Delayed alarm method and device of buzzer and storage medium

Abstract

The invention discloses a delayed alarm method and device for a buzzer and a storage medium, and the method comprises the steps: collecting multiple frames of original image data in a freezer when the freezer is opened; executing target detection in the original image data to obtain regional image data of a plurality of users in the freezer; performing classification operation on the original image data to obtain events of common operation of a plurality of users in the freezer; performing classification operation on the regional image data to obtain an action of independent operation of the user in the freezer; for the same frame of original image data, converting an event and a plurality of actions into an operation feature sequence; predicting reference time of continuous operation of the plurality of users in the freezer according to the operation feature sequence; correcting the reference time into target time according to the event; and when the freezer is kept open and a temperature alarm signal reported by the temperature sensor is received, the buzzer is controlled to give an alarm sound in a delayed manner according to the target time. According to the embodiment, the alarm accuracy of the buzzer in the freezer is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision, and in particular to a buzzer delayed alarm method, device and storage medium. Background Art

[0002] A buzzer and a temperature sensor are provided in the freezer. When the temperature sensor detects that the freezer is at a high temperature (eg, higher than 5°C), the buzzer sounds an alarm.

[0003] When the staff enters the freezer to move out goods, move in goods, count goods, clean the freezer, etc., the freezer will be opened for a long time, making the freezer at a high temperature. Currently, a delay circuit is set inside the buzzer. When the temperature sensor detects that the freezer is at a high temperature (such as higher than 5°C), the buzzer will delay for a period of time and sound an alarm to prompt the staff to complete the handling as soon as possible to avoid affecting other goods in the freezer.

[0004] However, there is uncertainty in the time taken for the staff to operate, while the delay time of the buzzer is fixed. The delay time of the buzzer is often shorter than the time taken for the staff to operate, resulting in false alarms of the buzzer. Summary of the invention

[0005] In view of this, the present invention provides a buzzer delayed alarm method, device and storage medium, which are used to improve the accuracy of buzzer alarm in a frozen warehouse.

[0006] A first aspect of the present invention provides a buzzer delayed alarm method, comprising:

[0007] When the freezer is opened, multiple frames of original image data are collected inside the freezer; a temperature sensor and a buzzer are installed inside the freezer;

[0008] Performing target detection in the original image data to obtain image data of the area where multiple users in the frozen warehouse are located;

[0009] Performing a classification operation on the original image data to obtain a plurality of events of common operations of the users in the frozen store;

[0010] Performing a classification operation on the regional image data to obtain the actions of the user operating individually in the frozen store;

[0011] For the original image data of the same frame, converting the event and the plurality of actions into an operation feature sequence;

[0012] Predicting a benchmark time for a plurality of users to continuously operate the frozen storage according to the operation feature sequence;

[0013] Correcting the reference time to the target time according to the event;

[0014] When the freezer remains open and receives a temperature alarm signal reported by the temperature sensor, the buzzer is controlled to delay emitting an alarm sound according to the target time.

[0015] A second aspect of the present invention provides a buzzer delayed alarm device, comprising:

[0016] The original image data acquisition module is used to collect multiple frames of original image data inside the freezer when the freezer is opened; a temperature sensor and a buzzer are installed inside the freezer;

[0017] A regional image data detection module, used to perform target detection in the original image data to obtain regional image data where multiple users in the frozen warehouse are located;

[0018] An event classification module, used to perform classification operations on the original image data to obtain multiple events of common operations of the users in the frozen warehouse;

[0019] An action classification module is used to perform a classification operation on the regional image data to obtain the actions of the user's individual operations in the frozen warehouse;

[0020] An operation feature sequence conversion module, used for converting the event and the plurality of actions into an operation feature sequence for the same frame of the original image data;

[0021] A reference time prediction module, used for predicting the reference time of continuous operation of a plurality of users in the frozen storage according to the operation feature sequence;

[0022] A target time correction module, used for correcting the reference time to the target time according to the event;

[0023] The delayed alarm control module is used to control the buzzer to delay the alarm sound according to the target time when the freezer remains open and receives the temperature alarm signal reported by the temperature sensor.

[0024] A third aspect of the present invention provides an electronic device, the electronic device comprising:

[0025] at least one processor; and

[0026] a memory communicatively connected to the at least one processor; wherein,

[0027] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the delayed alarm method of the buzzer as described in the first aspect above.

[0028] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the delayed alarm method of the buzzer as described in the first aspect above.

[0029] A fifth aspect of the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the delayed alarm method of the buzzer as described in the first aspect above is implemented.

[0030] In this embodiment, when the freezer is opened, multiple frames of original image data are collected inside the freezer; a temperature sensor and a buzzer are installed inside the freezer; target detection is performed in the original image data to obtain regional image data where multiple users are located in the freezer; classification operations are performed on the original image data to obtain events in which multiple users operate together in the freezer; classification operations are performed on the regional image data to obtain actions of users operating individually in the freezer; for the same frame of original image data, events and multiple actions are converted into operation feature sequences; the benchmark time for multiple users to continuously operate in the freezer is predicted based on the operation feature sequence; the benchmark time is corrected to the target time based on the event; when the freezer remains open and receives a temperature alarm signal reported by the temperature sensor, the buzzer is controlled to delay the alarm sound based on the target time. This embodiment identifies events of multiple users operating together in the frozen storage at the overall visual level and identifies actions of users operating individually in the frozen storage at the local visual level, thereby predicting the benchmark time of multiple users operating continuously in the frozen storage, and correcting the benchmark time to the target time based on prior knowledge. The time consumed by users operating in the frozen storage can be dynamically identified, making the target time highly accurate, thereby effectively improving the accuracy of the buzzer alarm in the frozen storage.

