Intelligent nursing video monitoring processing method and system

Abstract

The invention provides an intelligent nursing video monitoring processing method and system. The method comprises the following steps: acquiring first video monitoring data of a to-be-monitored nursing area at a first resolution, and if it is determined that a target abnormal behavior exists in the first video monitoring data, switching to a second resolution with higher definition to acquire second video monitoring data of the to-be-monitored nursing area; therefore, the video resolution is improved only when the abnormal behavior is detected, and unnecessary waste of data storage resources is avoided.

Description

Technical Field

[0001] The present application relates to data processing technology, and more particularly to an intelligent childcare video monitoring and processing method and system. Background Art

[0002] As the demand for childcare services grows, the safety and management of childcare institutions are receiving increasing attention. In order to ensure the safety and health of children in childcare environments, real-time monitoring has become an indispensable management tool. Traditional video surveillance systems usually use fixed resolution for continuous recording. Although this method can provide uninterrupted video recording, it also has obvious limitations.

[0003] On the one hand, although high-resolution video data can provide clearer and more detailed images, continuous recording consumes a lot of storage space and requires extremely high computing resources. For childcare institutions, this not only increases operating costs, but may also cause the system to be unable to operate stably for a long time due to limitations in storage space and computing power.

[0004] On the other hand, although low-resolution video data can reduce storage space usage and computing burden, during the monitoring process, when abnormal behavior or emergency situations occur, low-resolution video often cannot provide sufficiently clear and accurate image information, which brings difficulties to subsequent behavior analysis, event recording and evidence retention.

[0005] Therefore, how to ensure the real-time monitoring and economical data storage while quickly obtaining high-quality video data when abnormal behavior occurs has become a technical problem that needs to be solved in the field of childcare monitoring. Summary of the invention

[0006] The present application provides an intelligent childcare video monitoring and processing method and system, which avoids unnecessary waste of data storage resources by improving the video resolution only when abnormal behavior is detected.

[0007] In a first aspect, the present application provides an intelligent childcare video monitoring and processing method, comprising:

[0008] Acquire first video monitoring data of the nursery area to be monitored at a first resolution;

[0009] If it is determined that there is abnormal target behavior in the first video monitoring data, switch to a second resolution to obtain second video monitoring data of the nursery area to be monitored, wherein the second resolution is higher than the first resolution.

[0010] In the above scheme, the dual goals of efficient monitoring and accurate identification of abnormal target behavior are achieved by intelligently adjusting the resolution of video monitoring data. First, the first video monitoring data of the nursery area to be monitored is obtained at the first resolution. This step fully takes into account the demand for storage space and computing resources for continuous video monitoring. The selection of the first resolution is intended to balance the real-time monitoring and the economy of data storage. By adopting a relatively low resolution for preliminary monitoring, the amount of video data generated can be effectively reduced, thereby reducing the occupancy of storage space and reducing the computational burden of subsequent processing steps. When the model detects the presence of abnormal behavior, it will immediately switch to the second resolution to obtain the second video monitoring data of the nursery area to be monitored. The second resolution is higher than the first resolution and can provide clearer and more detailed video images. The design of this switching mechanism ensures that when abnormal behavior occurs, the system can quickly obtain high-quality video data, which provides strong support for subsequent behavior analysis, event recording and evidence retention. It can be seen that on the one hand, the above scheme avoids unnecessary waste of resources and improves the overall efficiency of the monitoring system by only increasing the video resolution when abnormal behavior is detected. On the other hand, the precise capture of high-resolution video data provides a richer and more accurate source of information for subsequent abnormal behavior analysis, helping to improve the accuracy and reliability of the analysis.

[0011] Optionally, after switching to the second resolution to obtain the second video monitoring data of the nursery area to be monitored, the method further includes:

[0012] If it is determined that the duration of the target abnormal behavior not existing in the second video monitoring data exceeds the preset duration threshold, switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored.

[0013] In the above scheme, if no abnormal behavior of the target is detected for a long time, continuing to monitor the video at a high resolution (second resolution) will cause unnecessary waste of storage space and computing resources. By switching back to the first resolution to obtain the third video monitoring data, the amount of data generated can be significantly reduced, thereby reducing storage costs, and reducing the computing burden of the system, improving overall operating efficiency. Despite switching back to the first resolution, the system still maintains the ability to continuously monitor the nursery area. This ensures that the system can promptly detect and respond to possible abnormal events even when resource consumption is low.

[0014] The risk of abnormal behavior in childcare areas varies in different time periods. For example, the probability of abnormal behavior is lower during infants’ lunch breaks or quiet activities. By dynamically adjusting the resolution, the system can flexibly adjust resource allocation based on actual monitoring conditions to meet the needs of different time periods.

[0015] Although the preset time threshold is used to determine when to switch back to the first resolution, the system still maintains the ability to monitor abnormal behavior in real time. Once abnormal behavior is detected again within the preset time threshold, the system can immediately switch back to the second resolution for high-precision monitoring to ensure a quick response to emergencies.

[0016] In addition, due to environmental interference or sensor errors, the system may misjudge abnormal behavior for a short period of time. By introducing a preset time threshold, the system can continuously observe monitoring data over a period of time to reduce false alarms caused by short-term interference.

[0017] Optionally, a first abnormal behavior monitoring model is used to monitor the target abnormal behavior from the first video monitoring data, and a second abnormal behavior monitoring model is used to monitor the target abnormal behavior from the second video monitoring data, wherein the recognition accuracy of the first abnormal behavior monitoring model is higher than that of the second abnormal behavior monitoring model.

