adjust the target device.

[0098] When k>1, T ave >T ue , the device performance of the target device is insufficient, and the device performance of the target device needs to be improved. Therefore, the hardware parameters of each computing unit can be increased, or the execution strategy of each functional component can be adjusted to a high-performance mode, or a combination of the above two adjustment methods can be used to further increase the computing power, which is not limited in this embodiment.

[0099] When k<1, T ave <T ue , there is redundancy in device performance, and the device power consumption of the target device needs to be reduced. Therefore, the hardware parameters of each computing unit can be reduced, or the execution strategy of each functional component can be adjusted to a low-power mode, or a combination of the above two adjustment methods can be used to further reduce the device energy consumption, which is not limited in this embodiment.

[0100] This embodiment controls the target device from the hardware and / or software dimensions, that is, adjusts the hardware parameters of each computing unit and / or the execution strategy of each functional component, thereby improving the device control effect and further achieving a better balance between device performance and device power consumption.

[0101] Furthermore, the execution strategy of each functional component includes the calculation frequency of each functional component; the execution strategy of each functional component is adjusted according to the power consumption adjustment coefficient, including: when the power consumption adjustment coefficient is greater than a preset threshold, selecting a secondary functional component from multiple functional components according to the importance of each functional component to the target device, and reducing the calculation frequency of the secondary functional component; when the power consumption adjustment coefficient is equal to the preset threshold, controlling each functional component to maintain the current calculation frequency; when the power consumption adjustment coefficient is less than the preset threshold, selecting a target functional component whose calculation frequency is lower than a preset normal value from each functional component, and increasing the calculation frequency of the target functional component.

[0102] It should be understood that in this embodiment, when the power consumption adjustment coefficient is greater than a preset threshold value (that is, when the device performance of the target device is insufficient), secondary functional components can be selected from multiple functional components based on the importance of each functional component to the target device, and the calculation frequency of the secondary functional components can be reduced to ensure the normal operation of the core functional components; when the power consumption adjustment coefficient is less than the preset threshold value (that is, when the device performance of the target device is redundant), target functional components with a calculation frequency lower than the preset normal value are selected from the various functional components, and the calculation frequency of the target functional components is increased to restore the normal operation of the secondary functional components.

[0103] For ease of understanding, the following examples are given, but the present application is not limited. As an example, it is assumed that the scenario involved in the method is an attendance scenario based on feature recognition, and the multiple functional components of the target device include a target recognition component, an attribute recognition component, and a trigger event component, etc. Since the main goal of the current scenario is target recognition, the target recognition component is more important to the target device than the attribute recognition component. When k>1, the device performance of the target device is insufficient, and the calculation frequency of the attribute recognition component can be reduced to ensure the normal operation of the target recognition component. When k<1, there is redundancy in the device performance of the target device. At this time, if it is detected that the calculation frequency of the attribute recognition component is lower than the preset normal value, the calculation frequency of the attribute recognition component can be increased to restore the normal operation of the attribute recognition component.

[0104] Furthermore, the execution strategy of each functional component includes a working mode of each functional component; the execution strategy of each functional component is adjusted according to the power consumption adjustment coefficient, including: when the power consumption adjustment coefficient is greater than a preset threshold, the working mode of at least one functional component is switched to a high-performance mode, and the functional component switched to the high-performance mode is migrated to the first computing unit for execution, wherein the first computing unit is the computing unit with the strongest performance among the multiple computing units; when the power consumption adjustment coefficient is equal to the preset threshold, each functional component is controlled to maintain the current working mode; when the power consumption adjustment coefficient is less than the preset threshold, the working mode of at least one functional component is switched to a low-energy consumption mode, and the functional component switched to the low-energy consumption mode is migrated to the second computing unit for execution, wherein the second computing unit is the computing unit with the lowest energy consumption among the multiple computing units.

[0105] It should be understood that when the power consumption adjustment coefficient is greater than a preset threshold value (i.e., when the device performance of the target device is insufficient), the working mode of at least one functional component can be switched to a high-performance mode first, and the functional component switched to the high-performance mode can be migrated to the computing unit with the strongest performance for execution, so as to improve the device performance of the target device; when the power consumption adjustment coefficient is less than a preset threshold value (i.e., when the device performance of the target device is redundant), the working mode of at least one functional component can be switched to a low-energy mode first, and the functional component switched to the low-energy mode can be migrated to the computing unit with the lowest energy consumption for execution, so as to reduce the device energy consumption of the target device.