[0031] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 This is a flow chart of a buzzer delayed alarm method provided in Embodiment 1 of the present invention.

[0034] Figure 2It is a schematic diagram of a time axis provided in Embodiment 1 of the present invention.

[0035] Figure 3 It is a structural diagram of an operation classification network provided in Example 1 of the present invention.

[0036] Figure 4 It is a structural schematic diagram of a buzzer delayed alarm device provided in Embodiment 2 of the present invention.

[0037] Figure 5 It is a structural schematic diagram of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can cover sequential implementations other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] Embodiment 1

[0041] See also Figure 1 , shows a flowchart of a buzzer delay alarm method provided by Embodiment 1 of the present invention, the method can be executed by a buzzer delay alarm device, the buzzer delay alarm device can be implemented in the form of hardware and / or software, and the buzzer delay alarm device can be configured in an electronic device. Figure 1 As shown, the method includes:

[0042] Step 101: When the freezer is opened, multiple frames of original image data are collected inside the freezer.

[0043] Generally, sensors such as switch sensors, cameras, temperature sensors and buzzers are installed inside the freezer to monitor the freezer and ensure its safe operation. This embodiment reuses existing sensors, adjusts their communication logic, and configures the server to perform calculations to implement the delayed alarm method of the buzzer, which can effectively reduce hardware adjustments and reduce hardware costs.

[0044] Among them, the switch sensor includes a door configuration Hall sensor, a door lock device, etc. The switch sensor is configured on the door of the freezer. When the door of the freezer is opened or closed, the switch sensor uploads a corresponding enable signal, indicating that the door of the freezer has been opened or closed.

[0045] When the switch sensor notifies that the freezer door is open, the staff (ie, the user) may enter the freezer to carry goods. At this time, the camera can be called to continuously collect multiple frames of original image data inside the freezer.

[0046] Step 102: Execute target detection in the original image data to obtain image data of the area where multiple users are located in the frozen warehouse.

[0047] In practical applications, each frame of original image data is traversed, and the user is taken as the detection target. Target detection is performed in each frame of original image data to obtain the range where multiple users are located in the frozen library, which is recorded as regional image data.

[0048] In one embodiment of the present invention, step 102 may include the following steps:

[0049] Step 1021: After the freezer is opened, the freezer is divided into an initial period and a non-initial period.

[0050] like Figure 2 As shown, on the time axis, the timestamp T0 when the switch sensor notifies the freezer door to open (i.e., receives the corresponding enable signal) is used as the starting point, and the initial period [T0, T1] and the non-initial period (T1, T2) are divided thereafter.

[0051] Generally, the length of the initial time period [T0, T1] is a default experience value, such as 3 minutes, and the non-initial time period (T1, T2) is the time period from the initial time period [T0, T1] to the current time T2.

[0052] Step 1022: In the initial period, query the reference image data collected inside the freezer before the freezer is opened.

[0053] During the initial period, the reference image data collected by the camera inside the freezer before the freezer door is opened may be queried from the database.

[0054] Generally, the reference image data is selected based on parameters such as clarity within a timestamp close to (eg, within 10 minutes) the time when the freezer door is opened.

[0055] Step 1023: Perform a differential operation on the original image data and the reference image data to obtain differential image data.

[0056] In actual applications, the interior decoration, goods and other objects in the freezer will basically not change before the freezer door is opened and when the freezer door is just opened and the user just enters the freezer. Therefore, a differential operation can be performed on the original image data and the reference image data to obtain differential image data representing the user.

[0057] Step 1024: extracting the image data of the area where multiple users are located in the frozen library from the original image data based on the differential image data.

[0058] In the differential image data, pixel points representing multiple users can be parsed, and with this as a reference, image data of the area where multiple users are located in the frozen library can be extracted from the original image data.

[0059] In the specific implementation, the interior decoration, goods and other objects in the frozen warehouse are used as background data (pixel points), and multiple users are used as foreground data. The foreground data (pixel points) are extracted from the differential image data, and a corrosion operation is performed on the foreground data to remove noise and obtain relatively pure foreground data. Since it will be expanded later, the impact of the corrosion operation on the foreground data can be eliminated.

[0060] If the erosion operation is completed, a minimum bounding rectangle is added to each independent foreground data. The so-called independence means that the foreground data belongs to the same connected area.

[0061] Taking the center of the minimum enclosing rectangle as the reference, the minimum enclosing rectangle is expanded in all directions to expand into the active area, and the active area is mapped to the original image data to obtain the regional image data of multiple users in the frozen library. In this way, some of the environment around the user can be included in the regional image data to provide a semantic reference for subsequent classification operations.

[0062] Step 1025: During the non-initial period, a tracking operation is performed on the image data of each region.

[0063] The regional image data of multiple users can be quickly located in a short initial period. Then, when entering the non-initial period, the tracking operation can be performed on each regional image data using algorithms such as feature point-based tracking and optical flow methods, thereby extracting the regional image data of multiple users in the frozen library from the subsequent original image data.

[0064] In practical applications, operations such as differentiation and tracking algorithms have lower computational complexity than deep learning, which can improve response speed in the initial period and have higher accuracy in specific scenarios when users enter the freezer.