[0018] In the above scheme, in the first video monitoring stage, the first abnormal behavior monitoring model with higher recognition accuracy is used for screening. This model can more accurately identify potential target abnormal behaviors and reduce misjudgments and missed judgments. This not only provides a more accurate target area for subsequent high-resolution monitoring, but also improves the accuracy of the entire monitoring system. It is worth noting that when switching to the second video monitoring stage, although the second abnormal behavior monitoring model with faster recognition speed is used, since the recognition process of the second abnormal behavior monitoring model has a preset time threshold as a buffer time, and this scheme is intended to record and save the monitoring video clips with target abnormal behaviors, and the second abnormal behavior monitoring model is used to determine the time end of high-resolution recording, and the recognition accuracy of the second abnormal behavior monitoring model can be compensated by the cache time, therefore, the recognition accuracy of the second abnormal behavior monitoring model can be set to be lower than the recognition accuracy of the first abnormal behavior monitoring model. Thereby, the efficiency of monitoring can be guaranteed, and the monitoring accuracy of the starting point of abnormal behavior and the monitoring reliability of the end point of abnormal behavior can be guaranteed.

[0019] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on a convolutional neural network model, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on a lightweight convolutional neural network model.

[0020] In the above scheme, in intelligent childcare monitoring, the first abnormal behavior monitoring model uses the CNN model, which can accurately identify the initial signs of abnormal behavior of the target. The second abnormal behavior monitoring model uses the Lightweight CNN model, which has lower computational complexity and faster processing speed while ensuring a certain recognition accuracy. After switching to the second video monitoring stage (higher resolution), the Lightweight CNN model can ensure a certain accuracy while reducing the consumption of computing resources. This process aims to save video clips with target abnormal behavior in high resolution. By adopting convolutional neural network models of different complexities, the system can flexibly adjust the allocation of computing resources according to actual needs, thereby ensuring a balance between the amount of stored data and the amount of recognition calculations.

[0021] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on multimodal behavior data fusion, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on single-modal behavior data.

[0022] In the above scheme, the first abnormal behavior monitoring model can capture the behavioral information of the childcare object more comprehensively by integrating behavioral data of multiple modalities. The complementarity of multimodal data enables the model to more accurately identify complex and subtle abnormal movements, reducing the information loss and misjudgment that may be caused by single modal data. Although the second abnormal behavior monitoring model is based only on unimodal behavioral data, its main purpose is to determine the end time of high-resolution video recording. Since there is a preset duration threshold as the identification cache period, the end time can be determined through unimodal behavioral data.

[0023] Optionally, the first abnormal behavior monitoring model is an abnormal behavior detection model determined based on body movement features and audio sound features, and the second abnormal behavior monitoring model is an abnormal behavior detection model based only on the body movement features.

[0024] In the above scheme, the first abnormal behavior monitoring model can more comprehensively capture the behavioral information of the childcare object by integrating body movement features and audio sound features. Body movement features provide spatial dimension information of the behavior, while audio sound features provide temporal dimension information of the behavior (such as the rhythm and intensity of the sound). The combination of the two can more accurately identify complex and subtle abnormal movements. For example, when infants and young children fall or have sudden illnesses, not only will there be abnormal body movements, but they will often be accompanied by abnormal crying or groaning. This multi-feature fusion method can significantly improve the accuracy of recognition. In the monitoring stage, after the potential abnormal behavior is identified through the multi-feature fusion of the first abnormal behavior monitoring model, the second abnormal behavior monitoring model quickly confirms the end time node based only on the body movement features.

[0025] Optionally, the preset duration threshold is determined by a preset behavior duration calibration list, and the preset behavior duration calibration list is used to establish a mapping relationship between the preset duration threshold and the target abnormal behavior.

[0026] In the above scheme, the preset behavior duration calibration list provides the system with an accurate duration threshold reference by recording and analyzing the time characteristics of various target abnormal behaviors in detail. This mechanism ensures that when the system determines abnormal behavior, it not only considers the type and characteristics of the behavior, but also fully considers the duration of the behavior, thereby improving the accuracy of the judgment. For example, for certain specific abnormal behaviors of infants and young children (such as continuous convulsions, long-term crying, etc.), the system can make accurate judgments based on the preset duration threshold to avoid misjudgment or missed judgments. Different abnormal behaviors often have different duration characteristics. Through the preset behavior duration calibration list, the system can set different duration thresholds for different types of abnormal behaviors, so as to more accurately adapt to the duration differences of different behaviors. This flexibility enables the system to make more accurate and reasonable judgments when faced with a variety of abnormal behaviors.

[0027] In a second aspect, the present application provides an intelligent childcare video monitoring and processing system, comprising:

[0028] An acquisition module, used for acquiring first video monitoring data of the nursery area to be monitored at a first resolution;

[0029] A processing module is used to switch to a second resolution to obtain second video monitoring data of the nursery area to be monitored when determining that there is abnormal target behavior in the first video monitoring data, wherein the second resolution is higher than the first resolution.

[0030] Optionally, the processing module is also used to determine that the duration of the absence of the target abnormal behavior in the second video monitoring data exceeds a preset duration threshold, and then switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored.

[0031] Optionally, a first abnormal behavior monitoring model is used to monitor the target abnormal behavior from the first video monitoring data, and a second abnormal behavior monitoring model is used to monitor the target abnormal behavior from the second video monitoring data, wherein the recognition accuracy of the first abnormal behavior monitoring model is higher than that of the second abnormal behavior monitoring model.

[0032] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on a convolutional neural network model, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on a lightweight convolutional neural network model.

[0033] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on multimodal behavior data fusion, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on single-modal behavior data.

[0034] Optionally, the first abnormal behavior monitoring model is an abnormal behavior detection model determined based on body movement features and audio sound features, and the second abnormal behavior monitoring model is an abnormal behavior detection model based only on the body movement features.

[0035] Optionally, the preset duration threshold is determined by a preset behavior duration calibration list, and the preset behavior duration calibration list is used to establish a mapping relationship between the preset duration threshold and the target abnormal behavior.

[0036] In a third aspect, the present application provides an electronic device, including:

[0037] processor; and,

[0038] A memory, configured to store executable instructions of the processor;

[0039] The processor is configured to perform any possible method described in the first aspect by executing the executable instructions.

[0040] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement any possible method described in the first aspect.