[0106] For ease of understanding, the following examples are given, but the present application is not limited. As an example, it is assumed that the scenario involved in the method is an attendance scenario based on feature recognition, and the multiple functional components of the target device include a target recognition component, an attribute recognition component, and a trigger event component, etc. The multiple computing units of the target device include a DSP and an NPU. The power consumption of the DSP is lower than that of the NPU, and the performance of the NPU is stronger than that of the DSP. Therefore, when k>1, the device performance of the target device is insufficient, and the working mode of the target recognition component can be switched to a high-performance mode, and the target recognition component can be migrated to the NPU for execution to improve the device performance of the target device. When k<1, the device performance of the target device is redundant, and the working mode of the attribute recognition component can be switched to a low-energy mode, and the attribute recognition component can be migrated to the DSP for execution to reduce the device energy consumption of the target device.

[0107] Further, adjusting the hardware parameters of each computing unit according to the power consumption adjustment coefficient includes:

[0108] Acquire current hardware parameters of each computing unit, and calculate parameter adjustment values ​​according to the power consumption adjustment coefficient and the current hardware parameters;

[0109] The current hardware parameters of each computing unit are adjusted according to the parameter adjustment value, wherein the current hardware parameters include a current computing frequency and / or a current voltage.

[0110] It should be understood that in order to improve the accuracy of hardware parameter adjustment, in this embodiment, the parameter adjustment value is calculated according to the power consumption adjustment coefficient and the current hardware parameter, and the current hardware parameter of each computing unit is adjusted according to the parameter adjustment value.

[0111] For ease of understanding, the following examples are given, but this application is not limited. As an example, the parameter adjustment value is first calculated based on the power consumption adjustment coefficient and the current hardware parameters. The specific calculation formula is: Δh = (k-1) * h, where Δh represents the parameter adjustment value, k represents the power consumption adjustment coefficient, and h represents the current hardware parameters. Then, the current hardware parameters of each computing unit are adjusted according to the parameter adjustment value Δh. In a specific implementation, when k>1, the device performance of the target device is insufficient, and it is necessary to improve the device performance of the target device. Since the calculated Δh>0, the current hardware parameters of each computing unit will be increased to improve the device performance of the target device. When k<1, there is redundancy in the device performance of the target device, and it is necessary to reduce the device power consumption of the target device. Since the calculated Δh<0, the current hardware parameters of each computing unit will be reduced to reduce the device energy consumption of the target device.

[0112] For ease of understanding, refer to Figure 3 This invention is provided for illustration, but is not intended to limit the present application. Figure 3 This is a specific flow chart of an embodiment of the device control method of the present application. In the figure, the device control method includes:

[0113] 1. Obtain multiple initial objects from multiple frames of images detected by the target device;

[0114] 2. Filter out short-time and recurring objects from multiple initial objects to obtain target objects, wherein short-time objects may refer to objects whose number of frames is less than a preset frame number threshold;

[0115] 3. Obtain motion information of multiple target objects from multiple frames of images detected by the target device;

[0116] 4. Calculate the average time required for the target device to process a single target object. The specific steps are: predict the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information; obtain the execution time of the complete process required for the target device to process a single frame of image; calculate the predicted time required for the target device to process each target object according to the number of existing frames and the execution time of the complete process, and calculate the average time required for the target device to process a single target object according to the predicted time;

[0117] 5. Calculate the power consumption adjustment coefficient based on the average time consumption and the target time consumption. The specific steps are: power consumption adjustment coefficient = average time consumption / target time consumption;

[0118] 6. Adjust the hardware parameters of each computing unit and / or the execution strategy of each functional component according to the power consumption adjustment coefficient. The specific steps are: adjust the current frequency and / or current voltage of each computing unit according to the power consumption adjustment coefficient; adjust the working mode or calculation frequency of each functional component according to the power consumption adjustment coefficient.

[0119] Reference Figure 4 , Figure 4 This is a flow chart of the third embodiment of the device control method of the present application. Based on the above embodiments, the third embodiment of the device control method of the present application is proposed.

[0120] In the third embodiment, the step S20 includes:

[0121] Step S201: predicting a set of motion position points of each target object according to the motion information.