[0065] Step 1026: If the tracking operation fails, the original image data is input into a preset target detection network to detect the image data of the area where multiple users are located in the frozen library.

[0066] In actual applications, the tracking operation fails due to factors such as overlap between users and users being obstructed by objects. In this case, the original image data can be input into a target detection network built and trained based on deep learning. The target detection network takes the user's body as the target and detects the image data of the area where multiple users (bodies) are located in the frozen storage.

[0067] Generally speaking, the target of the target detection network is the user's body, and the background is relatively pure, making the detection task relatively simple. Therefore, the target detection network can use a faster one-stage model, such as YOLO, SSD, RetainNet, etc., to improve the detection speed while ensuring the detection accuracy.

[0068] Step 1027: Expand the regional image data based on the center of the regional image data.

[0069] Generally speaking, the regional image data output by the target detection network is the range of its detection box. Therefore, the regional image data is expanded based on the center of the regional image data. In this way, some of the environment around the user can be included in the regional image data, providing a semantic reference for subsequent classification operations.

[0070] Step 103: Perform classification operations on the original image data to obtain events of joint operations by multiple users in the frozen storage.

[0071] In this embodiment, the semantics of the content recognition of the original image data can be used to perform multi-classification operations on the original image data to obtain events of joint operations by multiple users in the freezer, such as moving out goods, moving in goods, counting goods, cleaning the freezer, etc.

[0072] In the specific implementation, the operation classification network can be built and trained based on deep learning in advance, and when the frozen library is opened, the operation classification network can be loaded into the memory for operation.

[0073] like Figure 3 As shown, the operation classification network has a backbone structure Backbone, a branch structure Branch, a fusion structure Fusion, a first classification head structure ClassifyHead_1 and a second classification head structure ClassifyHead_2.

[0074] Among them, the backbone structure Backbone is used to extract features from the complete image data, the branch structure Branch is used to extract features from the local image data, the fusion structure Fusion is used to fuse the features of the complete image data with the features of the local image data, the first classification head structure ClassifyHead_1 is used to classify events in which multiple users operate together in the frozen warehouse, and the second classification head structure ClassifyHead_2 is used to classify the actions of users performing individual operations in the frozen warehouse.

[0075] Furthermore, both the backbone structure Backbone and the branch structure Branch can be CNN (Convolutional Neural Network) for extracting features from image data. The structures of the two are not limited to the partial structures in artificially designed neural networks for extracting features from image data, such as ResNet (residual network) and VGG (Visual Geometry Group) except the last layer or multiple layers of fully connected layers, etc., and can also be neural networks optimized by model quantization methods, neural networks searched for characteristics of frozen storage by NAS (Neural Architecture Search) methods, etc., and this embodiment does not impose any restrictions on this.

[0076] The fusion structure is responsible for the integration of features (such as dimensionality reduction, splicing, etc.) in order to better adapt to the task of classifying the actions of users performing individual operations in the freezer. It can include structures such as convolutional layers, pooling layers, fully connected layers, and downsampling layers.

[0077] The first classification head structure ClassifyHead_1 and the second classification head structure ClassifyHead_2 are both classification heads for multiple classifications of image data, which may include structures such as convolutional layers, fully connected layers, and activation functions (such as Softmax, etc.).

[0078] Since the backbone structure is reused, the training and operation of the classification network can be divided into the following two stages:

[0079] The first stage: Under the condition that the backbone structure Backbone and the first classification head structure ClassifyHead_1 are an independent classification model, the backbone structure and the first classification head structure are updated.

[0080] In the first stage, the branch structure Branch, the fusion structure Fusion and the second classification head structure ClassifyHead_2 are ignored, and the backbone structure Backbone and the first classification head structure ClassifyHead_1 are regarded as a complete, multi-classification classification model. The image data of events (i.e., labels) annotated with multiple users' common operations in the frozen warehouse are used as samples for supervised training of the classification model. During training, the cross entropy and other loss functions are used, and the parameters in the backbone structure Backbone and the first classification head structure ClassifyHead_1 are continuously iterated and updated according to the loss values ​​until the classification model is trained.

[0081] The second stage: take the backbone structure Backbone, branch structure Branch, fusion structure Fusion and the second classification head structure ClassifyHead_2 as an independent classification model, and update the branch structure Branch, fusion structure Fusion and the second classification head structure ClassifyHead_2 while keeping the backbone structure unchanged.

[0082] In the second stage, the first classification head structure ClassifyHead_1 is ignored, and the backbone structure Backbone, branch structure Branch, fusion structure Fusion and the second classification head structure ClassifyHead_2 are regarded as a complete, multi-classification classification model. The image data of the user's individual operation actions (i.e., labels) in the frozen storage are used as samples for supervised training of the classification model. During training, the cross entropy and other loss functions are used. The parameters in the backbone structure Backbone, branch structure Branch, fusion structure Fusion and the second classification head structure ClassifyHead_2 are continuously iterated and updated according to the loss values. The update is terminated when the backbone structure Backbone is reached, and the parameters of the backbone structure Backbone are kept unchanged until the classification model is trained.

[0083] In this embodiment, the original image data can be input into the backbone structure Backbone to extract features, which are recorded as original image features. The original image features are input into the first classification head structure ClassifyHead_1 to classify events of multiple users' joint operations in the frozen warehouse.

[0084] Generally, the first classification head structure ClassifyHead_1 identifies the probability that the content of the original image data belongs to multiple predetermined events, and determines that the event with the highest probability is an event in which multiple users jointly operate in the frozen library.