[0041] The intelligent childcare video monitoring and processing method and system provided in the present application obtains first video monitoring data of the childcare area to be monitored at a first resolution. If it is determined that target abnormal behavior exists in the first video monitoring data, the system switches to a second resolution with higher definition to obtain second video monitoring data of the childcare area to be monitored. This avoids unnecessary waste of data storage resources by improving the video resolution only when abnormal behavior is detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] 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.

[0043] Figure 1 This is a flowchart of an intelligent childcare video monitoring and processing method according to an example embodiment of the present application;

[0044] Figure 2 is a flowchart of an intelligent childcare video monitoring and processing method according to another exemplary embodiment of the present application;

[0045] Figure 3 is a structural schematic diagram of an intelligent childcare video monitoring and processing system according to an exemplary embodiment of the present application;

[0046] Figure 4 It is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present application.

[0047] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0048] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0049] Figure 1 FIG. 1 is a flow chart of an intelligent childcare video monitoring and processing method according to an exemplary embodiment of the present application. Figure 1 As shown, the method provided in this embodiment includes:

[0050] S101. Acquire first video monitoring data of a childcare area to be monitored at a first resolution.

[0051] In this step, the system first obtains the first video monitoring data of the nursery area to be monitored at a relatively low first resolution. The selection of the first resolution is intended to balance the real-time monitoring and the economy of data storage. Since normal monitoring data is the main data volume, the use of a relatively low resolution for preliminary monitoring can effectively reduce the amount of video data generated, thereby reducing the storage space occupied. It is worth mentioning that the collected monitoring data can be stored using an encrypted storage method in the prior art (for example, by existing methods such as password encryption, face encryption, and intranet encryption) to prevent data leakage and protect privacy.

[0052] Specifically, the system can collect video data in real time at a preset first resolution (such as 720P) through the camera equipment configured in the nursery area. The collected video data will be transmitted to the system's video processing module for processing and storage.

[0053] S102: If it is determined that there is abnormal target behavior in the first video monitoring data, switch to the second resolution to obtain second video monitoring data of the nursery area to be monitored.

[0054] If it is determined that there is abnormal target behavior in the first video monitoring data, switch to the second resolution to obtain the second video monitoring data of the nursery area to be monitored, wherein the second resolution is higher than the first resolution.

[0055] In this step, the system performs real-time analysis on the first video monitoring data to detect whether there is a target abnormal behavior. The target abnormal behavior can be one or more of a plurality of pre-set abnormal behaviors, such as falling, crying, and vigorous exercise of infants and young children.

[0056] In order to detect abnormal behavior, the system can use an abnormal behavior monitoring model based on machine learning. Specifically, the system can first use the training data to train the abnormal behavior monitoring model so that it can accurately identify the target abnormal behavior. During the training process, the system can extract key features from the video data (such as body movement features, audio sound features, etc.) and input these features into the abnormal behavior monitoring model for learning and optimization.

[0057] During the real-time monitoring process, the system analyzes the first video monitoring data frame by frame, extracts key features, and inputs these features into the trained abnormal behavior monitoring model. The model outputs the target behavior probability of the target abnormal behavior based on the input features. If the target behavior probability is greater than the preset abnormal behavior probability threshold, the system determines that the target abnormal behavior exists in the first video monitoring data.

[0058] Once abnormal behavior of the target is detected, the system immediately switches to the second resolution to obtain the second video monitoring data of the nursery area to be monitored. The second resolution is higher than the first resolution (such as 1080P or higher), which can provide clearer and more detailed video images. The design of this switching mechanism ensures that when abnormal behavior occurs, the system can quickly obtain high-quality video data, providing strong support for subsequent behavior analysis, event recording and evidence retention.

[0059] Specifically, the system can switch from the first resolution to the second resolution by adjusting the resolution setting of the camera device or performing resolution enhancement processing on the collected video data through the video processing module. The switched second video monitoring data will be transmitted to the video storage module of the system for storage for subsequent analysis and use.

[0060] In this embodiment, by acquiring first video monitoring data of the nursery area to be monitored with a first resolution, if it is determined that target abnormal behavior exists in the first video monitoring data, switching to a second resolution with higher definition to acquire second video monitoring data of the nursery area to be monitored, thereby avoiding unnecessary waste of data storage resources by improving the video resolution only when abnormal behavior is detected.

[0061] Figure 2 FIG. 1 is a flow chart of an intelligent childcare video monitoring and processing method according to another exemplary embodiment of the present application. Figure 2 As shown, the intelligent childcare video monitoring and processing method provided in this embodiment includes:

[0062] S201. Acquire first video monitoring data of a childcare area to be monitored at a first resolution.

[0063] In this step, the system first obtains the first video monitoring data of the nursery area to be monitored at a relatively low first resolution. The selection of the first resolution is intended to balance the real-time nature of monitoring with the economy of data storage. Since normal monitoring data is the main amount of data, by using a relatively low resolution for preliminary monitoring, the amount of video data generated can be effectively reduced, thereby reducing the storage space occupied.

[0064] S202: If it is determined that there is abnormal target behavior in the first video monitoring data, switch to the second resolution to obtain second video monitoring data of the nursery area to be monitored.

[0065] Optionally, the first abnormal behavior monitoring model is used to monitor the target abnormal behavior in the first video monitoring data, and the second abnormal behavior monitoring model is used to monitor the target abnormal behavior from the second video monitoring data, wherein the recognition accuracy of the first abnormal behavior monitoring model is higher than the recognition accuracy of the second abnormal behavior monitoring model.

[0066] Specifically, in the first video monitoring stage, the system uses a first abnormal behavior monitoring model with a high recognition accuracy. The system analyzes the first video monitoring data frame by frame, extracts key features (such as body movement features, audio sound features, etc.), and inputs these features into the first abnormal behavior monitoring model for recognition.

[0067] After switching to the second video monitoring stage, the system adopts the second abnormal behavior monitoring model with faster recognition speed but slightly lower recognition accuracy. This model has lower computational complexity and faster processing speed while ensuring a certain recognition accuracy. The system continues to analyze the second video monitoring data frame by frame, extract key features, and input these features into the second abnormal behavior monitoring model for rapid recognition. Although the recognition accuracy of this model is lower than that of the first model, it can still meet the monitoring needs because it has been determined by the first model before switching to the second resolution, and there is a preset time threshold as a buffer in the second monitoring stage.