[0122] It should be understood that in order to improve the accuracy of the number of existing frames, in this embodiment, the motion position point set of each target object is first predicted based on the motion information of each target object, and then the existence time of each target object in the detection area corresponding to the target device is predicted based on the motion position point set, and then the number of existence frames of each target object in the detection area corresponding to the target device is calculated based on the existence time and the number of image frames processed by the target device per second.

[0123] It can be understood that predicting the set of motion position points of each target object based on motion information can be predicting the set of motion position points of each target object based on the motion information through a preset motion model, wherein the preset motion model is used to estimate the position of the target object in the future frame based on the motion information, and the preset motion model can be a linear model, a polynomial model, a Kalman filter, etc., and this embodiment is not limited to this.

[0124] Step S202: predicting the existence time of each target object in the detection area corresponding to the target device according to the set of motion position points.

[0125] For ease of understanding, refer to Figure 5 This invention is provided for illustration, but is not intended to limit the present application. Figure 5This is a scene diagram of an embodiment of the device control method of the present application. As an example, in the figure, the square patterns are target objects, the lines with arrows are the sets of motion position points of the target objects, and the dotted box area is the detection area corresponding to the target device. Predicting the existence time of each target object in the detection area corresponding to the target device based on the motion position point set can be determining the time when each target object enters the detection area and the time when it leaves the detection area based on the motion position point set, and calculating the time difference from the target object entering the detection area to leaving the detection area, which is the existence time of the target object in the detection area.

[0126] Step S203: Calculate the number of existence frames of each target object in the detection area corresponding to the target device based on the existence time and the number of image frames processed per second by the target device.

[0127] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the existence time of the mth target object in the detection area corresponding to the target device is T m , the number of image frames processed by the target device per second is FPS, then the number of frames in which each target object exists in the detection area corresponding to the target device can be calculated by the following formula:

[0128] OTK m =T m *FPS

[0129] Where, OTK m represents the number of frames in which the mth target object exists in the detection area corresponding to the target device, T m It indicates the existence time of the mth target object in the detection area corresponding to the target device, and FPS indicates the number of image frames processed per second by the target device.

[0130] This embodiment first predicts the set of motion position points of each target object based on the motion information of each target object, then predicts the existence time of each target object in the detection area corresponding to the target device based on the set of motion position points, and then calculates the number of existence frames of each target object in the detection area corresponding to the target device based on the existence time and the number of image frames processed by the target device per second, thereby improving the accuracy of the number of existence frames and further improving the accuracy of power consumption adjustment.

[0131] In the third embodiment, the step S30 includes:

[0132] Step S301: obtaining the execution time consumed by each computing unit when executing each functional component to process a single frame of image, wherein the target device includes a plurality of computing units for heterogeneous computing.

[0133] It should be understood that in this embodiment, a computing unit of heterogeneous computing can be used in the target device, that is, when the target device needs to execute a certain functional component, it will be assigned to the most suitable computing unit for processing according to the type and characteristics of the functional component to achieve the best computing performance and efficiency. For example, in a scene of attendance based on feature recognition, the target device can use DSP (Digital Signal Processor) and NPU (Neural-network Process Units) to process multiple frames of images at the same time.

[0134] Taking into account that different computing units will have different execution times when performing different functions, in order to improve the accuracy of the execution time of the complete process, in this embodiment, the execution time of each computing unit when executing each functional component to process a single frame image is first obtained, and then the execution time is summed to obtain the complete process execution time required for the target device to process a single frame image.

[0135] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the execution time of the i-th computing unit executing the j-th functional component to process a single frame of image is represented by TD i,j , the first computing unit is DSP, the second computing unit is NPU, the first functional component is target detection component, and the second functional component is object filtering component. Then the execution time of DSP executing target detection component to process a single frame image is TD 1,1 The execution time of NPU executing the target detection component to process a single frame image is TD 2,1 , the execution time of DSP executing object filtering component to process a single frame image is TD 1,2 The execution time of NPU executing object filtering component to process a single frame image is TD 2,2 .

[0136] Step S302: summing up the execution times to obtain the execution time of the complete process required for the target device to process a single frame of image.

[0137] For ease of understanding, the following examples are given, but are not intended to limit the present application. As an example, assume that the execution time of the i-th computing unit executing the j-th functional component to process a single frame of image is represented by TD i,j , the time required for the target device to process a single frame of image can be calculated using the following formula:

[0138]

[0139] In the formula, TD i,jIt represents the execution time of the i-th computing unit executing the j-th functional component to process a single frame image, and TE represents the execution time of the complete process required for the target device to process a single frame image.