[0085] Step 104: Perform classification operations on the regional image data to obtain the user's individual operation actions in the frozen warehouse.

[0086] In this embodiment, the semantics of the content recognition of each area image data can be performed to perform multi-classification operations on the area image data to obtain the user's individual operation actions in the freezer, such as standing, squatting, lifting, writing, etc.

[0087] In a specific implementation, each regional image data is sequentially input into the branch structure Branch to extract features, which are recorded as regional image features.

[0088] The original image features and the regional image features are input into the fusion structure Fusion to jointly extract features, which are recorded as fused image features.

[0089] The fused image features are input into the second classification head structure ClassifyHead_2 to classify the user's individual operation actions in the frozen warehouse.

[0090] Generally, the second classification head structure ClassifyHead_2 identifies the probability that the content of the regional image data belongs to multiple predetermined actions, and determines that the action with the highest probability is the action of the user operating alone in the frozen warehouse.

[0091] In one design, Figure 3 As shown, the fusion structure Fusion includes a first convolution module ConvModule_1, a second convolution module ConvModule_2, a third convolution module ConvModule_3, a fourth convolution module ConvModule_4 and an atrous spatial convolution pooling layer (Atrous Spatial Pyramid Pooling, ASPP).

[0092] Among them, the first convolution module ConvModule_1, the second convolution module ConvModule_2, the third convolution module ConvModule_3, and the fourth convolution module ConvModule_4 are all modules that encapsulate CNN-related structures, such as convolution layers, pooling layers, normalization layers (such as BN (Batch Normlization)), activation functions (such as ReLU (Rectified Linear Unit)), etc.

[0093] The first convolution module ConvModule_1, the second convolution module ConvModule_2, the third convolution module ConvModule_3, and the fourth convolution module ConvModule_4 can reuse part of the structure of a third-party neural network, such as the C2f module and the convolution module ConvModule in YOLO, the conv2, conv3, conv4, and conv5 modules in ResNet, etc., or can be a custom structure, which is not limited in this embodiment.

[0094] In this design, the original image features are input into the first convolution module ConvModule_1 to extract features, which are recorded as the first candidate image features P1.

[0095] The first candidate image feature P1 is sequentially input into the second convolution module ConvModule_2 and the atrous spatial convolution pooling layer ASPP to extract features, which are recorded as the second candidate image feature P2.

[0096] A down-sampling operation DownSample is performed on the regional image feature to obtain a third candidate image feature P3.

[0097] The Concat function is used to concatenate the second candidate image feature P2 and the third candidate image feature P3 into a new feature, which is recorded as the fourth candidate image feature P4.

[0098] The fourth candidate image feature P4 is input into the third convolution module ConvModule_3 to extract features, which are recorded as the fifth candidate image feature P5.

[0099] A down-sampling operation DownSample is performed on the fifth candidate image feature P5 to obtain a sixth candidate image feature P6.

[0100] The Concat function is used to concatenate the first candidate image feature P1 and the sixth candidate image feature P6 into a new feature, which is recorded as the seventh candidate image feature P7.

[0101] The seventh candidate image feature P7 is input into the fourth convolution module to extract features, which are recorded as fused image features.

[0102] In this embodiment, the complete original image data contains complete environmental information and action information of each user. There is an intrinsic correlation between the action information of each user because they jointly operate a certain event in the frozen storage. The original image features output by the trunk structure and the regional image features output by the branch structure are integrated, which can effectively increase the amount of feature information and focus attention on the regional image features, thereby improving the accuracy of detecting the actions of users operating individually in the frozen storage.

[0103] Step 105: For the same frame of original image data, convert the event and multiple actions into an operation feature sequence.

[0104] In this embodiment, each frame of original image data can be traversed, and in the same frame of original image data, events and multiple actions can be vectorized and converted into a structured operation feature sequence to facilitate subsequent prediction of the benchmark time for multiple users to continuously operate in the freezer.

[0105] Exemplarily, events are mapped to first vectors and actions are mapped to second vectors.

[0106] In the same frame of original image data, the second vector is arranged after the first vector according to the distribution order of the regional image data (such as from top to bottom, from left to right), and an operation feature sequence is obtained.

[0107] The dimension of the operational feature sequence is compared with a preset threshold.

[0108] If the dimension of the operation feature sequence is equal to a preset threshold, it is determined that the operation feature sequence is valid.

[0109] If the dimension of the operation feature sequence is smaller than the preset threshold, a specified empty vector is added to the end of the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold.

[0110] If the dimension of the operation feature sequence is greater than a preset threshold, a merging operation is performed on the second vector in the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold.

[0111] The merge operation includes:

[0112] Any two area image data with the same action are combined into an image pair.

[0113] Calculates the distance between two regions of image data (represented by center points) in an image pair.

[0114] The second vectors corresponding to the two regional image data in the image pair are merged in sequence according to the distance, that is, the second vectors corresponding to the two regional image data in the image pair with the smallest distance are merged first, and then the second vectors corresponding to the two regional image data in the image pair with the second smallest distance are merged, and so on, until the dimension of the operation feature sequence is equal to the threshold.

[0115] Furthermore, when merging the second vectors corresponding to two regional image data in an image pair, any regional image data in the image pair and its corresponding second vector can be randomly selected to be retained so as to participate in subsequent merging operations, while the other regional image data in the image pair and its corresponding second vector can be discarded.

[0116] Step 106: predicting the benchmark time for multiple users to continuously operate in the frozen storage based on the operation feature sequence.