[0068] In order to train the first and second abnormal behavior monitoring models, video data of the childcare area can be collected and abnormal behaviors can be annotated. The annotated data should include information such as the type, time and location of the abnormal behavior. The two models are trained separately using the annotated data. The first model focuses on improving the recognition accuracy, and a more complex network structure and more training rounds can be used; while the second model focuses on improving the recognition speed, and a lightweight network structure and efficient training strategy can be used.

[0069] During the training process, the system needs to tune the model parameters to ensure that the model can perform optimally in different monitoring stages.

[0070] In the first video monitoring stage, the system uses the first abnormal behavior monitoring model to perform preliminary monitoring and screening of video data. If potential target abnormal behavior is detected, the resolution switching mechanism is triggered.

[0071] After switching to the second video monitoring stage, the system uses the second abnormal behavior monitoring model to quickly identify the high-resolution video data. If the target abnormal behavior is not detected within a certain period of time (i.e. within the preset time threshold), it will automatically switch back to the first resolution for subsequent monitoring.

[0072] In one possible implementation, the first abnormal behavior monitoring model is an abnormal action detection model determined based on a convolutional neural network model, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on a lightweight convolutional neural network model.

[0073] Based on the previous examples, this embodiment provides specific implementations of the first abnormal behavior monitoring model and the second abnormal behavior monitoring model, namely, respectively using an abnormal action detection model based on a convolutional neural network (CNN) and an abnormal action detection model based on a lightweight convolutional neural network (such as MobileNet, EfficientNet, etc.).

[0074] Among them, the first abnormal behavior monitoring model adopts a deep convolutional neural network structure, which includes multiple convolutional layers, pooling layers, fully connected layers, and activation function layers. The convolutional layer is used to extract spatial features in the video frame, such as edges and textures; the pooling layer is used to reduce the dimension of the feature map and reduce the amount of calculation; the fully connected layer is used to map the extracted features to the probability space of abnormal behavior. The activation function (such as ReLU) is used to increase the nonlinear ability of the model and improve the expressiveness of the model. The model can be trained using large-scale annotated video data of the nursery area, and the data set contains many types of abnormal behaviors. The cross entropy loss function is used as the optimization target, and the model parameters are updated through the back propagation algorithm. Data enhancement techniques (such as rotation, scaling, cropping, etc.) are used to increase the diversity of training data and improve the generalization ability of the model. The training process of the model is optimized by adjusting hyperparameters such as learning rate and batch size.

[0075] In the first video monitoring stage, the system processes the video data frame by frame and inputs each frame into the first abnormal behavior monitoring model. The model outputs the probability of the target abnormal behavior in each frame, and the system determines whether to trigger the resolution switching mechanism based on the preset probability threshold.

[0076] As for the second abnormal behavior monitoring model, it can be an abnormal action detection model based on a lightweight convolutional neural network (such as MobileNet). The second abnormal behavior monitoring model adopts a lightweight convolutional neural network structure, such as MobileNet, EfficientNet, etc. These network structures use technical means such as deep separable convolution and bottleneck structure to greatly reduce the computational complexity and parameter amount of the model while ensuring a certain recognition accuracy. The model also contains convolutional layers, pooling layers, fully connected layers, and activation function layers, but the specific structure and parameter settings are different from those of the first abnormal behavior monitoring model. The second model is trained using the same training data and optimization objectives as the first abnormal behavior monitoring model. Due to the characteristics of the lightweight network structure, the training process of the second model is more efficient and converges faster. Data enhancement technology and hyperparameter adjustment strategies are also used to optimize the training process of the model.

[0077] In the second video monitoring stage, the system processes the high-resolution video data frame by frame and inputs each frame into the second abnormal behavior monitoring model. The model quickly outputs the probability of the target abnormal behavior in each frame, and the system determines whether to switch back to the first resolution or trigger the abnormal alarm mechanism based on the preset probability threshold and preset duration threshold.

[0078] In the first video monitoring stage, the system processes video data in real time and determines whether there is potential target abnormal behavior based on the output of the first abnormal behavior monitoring model. If potential abnormal behavior is detected, the system immediately switches to the second video monitoring stage and uses the second abnormal behavior monitoring model for high-resolution video monitoring. In the second video monitoring stage, the system continuously monitors video data and determines whether to switch back to the first resolution or trigger an abnormal alarm based on the output of the second abnormal behavior monitoring model and the preset duration threshold.

[0079] Furthermore, the first abnormal behavior monitoring model is an abnormal action detection model determined based on multimodal behavior data fusion, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on unimodal behavior data.

[0080] In one possible implementation, to determine the presence of target abnormal behavior in the first video monitoring data, a skeleton key point sequence of the first video monitoring data can be obtained, and a spatiotemporal graph convolution is performed on the skeleton key point sequence to determine a skeleton feature vector. At the same time, an audio signal of the first video monitoring data is obtained to extract an acoustic feature vector from the audio signal, and then a fused feature vector is generated based on the skeleton feature vector and the acoustic feature vector. The fused feature vector is input into a timing model to output a target behavior probability of the presence of target abnormal behavior. If it is determined that the target behavior probability is greater than a preset abnormal behavior probability threshold, it is determined that the target abnormal behavior exists in the first video monitoring data.

[0081] Specifically, it can be to obtain the first video monitoring data The skeleton key point sequence of the parent object in the frame ,in, For the The 3D coordinates of the preset joints, The total number of preset joints for the preset joint template.

[0082] Specifically, first video monitoring data of the nursery area to be monitored is acquired at a first resolution. Subsequently, computer vision technology, especially a human posture estimation algorithm based on deep learning, is used to extract a sequence of skeletal key points of the nursery object from each frame of the first video monitoring data. These key points usually correspond to the main joints of the human body, such as the head, shoulders, elbows, wrists, hips, knees and ankles. In this embodiment, we assume that the total number of preset joints of the preset joint template is n, then for each frame of video data, we can obtain a sequence of skeletal key points containing n three-dimensional coordinates, where each coordinate represents the position of a preset joint in three-dimensional space.