[0140] This embodiment first obtains the execution time of each computing unit when executing each functional component to process a single frame image, and then sums the execution time to obtain the complete process execution time required for the target device to process the single frame image, thereby improving the accuracy of the complete process execution time, and further improving the accuracy of power consumption regulation.

[0141] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the device control method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0142] This application also provides a device control device, please refer to Figure 6 , the equipment control device comprises:

[0143] The information acquisition module 10 is used to acquire multiple target objects and motion information of each target object from multiple frames of images detected by the target device;

[0144] A frame number prediction module 20, configured to predict the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information;

[0145] The time consumption acquisition module 30 is used to acquire the time consumption of the complete process execution required for the target device to process a single frame image;

[0146] A time calculation module 40 is used to calculate the predicted time required for the target device to process each target object according to the number of existing frames and the execution time of the complete process, and calculate the average time required for the target device to process a single target object according to the predicted time;

[0147] The device control module 50 is used to calculate a power consumption adjustment coefficient according to the average time consumption and the target time consumption, and control the target device according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption that satisfies the user demand corresponding to the target device.

[0148] The device control device provided by the present application adopts the device control method in the above embodiment, which can solve the technical problem that the related device control method takes improving the processing performance of the processing device as the control target, resulting in high device energy consumption and imbalance between device performance and device power consumption. Compared with the related technology, the beneficial effects of the device control device provided by the present application are the same as the beneficial effects of the device control method provided by the above embodiment, and the other technical features in the device control device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0149] The present application provides a device control device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the device control method in the above-mentioned embodiment one.

[0150] Reference below Figure 7 , which shows a schematic diagram of the structure of a device control device suitable for implementing the embodiment of the present application. The device control device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The device control device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0151] like Figure 7 As shown, the device control device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a ROM (Read Only Memory) 1002 or a program loaded from a storage device 1003 to a RAM (Random Access Memory) 1004. In RAM 1004, various programs and data required for the operation of the device control device are also stored. The processing device 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, an LCD (Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the device control device to communicate with other devices wirelessly or wired to exchange data. Although the figure shows a device control device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0152] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0153] The device control device provided by the present application adopts the device control method in the above embodiment, which can solve the technical problem that the related device control method takes improving the processing performance of the processing device as the control target, resulting in high device energy consumption and imbalance between device performance and device power consumption. Compared with the related technology, the beneficial effects of the device control device provided by the present application are the same as the beneficial effects of the device control method provided by the above embodiment, and the other technical features in the device control device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0154] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0155] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0156] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the device control method in the above-mentioned embodiment.

[0157] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, or infrared systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory), optical fiber, CD-ROM (CD-Read Only Memory, portable compact disk read-only memory), optical storage device, magnetic storage device, or any suitable combination of the above. In the present embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0158] The computer-readable storage medium may be included in the device control device; or may exist independently without being assembled into the device control device.

[0159] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the device control device, the device control device executes the device control method in the above embodiment.

[0160] The computer program code for performing the operation of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a LAN (Local Area Network) or a WAN (Wide Area Network), or it can be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0161] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0162] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0163] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned device control method, and can solve the technical problem that the related device control method takes improving the processing performance of the processing device as the control target, resulting in high device energy consumption and imbalance between device performance and device power consumption. Compared with the related art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the device control method provided by the above-mentioned embodiment, and will not be repeated here.

[0164] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A device control method, characterized in that: The device control method comprises: Acquire multiple target objects and motion information of each target object from multiple frames of images detected by the target device; Predicting the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information; Obtaining the execution time of the complete process required for the target device to process a single frame of image; Calculate the predicted time consumed by the target device to process each target object according to the number of existing frames and the execution time consumed by the complete process, and calculate the average time consumed by the target device to process a single target object according to the predicted time consumed; A power consumption adjustment coefficient is calculated according to the average time consumption and the target time consumption, and the target device is controlled according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption required to meet the user demand corresponding to the target device.

2. The device control method according to claim 1, characterized in that: The controlling the target device according to the power consumption adjustment coefficient includes: Acquiring hardware parameters of each computing unit and / or execution strategy of each functional component, wherein the target device includes multiple computing units and multiple functional components; The hardware parameters of each computing unit and / or the execution strategy of each functional component are adjusted according to the power consumption adjustment coefficient.