[0117] In this embodiment, a lightweight time series prediction model can be constructed based on deep learning, such as LSTM (Long Short Term Memory) and GRU (Gated Recurrent Unit) to improve the response speed.

[0118] In an offline environment, the time series prediction model can be trained by defining historical operation feature sequences as samples and the actual time of multiple users' continuous operations in the frozen storage as labels.

[0119] The current operation feature sequence is input into the time series prediction model for processing, so as to predict the continuous operation time of multiple users in the frozen warehouse, which is recorded as the benchmark time.

[0120] Step 107: Correct the reference time to the target time according to the event.

[0121] In actual applications, when the freezer door is just opened, the operations of multiple users are relatively limited, the amount of information in the operation feature sequence is low, and the benchmark time may have a certain deviation. As multiple users gradually carry out operations, the amount of information in the operation feature sequence gradually becomes richer, and the benchmark time will gradually stabilize.

[0122] Therefore, when the door of the freezer is just opened, the reference time can be corrected to the target time based on the prior knowledge of handling various events, thereby reducing the fluctuation of the target time and improving the accuracy of the target time.

[0123] In a specific implementation, the number of events of each type is counted, and the event (type) with the highest number is selected from all types of events.

[0124] Query the standard time, step size and fluctuation range configured for the highest number of events; the standard time is the average time consumed from opening to closing the freezer under the same type of events in history; the fluctuation range is the range of time consumed from opening to closing the freezer under the same type of events in history, which has an upper limit and a lower limit.

[0125] The base time is compared to the standard time, and the difference between the base time and the standard time is compared to the step size.

[0126] In one case, if the reference time is greater than the standard time and the difference between the reference time and the standard time is less than or equal to the step size, it means that the deviation between the reference time and the standard time is small, and the reference time is set as the target time.

[0127] In one case, if the base time is greater than the standard time, and the difference between the base time and the standard time is greater than the step size, indicating that the deviation between the base time and the standard time is large, the base time is subtracted from the first offset time to obtain the target time; wherein the first offset time is a specified ratio of the step size multiplied by the product of the difference between the standard time and the lower limit of the fluctuation range.

[0128] In one case, if the reference time is less than the standard time and the difference between the standard time and the reference time is less than or equal to the step size, it means that the deviation between the reference time and the standard time is small, and the reference time is set as the target time.

[0129] In one case, if the base time is less than the standard time and the difference between the standard time and the base time is greater than the step size, indicating that the deviation between the base time and the standard time is large, the base time is added with the second offset time to obtain the target time; wherein the second offset time is the step size multiplied by the product of the upper limit value of the fluctuation range and the difference between the standard time and a specified ratio.

[0130] Then, the correction process is expressed as:

[0131] If Tbase>Tsta, Tbase-Tsta≤Step, then Ttarget=Tbase.

[0132] If Tbase>Tsta, Tbase-Tsta>Step, then Ttarget=Tbase-Step×(Tsta-DownLimit) / w.

[0133] If Tbase<Tsta, Tsta-Tbase≤Step, then Ttarget=Tbase.

[0134] If Tbase<Tsta, Tsta-Tbase>Step, then Ttarget=Tbase+Step×(UpLimit-Tsta) / w.

[0135] Among them, Tbase is the base time, Tsta is the standard time, Ttarget is the target time, Step is the step size, UpLimit is the upper limit of the fluctuation range, DownLimit is the lower limit of the fluctuation range, and w is the ratio.

[0136] In this embodiment, the standard time, step size, fluctuation range, and ratio are used to constrain the reference time to converge to the standard time within a limited range, which can effectively reduce the fluctuation of the reference time.

[0137] Step 108: When the freezer remains open and a temperature alarm signal reported by the temperature sensor is received, the buzzer is controlled to delay the alarm sound according to the target time.

[0138] In the process of determining whether the freezer door remains open based on the enable signal of the switch sensor, if a temperature alarm signal is received from the temperature sensor when it detects that the temperature value in the freezer is greater than a preset threshold (such as 5°C), the buzzer can be controlled to delay the alarm sound according to the target time.

[0139] If there is no delay circuit inside the buzzer, the temperature sensor can be controlled to energize the buzzer when the target time is reached, so that the buzzer emits an alarm sound.

[0140] If a delay circuit is set inside the buzzer, the historical time from the opening of the freezer to the receipt of the temperature alarm signal can be counted.

[0141] Subtract the historical time from the target time control to obtain the margin time, which indicates the remaining delay time. The margin time is compared with the delay time provided by the delay circuit in the buzzer.

[0142] If the margin time is less than or equal to the delay time provided by the delay circuit in the buzzer, the temperature sensor is immediately controlled to energize the buzzer so that an alarm sounds when the delay time is reached. The delay time provided by the delay circuit in the buzzer is relatively short (such as 10 minutes). The delay time provided by the delay circuit in the buzzer exceeds the margin time by a short amount, and the impact on the safety monitoring of the freezer is relatively limited and can be ignored.

[0143] If the margin time is greater than the delay time provided by the delay circuit in the buzzer, the delay time is subtracted from the margin time to obtain the waiting time.

[0144] When the waiting time is reached, the temperature sensor is controlled to attract the buzzer and energize it, so that an alarm sound is issued when the delay time is reached. At this time, the sum of the delay time, the waiting time and the margin time is the target time, and the alarm time can be accurately controlled.