[0083] Using formula (1), and the skeleton key point sequence Perform spatiotemporal graph convolution to determine the skeleton feature vector , where formula (1) is:

[0084]

[0085] in, is the spatiotemporal graph convolution weight matrix, For the A spatiotemporal adjacency matrix, is the ReLU activation function, is the order of spatiotemporal graph convolution;

[0086] Specifically, the extracted skeleton key point sequence can be processed using spatiotemporal graph convolution to determine the skeleton feature vector. Spatiotemporal graph convolution is a convolution operation specifically used to process graph structure data, which can capture spatial and temporal dependencies at the same time. In this embodiment, formula (1) is used to calculate the skeleton feature vector, which includes parameters such as the spatiotemporal graph convolution weight matrix, the spatiotemporal adjacency matrix, the ReLU activation function, and the spatiotemporal graph convolution order. Through this step, the system can convert the original skeleton key point sequence into a more expressive feature vector, providing a basis for subsequent behavior analysis.

[0087] For formula (1), the skeleton key point sequence is processed by spatiotemporal graph convolution to effectively extract the body movement features of the child. This process not only considers the spatial relationship between the skeleton points, but also incorporates the information of the time dimension, making the extracted feature vector more comprehensive and accurate. The application of spatiotemporal graph convolution can capture subtle changes and trends in body movements, thereby improving the accuracy of identifying abnormal behaviors. This is especially important for monitoring the behavior of infants and young children, because their movements are often more complex and changeable.

[0088] The above-mentioned spatiotemporal graph convolution weight matrix can be determined by following the steps below:

[0089] Initialize the weight matrix: Before training begins, the weight matrix is ​​randomly initialized or uses some initialization strategy (such as Xavier initialization, He initialization, etc.).

[0090] Forward propagation: Multiply the input feature matrix by the weight matrix and obtain the output feature matrix through the activation function.

[0091] Calculate loss: According to the task objective (such as classification, regression, etc.), calculate the difference between the predicted value and the true value to obtain the loss function value.

[0092] Backpropagation: Calculate the gradient of the loss function with respect to the weight matrix through the chain rule, and use an optimization algorithm (such as SGD, Adam, etc.) to update the weight matrix.

[0093] Repeated training: Repeat the process of forward propagation, loss calculation, and backpropagation until the loss function converges or reaches a predetermined number of training rounds.

[0094] The above-mentioned spatiotemporal adjacency matrix can be determined by the following steps:

[0095] Spatial adjacency matrix: For skeletal joints, an adjacency matrix can be constructed based on their connection relationship. If node i and node j are directly connected, the corresponding element in the adjacency matrix is ​​1, otherwise it is 0.

[0096] Temporal adjacency matrix: In spatiotemporal graph convolution, the temporal adjacency matrix is ​​used to describe the connection relationship between nodes at different time steps. Nodes in adjacent time steps are considered to be directly connected, that is, the temporal adjacency matrix is ​​a matrix with 1 on the diagonal and 0 on the rest (or with a certain smoothing coefficient added).

[0097] Spatiotemporal adjacency matrix: The spatiotemporal adjacency matrix combines the information of the spatial adjacency matrix and the temporal adjacency matrix, which can be directly added or multiplied to obtain the final spatiotemporal adjacency matrix.

[0098] The above-mentioned spatiotemporal graph convolution order can be determined by following the steps below:

[0099] Choose the appropriate order according to the needs of the specific task. For example, in the action recognition task, it may be necessary to capture the dependencies between distant skeletal joints, so a higher order can be selected. The higher the order, the higher the computational cost. Therefore, a trade-off needs to be made between model performance and computational cost.

[0100] The performance of the model under different orders can be compared experimentally, and the order with the best performance can be selected. The order can also be determined by using a heuristic method based on the structure of the graph and the characteristics of the task. For example, in a bone joint point graph, the appropriate order can be determined based on the physiological structure of the human skeleton. The dynamic order method can also be used, that is, different orders are used at different time steps or different nodes.

[0101] Next, obtain the first video monitoring data The audio signal of the object in the frame is used to extract the acoustic feature vector from the audio signal .

[0102] Specifically, the audio signal of the childcare object is also extracted from each frame of the first video monitoring data. These audio signals contain various sounds made by the childcare object, such as speaking, laughing, crying, etc. In order to extract useful acoustic features from these audio signals, the system uses short-time Fourier transform to generate a Mel-spectrogram and determine the acoustic feature vector. These acoustic feature vectors can reflect the changes in the frequency and time of the audio signal, providing additional information for subsequent behavioral analysis.

[0103] Using formula (2), and according to the bone feature vector and the acoustic eigenvector Determining attention weights , where formula (2) is:

[0104]

[0105] in, is the query matrix, is the key matrix, is the feature dimension.

[0106] Specifically, after obtaining the skeleton feature vector and the acoustic feature vector, the system uses formula (2) to calculate the attention weight. The attention mechanism can determine the importance of each feature based on the correlation between the skeleton feature vector and the acoustic feature vector.

[0107] For formula (2), the attention mechanism can be used to calculate the attention weight between the skeletal feature vector and the acoustic feature vector to achieve effective evaluation of feature importance. This step can highlight the role of key features in behavior recognition and reduce the interference of noise and irrelevant information. By introducing attention weights, the system can focus more on features that are important for behavior recognition, thereby improving the robustness and stability of recognition.

[0108] The query matrix and the key matrix may be determined by the following steps:

[0109] Initialization: The query matrix and key matrix are usually obtained by random initialization. Before model training begins, you can use initialization methods such as Xavier or He to initialize these matrices according to the dimensions of the input features.

[0110] Feature extraction: The features of the query matrix and the key matrix are extracted from the bone feature vector and the acoustic feature vector. Therefore, it is necessary to first use a deep learning model (such as a convolutional neural network) to process the bone data and the audio data respectively to extract high-level feature representations.