3. The device control method according to claim 2, characterized in that: The execution strategy of each functional component includes the calculation frequency of each functional component; The step of adjusting the execution strategy of each functional component according to the power consumption adjustment coefficient includes: When the power consumption adjustment coefficient is greater than a preset threshold, selecting a secondary functional component from a plurality of functional components according to the importance of each functional component to the target device, and reducing the calculation frequency of the secondary functional component; When the power consumption adjustment coefficient is equal to a preset threshold, controlling each functional component to maintain the current calculation frequency; When the power consumption adjustment coefficient is less than a preset threshold, a target functional component whose calculation frequency is lower than a preset normal value is selected from each functional component, and the calculation frequency of the target functional component is increased.

4. The device control method according to claim 2, characterized in that: The execution strategy of each functional component includes the working mode of each functional component; The step of adjusting the execution strategy of each functional component according to the power consumption adjustment coefficient includes: When the power consumption adjustment coefficient is greater than a preset threshold, the working mode of at least one functional component is switched to a high-performance mode, and the functional component switched to the high-performance mode is migrated to a first computing unit for execution, wherein the first computing unit is the computing unit with the strongest performance among the multiple computing units; When the power consumption adjustment coefficient is equal to a preset threshold, controlling each functional component to maintain a current working mode; When the power consumption adjustment coefficient is less than a preset threshold, the working mode of at least one functional component is switched to a low energy consumption mode, and the functional component switched to the low energy consumption mode is migrated to a second computing unit for execution, wherein the second computing unit is a computing unit with the lowest energy consumption among multiple computing units.

5. The device control method according to claim 2, characterized in that: The adjusting the hardware parameters of each computing unit according to the power consumption adjustment coefficient includes: Acquire current hardware parameters of each computing unit, and calculate parameter adjustment values ​​according to the power consumption adjustment coefficient and the current hardware parameters; The current hardware parameters of each computing unit are adjusted according to the parameter adjustment value, wherein the current hardware parameters include a current computing frequency and / or a current voltage.

6. The device control method according to any one of claims 1 to 5, characterized in that: The step of predicting the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information includes: Predicting a set of motion position points of each target object according to the motion information; Predicting the existence time of each target object in the detection area corresponding to the target device according to the set of motion position points; The number of frames in which each target object exists in the detection area corresponding to the target device is calculated based on the existence time and the number of image frames processed per second by the target device.

7. The device control method according to any one of claims 1 to 5, characterized in that: The time-consuming execution of the complete process required for obtaining the target device to process a single-frame image includes: Obtaining the execution time consumed by each computing unit when executing each functional component to process a single frame of image, wherein the target device includes a plurality of computing units for heterogeneous computing; The execution times are summed to obtain the complete process execution time required for the target device to process a single frame of image.

8. The device control method according to any one of claims 1 to 5, characterized in that: The calculating the power consumption adjustment coefficient according to the average time consumption and the target time consumption includes: Calculate the power consumption adjustment coefficient according to the average time consumption and the target time consumption by using the adjustment coefficient calculation formula; The adjustment coefficient calculation formula is: Where k is the power consumption adjustment coefficient, T ave is the average time, T ue Time consuming for the goal.

9. A device control device, characterized in that: The equipment control device comprises: An information acquisition module, used to acquire multiple target objects and motion information of each target object from multiple frames of images detected by the target device; A frame number prediction module, used to predict the number of frames in which each target object exists in the detection area corresponding to the target device according to the motion information; A time-consuming acquisition module is used to acquire the time-consuming execution of the complete process required for the target device to process a single frame of image; A time calculation module, used to calculate the predicted time required for the target device to process each target object according to the number of existing frames and the execution time of the complete process, and calculate the average time required for the target device to process a single target object according to the predicted time; The device control module is used to calculate the power consumption adjustment coefficient according to the average time consumption and the target time consumption, and control the target device according to the power consumption adjustment coefficient, wherein the target time consumption is the time consumption that meets the user demand corresponding to the target device.

10. A device control device, characterized in that: The device control device comprises: a memory, a processor, and a device control program stored in the memory and executable on the processor, wherein the device control program implements the device control method according to any one of claims 1 to 8 when executed by the processor.

11. A storage medium, characterized in that: The storage medium stores a device control program, and when the device control program is executed by the processor, the device control method according to any one of claims 1 to 8 is implemented.

Images