[0145] In this embodiment, when the freezer is opened, multiple frames of original image data are collected inside the freezer; a temperature sensor and a buzzer are installed inside the freezer; target detection is performed in the original image data to obtain regional image data where multiple users are located in the freezer; classification operations are performed on the original image data to obtain events in which multiple users operate together in the freezer; classification operations are performed on the regional image data to obtain actions of users operating individually in the freezer; for the same frame of original image data, events and multiple actions are converted into operation feature sequences; the benchmark time for multiple users to continuously operate in the freezer is predicted based on the operation feature sequence; the benchmark time is corrected to the target time based on the event; when the freezer remains open and receives a temperature alarm signal reported by the temperature sensor, the buzzer is controlled to delay the alarm sound based on the target time. This embodiment identifies events of multiple users operating together in the frozen storage at the overall visual level and identifies actions of users operating individually in the frozen storage at the local visual level, thereby predicting the benchmark time of multiple users operating continuously in the frozen storage, and correcting the benchmark time to the target time based on prior knowledge. The time consumed by users operating in the frozen storage can be dynamically identified, making the target time highly accurate, thereby effectively improving the accuracy of the buzzer alarm in the frozen storage.

[0146] Embodiment 2

[0147] See also Figure 4 , shows a schematic diagram of the structure of a buzzer delay alarm device provided by Embodiment 2 of the present invention. Figure 4 As shown, the device comprises:

[0148] The original image data acquisition module 401 is used to collect multiple frames of original image data inside the freezer when the freezer is opened; a temperature sensor and a buzzer are installed inside the freezer;

[0149] The regional image data detection module 402 is used to perform target detection in the original image data to obtain regional image data where multiple users in the frozen warehouse are located;

[0150] An event classification module 403 is used to perform a classification operation on the original image data to obtain multiple events of common operations of the users in the frozen warehouse;

[0151] An action classification module 404 is used to perform a classification operation on the regional image data to obtain the actions of the user's individual operation in the freezer;

[0152] An operation feature sequence conversion module 405 is used to convert the event and the plurality of actions into an operation feature sequence for the same frame of the original image data;

[0153] A reference time prediction module 406, used for predicting the reference time of continuous operation of the plurality of users in the freezer according to the operation feature sequence;

[0154] A target time correction module 407, used to correct the reference time to the target time according to the event;

[0155] The delayed alarm control module 408 is used to control the buzzer to delay the alarm sound according to the target time when the freezer remains open and receives the temperature alarm signal reported by the temperature sensor.

[0156] In one embodiment of the present invention, the regional image data detection module 402 includes:

[0157] A time period division module, used for dividing the initial time period and the non-initial time period after the freezer is opened;

[0158] A reference image data query module, used to query the reference image data collected inside the frozen store before the frozen store is opened during the initial period;

[0159] A differential image data calculation module, used for performing a differential operation on the original image data and the reference image data to obtain differential image data;

[0160] A regional image data extraction module, used for extracting regional image data where multiple users in the frozen warehouse are located from the original image data according to the differential image data;

[0161] A region tracking module, used for performing a tracking operation on each of the region image data during the non-initial period;

[0162] A target detection module, for inputting the original image data into a preset target detection network to detect image data of the area where multiple users are located in the frozen warehouse if the tracking operation fails;

[0163] The regional image data expansion module is used to expand the regional image data based on the center of the regional image data.

[0164] In one embodiment of the present invention, the regional image data extraction module is further used for:

[0165] extracting foreground data from the differential image data;

[0166] performing an erosion operation on the foreground data;

[0167] If the erosion operation is completed, a minimum bounding rectangle is added to each independent foreground data;

[0168] Taking the center of the minimum bounding rectangle as a reference, expanding the minimum bounding rectangle into an active area;

[0169] The activity area is mapped to the original image data to obtain the area image data where multiple users in the frozen warehouse are located.

[0170] In one embodiment of the present invention, the event classification module 403 is further used to:

[0171] Loading an operation classification network; the operation classification network has a trunk structure, a branch structure, a fusion structure, a first classification head structure and a second classification head structure;

[0172] Inputting the original image data into the backbone structure to extract original image features;

[0173] Inputting the original image features into the first classification head structure to classify multiple events of common operations of the users in the frozen warehouse;

[0174] The action classification module 404 includes:

[0175] A regional image feature extraction module, used for inputting the regional image data into the branch structure to extract regional image features;

[0176] An image feature fusion module, used for inputting the original image features and the regional image features into the fusion structure to extract fused image features;

[0177] The image feature classification module is used to input the fused image features into the second classification head structure to classify the actions of the user performing individual operations in the frozen warehouse.

[0178] In one embodiment of the present invention, it also includes:

[0179] A first model updating module, used for updating the main structure and the first classification head structure under the condition that the main structure and the first classification head structure are an independent classification model;

[0180] The second model updating module is used to take the trunk structure, the branch structure, the fusion structure and the second classification head structure as an independent classification model, and update the branch structure, the fusion structure and the second classification head structure under the condition of maintaining the trunk structure unchanged.