[0111] Dimension matching: The dimensions of the query matrix and the key matrix need to match for subsequent dot product operations. If the extracted feature dimensions do not match, the dimensions can be adjusted through linear transformations (such as fully connected layers).

[0112] Training updates: During model training, the query matrix and key matrix are updated through the back-propagation algorithm. Based on the attention weights calculated by the attention mechanism and the loss function of the model, the gradients of these matrices can be calculated and their values ​​can be updated using optimization algorithms (such as Adam, SGD, etc.).

[0113] The determination of the above-mentioned feature dimensions can be carried out according to the following steps:

[0114] In intelligent childcare monitoring, the goal is to accurately identify abnormal behaviors of childcare objects, which requires a higher feature dimension to capture more detailed information. The higher the feature dimension, the higher the complexity of the model and the greater the amount of computation. Therefore, a trade-off needs to be made between model performance and computational cost.

[0115] The performance of the model under different feature dimensions can be verified through experiments. In practical applications, multiple candidate values ​​of feature dimensions can be set, and the models can be trained separately, and their performance on the validation set can be compared.

[0116] In addition to experimental verification, you can also refer to the experience and research results in related fields to determine the feature dimensions. For example, in the study of attention mechanisms, there are some commonly used feature dimension settings that can be used as reference.

[0117] Next, using formula (3), and according to the bone feature vector and the acoustic eigenvector Generate fused feature vector , where formula (3) is:

[0118]

[0119] For the above formula (3), the skeletal feature vector and the acoustic feature vector are fused to generate a fused feature vector containing richer information. This process achieves a comprehensive description of the behavior of the child, including both body movements and sound features, providing a more accurate and comprehensive information basis for subsequent abnormal behavior judgment. The generation of fused feature vectors enables the system to comprehensively utilize information from multiple modalities for behavior recognition, thereby enhancing the recognition capability and accuracy.

[0120] The fused feature vector Input to the time series model to output the probability of target behavior with target abnormal behavior , the timing model includes formula (4), which is:

[0121]

[0122] in, is the weight matrix of the fully connected layer, is the bias term of the fully connected layer.

[0123] To determine the above weight matrix and bias term, you can refer to the following steps:

[0124] First, the weight matrix and bias term need to be initialized before the model training begins. Common initialization methods include random initialization, zero initialization, normal distribution initialization, Xavier initialization, etc. In the above scheme, considering the processing ability of the time series model for long sequence data and the need to avoid gradient disappearance or explosion, Xavier initialization or similar strategies can be used to adaptively adjust the scale of the initial weights according to the number of input and output neurons, which helps to keep the distribution of activation values ​​and gradients in the entire network relatively stable, thereby improving the training efficiency and convergence of the model.

[0125] Once the weight matrix and bias terms are initialized, they will be continuously adjusted and optimized during the model training process through the back-propagation algorithm. The training data includes a large amount of annotated video monitoring data, which contains various behavior scenarios of the childcare objects, including normal and abnormal behaviors. This data is used to calculate the loss function, which measures the difference between the target behavior probability predicted by the model and the actual label.

[0126] In order to minimize the loss function, optimization algorithms such as stochastic gradient descent (SGD), Adam, RMSprop, etc. are usually used to update the weight matrix and bias terms. These algorithms gradually adjust the parameter values ​​according to a certain learning rate based on the gradient information of the loss function to the parameters until the performance of the model on the validation set reaches the optimal or converges.

[0127] During the training process, in order to prevent the model from overfitting to the training data and reducing the generalization ability, a regularization strategy needs to be introduced. Common regularization methods include L1 regularization, L2 regularization, Dropout, etc. In this embodiment, an L2 regularization term may be applied to the weight matrix, that is, the sum of the squares of the weights multiplied by a regularization coefficient is added to the loss function, thereby encouraging the model to learn smaller weight values, reducing the complexity of the model and improving the generalization ability.

[0128] Finally, the final determination of the weight matrix and bias term also depends on the selection of a series of hyperparameters, such as learning rate, batch size, regularization coefficient, number of training rounds, etc. The selection of these hyperparameters has an important impact on the performance of the model and usually needs to be tuned through strategies such as cross-validation, grid search or random search.

[0129] For the above formula (4), the fused feature vector is processed by the time series model to output the target behavior probability of the target abnormal behavior. This process realizes the prediction and judgment of the behavior trend of the childcare object, and can timely discover and respond to abnormal behavior. The application of the time series model enables the system to quickly output the behavior probability, thereby improving the response speed and processing efficiency to abnormal behavior. This is particularly important for the monitoring of infant and young child care environments, because timely response and processing can maximize the safety and health of infants and young children.

[0130] In the above scheme, by obtaining the skeleton key point sequence of the childcare object in each frame of the first video monitoring data, the body movement information of the childcare object can be accurately captured. The extraction of the skeleton key point sequence fully considers the characteristics of infant behavior, such as large movement amplitude and fast change, thereby ensuring the complete capture of key information.

[0131] Secondly, the skeleton key point sequence is processed using spatiotemporal graph convolution to determine the skeleton feature vector. Spatiotemporal graph convolution can simultaneously consider information in both time and space dimensions, thereby more comprehensively reflecting the action characteristics of the childcare object. The application of this step significantly improves the accuracy and robustness of behavior recognition, providing strong support for subsequent behavior judgment.

[0132] Furthermore, the acoustic feature vector is extracted from the audio signal and combined with the skeleton feature vector, and the fused feature vector is determined using the attention mechanism. The introduction of the acoustic feature vector provides an additional source of information for behavior recognition, such as the rhythm and intensity of the sound. This information complements the body movement features and together constitutes a comprehensive description of the behavior. The application of the attention mechanism further highlights the role of key features in behavior recognition and improves the accuracy of recognition.

[0133] Finally, the fused feature vector is input into the time series model to output the target behavior probability of the target abnormal behavior. The time series model can take into account the time series characteristics of the behavior, so as to predict and determine the changing trend of the behavior. By comparing the target behavior probability with the preset abnormal behavior probability threshold, it is possible to accurately determine whether the target abnormal behavior exists in the first video monitoring data.