[0181] In one embodiment of the present invention, the fusion structure includes a first convolution module, a second convolution module, a third convolution module, a fourth convolution module and a dilated spatial convolution pooling layer;

[0182] The image feature fusion module is also used for:

[0183] Inputting the original image features into the first convolution module to extract first candidate image features;

[0184] Inputting the first candidate image features into the second convolution module and the atrous spatial convolution pooling layer in sequence to extract second candidate image features;

[0185] Performing a downsampling operation on the regional image feature to obtain a third candidate image feature;

[0186] splicing the second candidate image feature and the third candidate image feature into a fourth candidate image feature;

[0187] Inputting the fourth candidate image feature into the third convolution module to extract the fifth candidate image feature;

[0188] Performing a downsampling operation on the fifth candidate image feature to obtain a sixth candidate image feature;

[0189] splicing the first candidate image feature and the sixth candidate image feature into a seventh candidate image feature;

[0190] The seventh candidate image feature is input into the fourth convolution module to extract the fused image feature.

[0191] In one embodiment of the present invention, the operation feature sequence conversion module 405 is further used to:

[0192] Mapping the event into a first vector and mapping the action into a second vector respectively;

[0193] In the original image data of the same frame, the second vector is arranged after the first vector according to the distribution order of the regional image data to obtain an operation feature sequence;

[0194] If the dimension of the operation feature sequence is equal to a preset threshold, then determining that the operation feature sequence is valid;

[0195] If the dimension of the operation feature sequence is less than a preset threshold, a specified empty vector is added to the end of the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold;

[0196] If the dimension of the operation feature sequence is greater than a preset threshold, performing a merging operation on the second vector in the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold;

[0197] The merging operation includes:

[0198] Combining any two of the area image data with the same action into an image pair;

[0199] Calculating the distance between the image data of two regions in the image pair;

[0200] The second vectors corresponding to the two area image data in the image pair are merged in sequence according to the distance.

[0201] In one embodiment of the present invention, the target time correction module 407 is further used to:

[0202] Counting the number of various events;

[0203] Query the standard time, step size and fluctuation range configured for the event with the highest number; the standard time is the average time consumed from opening to closing of the frozen warehouse under the event in history; the fluctuation range is the interval of time consumed from opening to closing of the frozen warehouse under the event in history;

[0204] If the reference time is greater than the standard time, and the difference between the reference time and the standard time is less than or equal to the step length, setting the reference time as the target time;

[0205] If the reference time is greater than the standard time, and the difference between the reference time and the standard time is greater than the step length, the first offset time is subtracted from the reference time to obtain the target time; the first offset time is a specified ratio of the product of the step length multiplied by the difference between the standard time and the lower limit of the fluctuation range;

[0206] If the reference time is less than the standard time, and the difference between the standard time and the reference time is less than or equal to the step length, setting the reference time as the target time;

[0207] If the reference time is less than the standard time, and the difference between the standard time and the reference time is greater than the step length, the reference time is added with a second offset time to obtain the target time; the second offset time is a specified ratio of the product of the step length multiplied by the difference between the upper limit of the fluctuation range and the standard time;

[0208] The delay alarm control module 408 is also used for:

[0209] Counting the historical time from the opening of the freezer to the receipt of the temperature alarm signal;

[0210] Subtract the historical time from the target time control to obtain the margin time;

[0211] If the margin time is less than or equal to the delay time provided by the buzzer, the temperature sensor is controlled to energize the buzzer so as to emit an alarm sound when the delay time is reached;

[0212] If the margin time is greater than the delay time provided by the buzzer, the delay time is subtracted from the margin time to obtain the waiting time;

[0213] When the waiting time is reached, the temperature sensor is controlled to attract the buzzer and energize it, so as to issue an alarm sound when the delay time is reached.

[0214] The delayed alarm device for a buzzer provided in the embodiment of the present invention can execute the delayed alarm method for a buzzer provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the delayed alarm method for a buzzer.

[0215] Embodiment 3

[0216] See also Figure 5, shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0217] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0218] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0219] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the delayed alarm method of the buzzer.

[0220] In some embodiments, the delayed alarm method of the buzzer can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the delayed alarm method of the buzzer described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the delayed alarm method of the buzzer in any other appropriate manner (for example, by means of firmware).

[0221] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0222] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0223] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0224] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0225] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0226] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0227] Embodiment 4

[0228] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the delayed alarm method of the buzzer provided in any embodiment of the present invention is implemented.

[0229] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, 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 a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a 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 local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0230] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0231] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A buzzer delayed alarm method, characterized in that: include: When the freezer is opened, multiple frames of original image data are collected inside the freezer; A temperature sensor and a buzzer are installed inside the freezer; Performing target detection in the original image data to obtain image data of the area where multiple users in the frozen warehouse are located; Performing a classification operation on the original image data to obtain a plurality of events of common operations of the users in the frozen store; Performing a classification operation on the regional image data to obtain the actions of the user operating individually in the frozen store; For the original image data of the same frame, converting the event and the plurality of actions into an operation feature sequence; Predicting a benchmark time for a plurality of users to continuously operate the frozen storage according to the operation feature sequence; Correcting the reference time to the target time according to the event; When the freezer remains open and receives a temperature alarm signal reported by the temperature sensor, the buzzer is controlled to delay emitting an alarm sound according to the target time.

2. The method according to claim 1, characterized in that The performing of target detection in the original image data to obtain image data of the area where multiple users in the frozen warehouse are located includes: After the freezer is opened, an initial period and a non-initial period are divided; In the initial period, querying the reference image data collected inside the frozen storage before the frozen storage is opened; Performing a difference operation on the original image data and the reference image data to obtain difference image data; Extracting image data of regions where multiple users in the frozen warehouse are located from the original image data according to the differential image data; During the non-initial period, performing a tracking operation on each of the regional image data; If the tracking operation fails, the original image data is input into a preset target detection network to detect image data of the area where multiple users are located in the frozen warehouse; The regional image data is expanded based on the center of the regional image data.