[0134] If the target behavior probability is determined Greater than the preset abnormal behavior probability threshold , it is determined that there is target abnormal behavior in the first video monitoring data.

[0135] Specifically, the system inputs the fused feature vector into a time series model (such as a recurrent neural network, a long short-term memory network or a Transformer, etc.) to output the target behavior probability of the target abnormal behavior. The time series model can capture the temporal dependency of the feature vector, thereby more accurately identifying abnormal behavior. In this embodiment, formula (4) (which includes the weight matrix and bias term of the fully connected layer) is used to calculate the target behavior probability. If the target behavior probability is greater than the preset abnormal behavior probability threshold, the system determines that the target abnormal behavior exists in the first video monitoring data and triggers the subsequent high-resolution monitoring process.

[0136] Furthermore, in determining the probability of the target behavior Greater than the preset abnormal behavior probability threshold Before, we can also use formula (5) and according to the The skeleton key point sequence of the parent object in the frame , No. The skeleton key point sequence of the parent object in the frame and The audio signal in the frame determines the above abnormal behavior probability threshold , thereby realizing dynamic adjustment of the probability threshold, where formula (5) is:

[0137]

[0138] in, is the preset basic judgment threshold, is the first dynamic threshold adjustment coefficient, is the second dynamic threshold adjustment coefficient, The skeleton key point sequence of the parent object in the frame The coordinates of the i-th key point in , No. The skeleton key point sequence of the parent object in the frame The coordinates of the i-th key point in , It is the energy ratio of the target sound to the background noise in the audio signal.

[0139] Among them, the above-mentioned preset basic judgment threshold can be the distribution of the statistical normal behavior feature vector in the training set (such as mean ± 3 times the standard deviation), and the threshold is set to cover more than 90% of normal samples to reduce false positives. As for the first dynamic threshold adjustment coefficient and the second dynamic threshold adjustment coefficient, the characteristic differences between normal behavior and abnormal behavior can be found by analyzing the infant behavior data set, so as to determine the appropriate threshold and adjustment coefficient. It can also be in a laboratory environment to simulate various behavioral scenarios of infants and young children, and by adjusting the threshold and adjustment coefficient, observe the changes in the recognition effect, so as to find the optimal parameter combination. It can also be based on the knowledge and experience of experts in the field to preliminarily set the threshold and adjustment coefficient, and fine-tune it through actual monitoring data.

[0140] In the above scheme, a dynamic threshold adjustment mechanism is introduced, which dynamically adjusts the preset basic judgment threshold by comprehensively considering the changes in the skeleton key point sequence of the childcare object and the energy ratio in the audio signal. The technical effect of this step is that it can flexibly adjust the threshold for abnormal behavior judgment according to the actual behavior characteristics of the childcare object and the environmental noise conditions, thereby improving the accuracy and adaptability of the judgment.

[0141] Specifically, by calculating the change in the coordinates of key points in the skeleton key point sequence, the system can capture subtle changes in the body movements of the childcare object. This change reflects the activeness of the infant's behavior. When the behavior becomes abnormally intense or changes suddenly, the change will increase significantly. At the same time, the energy ratio of the target sound to the background noise in the audio signal provides information about the sound dimension. When the infant makes abnormal sounds (such as crying or screaming), the energy ratio will change.

[0142] Combining these two pieces of information, the system can more accurately determine whether the infant is in an abnormal state. When both behavioral changes and sound characteristics indicate that there may be an abnormality, the dynamic threshold will be lowered, making the system more sensitive to abnormal behavior. Conversely, when the behavior is stable and the sound is normal, the dynamic threshold will be relatively higher, reducing the possibility of misjudgment.

[0143] In addition, this dynamic threshold adjustment mechanism also enhances the robustness of the system. In practical applications, factors such as environmental noise and light changes may interfere with monitoring data and lead to misjudgment. By introducing dynamic threshold adjustment, the system can adaptively adjust the judgment threshold according to real-time monitoring data, thereby reducing the impact of these interference factors on the judgment results.

[0144] Furthermore, this dynamic adjustment mechanism also helps improve the overall performance of the system. In the process of abnormal behavior monitoring, the system needs to balance the judgment accuracy and computing resource consumption. By dynamically adjusting the threshold, the system can reduce unnecessary calculations and data storage as much as possible while ensuring the accuracy of judgment, thereby improving the system's operating efficiency and resource utilization.

[0145] In addition, the above-mentioned acoustic feature vector is extracted from the audio signal , it can include:

[0146] Perform a short-time Fourier transform on the audio signal to generate a mel-spectrogram ;

[0147] Using formula (6), and according to the Mel spectrum Determine the acoustic eigenvector , where formula (6) is:

[0148]

[0149] in, is the maximum pooling operation, It is a one-dimensional convolution operation.

[0150] In the above scheme, the Mel spectrogram is generated by performing a short-time Fourier transform on the audio signal. This step converts the audio signal from the time domain to the frequency domain and performs nonlinear compression on the frequency scale based on the auditory characteristics of the human ear, thereby more accurately reflecting the key features of the audio signal. The Mel spectrogram not only retains the time information of the audio signal, but also highlights the frequency components that are more relevant to human auditory perception.

[0151] Next, the Mel-spectrogram is further processed using a combination of a maximum pooling operation and a one-dimensional convolution operation (as shown in formula (5)) to determine the acoustic feature vector. The maximum pooling operation can reduce the dimension of the data and the amount of calculation while retaining the most important feature information. The one-dimensional convolution operation can extract local features and achieve effective traversal and feature extraction of the entire Mel-spectrogram by sliding the convolution kernel. This combined operation not only improves the efficiency of feature extraction, but also ensures that the extracted acoustic feature vector can accurately reflect the key information in the audio signal.

[0152] It can be seen that the acoustic feature vector extracted from the audio signal can be used as one of the important inputs of the subsequent abnormal behavior monitoring model, and combined with the limb movement features such as the skeleton key point sequence, it can form multimodal behavior data, which can be used to more comprehensively capture and identify abnormal behaviors in the childcare objects. The fusion of this multimodal data can significantly improve the accuracy and reliability of abnormal behavior monitoring.