3. The method according to claim 2, characterized in that The step of extracting the image data of the area where the multiple users in the frozen warehouse are located from the original image data according to the differential image data includes: extracting foreground data from the differential image data; performing an erosion operation on the foreground data; If the erosion operation is completed, a minimum bounding rectangle is added to each independent foreground data; Taking the center of the minimum circumscribed rectangle as a reference, expanding the minimum circumscribed rectangle into an active area; The activity area is mapped to the original image data to obtain the area image data where multiple users in the frozen warehouse are located.

4. The method according to claim 1, characterized in that: The performing of classification operation on the original image data to obtain a plurality of events of common operations of the users in the frozen storehouse includes: Loading an operation classification network; the operation classification network has a trunk structure, a branch structure, a fusion structure, a first classification head structure and a second classification head structure; Inputting the original image data into the backbone structure to extract original image features; Inputting the original image features into the first classification head structure to classify multiple events of common operations of the users in the frozen warehouse; The performing of classification operation on the regional image data to obtain the action of the user's individual operation in the freezer includes: Inputting the regional image data into the branch structure to extract regional image features; Inputting the original image features and the regional image features into the fusion structure to extract fused image features; The fused image features are input into the second classification head structure to classify the actions performed by the user individually in the frozen store.

5. The method according to claim 4, characterized in that Also includes: Under the condition that the trunk structure and the first classification head structure are an independent classification model, updating the trunk structure and the first classification head structure; The trunk structure, the branch structure, the fusion structure and the second classification head structure are taken as an independent classification model, and the branch structure, the fusion structure and the second classification head structure are updated under the condition of keeping the trunk structure unchanged.

6. The method according to claim 4, characterized in that The fusion structure includes a first convolution module, a second convolution module, a third convolution module, a fourth convolution module and a dilated spatial convolution pooling layer; The step of inputting the original image features and the regional image features into the fusion structure to extract the fused image features comprises: Inputting the original image features into the first convolution module to extract first candidate image features; Inputting the first candidate image features into the second convolution module and the atrous spatial convolution pooling layer in sequence to extract second candidate image features; Performing a downsampling operation on the regional image feature to obtain a third candidate image feature; splicing the second candidate image feature and the third candidate image feature into a fourth candidate image feature; Inputting the fourth candidate image feature into the third convolution module to extract the fifth candidate image feature; Performing a downsampling operation on the fifth candidate image feature to obtain a sixth candidate image feature; splicing the first candidate image feature and the sixth candidate image feature into a seventh candidate image feature; The seventh candidate image feature is input into the fourth convolution module to extract the fused image feature.

7. The method according to any one of claims 1 to 6, characterized in that The converting the event and the plurality of actions into an operation feature sequence for the same frame of the original image data comprises: Mapping the event into a first vector and mapping the action into a second vector respectively; In the original image data of the same frame, the second vector is arranged after the first vector according to the distribution order of the regional image data to obtain an operation feature sequence; If the dimension of the operation feature sequence is equal to a preset threshold, then determining that the operation feature sequence is valid; If the dimension of the operation feature sequence is less than a preset threshold, a specified empty vector is added to the end of the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold; If the dimension of the operation feature sequence is greater than a preset threshold, performing a merging operation on the second vector in the operation feature sequence until the dimension of the operation feature sequence is equal to the threshold; The merging operation includes: Combining any two of the area image data with the same action into an image pair; Calculating the distance between the image data of two regions in the image pair; The second vectors corresponding to the two area image data in the image pair are merged in sequence according to the distance.

8. The method according to any one of claims 1 to 6, characterized in that The step of correcting the reference time to the target time according to the event includes: Counting the number of various events; Query the standard time, step size and fluctuation range configured for the event with the highest number; the standard time is the average time consumed from opening to closing of the frozen warehouse under the event in history; the fluctuation range is the interval of time consumed from opening to closing of the frozen warehouse under the event in history; If the reference time is greater than the standard time, and the difference between the reference time and the standard time is less than or equal to the step length, setting the reference time as the target time; If the reference time is greater than the standard time, and the difference between the reference time and the standard time is greater than the step length, the first offset time is subtracted from the reference time to obtain the target time; the first offset time is a specified ratio of the product of the step length multiplied by the difference between the standard time and the lower limit of the fluctuation range; If the reference time is less than the standard time, and the difference between the standard time and the reference time is less than or equal to the step length, setting the reference time as the target time; If the reference time is less than the standard time, and the difference between the standard time and the reference time is greater than the step length, the reference time is added with a second offset time to obtain the target time; the second offset time is a specified ratio of the product of the step length multiplied by the difference between the upper limit of the fluctuation range and the standard time; The step of controlling the buzzer to delay the alarm sound according to the target time includes: Counting the historical time from the opening of the freezer to the receipt of the temperature alarm signal; Subtract the historical time from the target time control to obtain the margin time; If the margin time is less than or equal to the delay time provided by the buzzer, the temperature sensor is controlled to energize the buzzer so as to emit an alarm sound when the delay time is reached; If the margin time is greater than the delay time provided by the buzzer, the delay time is subtracted from the margin time to obtain the waiting time; When the waiting time is reached, the temperature sensor is controlled to attract the buzzer and energize it, so as to issue an alarm sound when the delay time is reached.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the delayed alarm method of the buzzer according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the delayed alarm method of the buzzer according to any one of claims 1 to 8 is implemented.

Images