[0153] S203: If it is determined that the duration of no abnormal behavior of the target in the second video monitoring data exceeds a preset duration threshold, switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored.

[0154] In this step, after switching to the second resolution for high-resolution video monitoring, the system will continue to analyze the second video monitoring data to detect whether there is abnormal behavior of the target. If the system does not detect abnormal behavior of the target within a certain period of time (i.e. within the preset time threshold), it will automatically switch back to the first resolution to continue to obtain the third video monitoring data.

[0155] Figure 3 FIG. 1 is a schematic diagram of the structure of an intelligent childcare video monitoring and processing system according to an exemplary embodiment of the present application. Figure 3 As shown, the intelligent childcare video monitoring and processing system 300 provided in this embodiment includes:

[0156] An acquisition module 310 is used to acquire first video monitoring data of the nursery area to be monitored at a first resolution;

[0157] The processing module 320 is used to switch to a second resolution to obtain second video monitoring data of the nursery area to be monitored when it is determined that there is abnormal target behavior in the first video monitoring data, wherein the second resolution is higher than the first resolution.

[0158] Optionally, the processing module 320 is also used to determine that the duration of the absence of the target abnormal behavior in the second video monitoring data exceeds a preset duration threshold, and then switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored.

[0159] Optionally, a first abnormal behavior monitoring model is used to monitor the target abnormal behavior from the first video monitoring data, and a second abnormal behavior monitoring model is used to monitor the target abnormal behavior from the second video monitoring data, wherein the recognition accuracy of the first abnormal behavior monitoring model is higher than that of the second abnormal behavior monitoring model.

[0160] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on a convolutional neural network model, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on a lightweight convolutional neural network model.

[0161] Optionally, the first abnormal behavior monitoring model is an abnormal action detection model determined based on multimodal behavior data fusion, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on single-modal behavior data.

[0162] Optionally, the first abnormal behavior monitoring model is an abnormal behavior detection model determined based on body movement features and audio sound features, and the second abnormal behavior monitoring model is an abnormal behavior detection model based only on the body movement features.

[0163] Optionally, the preset duration threshold is determined by a preset behavior duration calibration list, and the preset behavior duration calibration list is used to establish a mapping relationship between the preset duration threshold and the target abnormal behavior.

[0164] Figure 4 is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present application. Figure 4 As shown, an electronic device 400 provided in this embodiment includes: a processor 401 and a memory 402; wherein:

[0165] The memory 402 is used to store computer programs, and the memory may also be a flash memory.

[0166] The processor 401 is used to execute the execution instructions stored in the memory to implement each step in the above method. For details, please refer to the relevant description in the above method embodiment.

[0167] Optionally, the memory 402 may be independent or integrated with the processor 401 .

[0168] When the memory 402 is a device independent of the processor 401, the electronic device 400 may further include:

[0169] The bus 403 is used to connect the memory 402 and the processor 401 .

[0170] This embodiment further provides a readable storage medium, in which a computer program is stored. When at least one processor of an electronic device executes the computer program, the electronic device executes the methods provided in the above-mentioned various implementation modes.

[0171] This embodiment also provides a program product, which includes a computer program stored in a readable storage medium. At least one processor of the electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device implements the methods provided in the above various embodiments.

[0172] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the claims.

[0173] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. An intelligent childcare video monitoring and processing method, characterized in that: include: Acquire first video monitoring data of the nursery area to be monitored at a first resolution; If it is determined that there is abnormal behavior of the target in the first video monitoring data, switching to a second resolution to obtain second video monitoring data of the nursery area to be monitored, wherein the second resolution is higher than the first resolution; If it is determined that the duration of the target abnormal behavior not existing in the second video monitoring data exceeds the preset duration threshold, switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored.

2. The intelligent childcare video monitoring and processing method according to claim 1 is characterized in that: The first abnormal behavior monitoring model is used to monitor the target abnormal behavior from the first video monitoring data, and the second abnormal behavior monitoring model is used to monitor the target abnormal behavior from the second video monitoring data, wherein the recognition accuracy of the first abnormal behavior monitoring model is higher than the recognition accuracy of the second abnormal behavior monitoring model.

3. The intelligent childcare video monitoring and processing method according to claim 2 is characterized in that: The first abnormal behavior monitoring model is an abnormal action detection model determined based on a convolutional neural network model, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on a lightweight convolutional neural network model.

4. The intelligent childcare video monitoring and processing method according to claim 2 is characterized in that: The first abnormal behavior monitoring model is an abnormal action detection model determined based on multimodal behavior data fusion, and the second abnormal behavior monitoring model is an abnormal action detection model determined based on single-modal behavior data.

5. The intelligent childcare video monitoring and processing method according to claim 4 is characterized in that: The first abnormal behavior monitoring model is an abnormal behavior detection model determined based on body movement features and audio sound features, and the second abnormal behavior monitoring model is an abnormal behavior detection model based only on the body movement features.

6. The intelligent childcare video monitoring and processing method according to any one of claims 3 to 5, characterized in that: The preset duration threshold is determined by a preset behavior duration calibration list, and the preset behavior duration calibration list is used to establish a mapping relationship between the preset duration threshold and the target abnormal behavior.

7. An intelligent childcare video monitoring and processing system, characterized in that: include: An acquisition module, used for acquiring first video monitoring data of the nursery area to be monitored at a first resolution; A processing module, configured to switch to a second resolution to obtain second video monitoring data of the nursery area to be monitored when determining that there is abnormal target behavior in the first video monitoring data, wherein the second resolution is higher than the first resolution; The processing module is also used to switch to the first resolution to obtain the third video monitoring data of the nursery area to be monitored when it is determined that the duration during which the target abnormal behavior does not exist in the second video monitoring data exceeds a preset duration threshold.

8. An electronic device, characterized in that: include: processor; as well as, A memory, configured to store executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

Images