Children room monitoring system and application method

Abstract

The invention discloses a children's room monitoring system and an application method, and relates to the technical field of monitoring systems. The device at least comprises an AI box, an image sensor, a sound sensor, an auxiliary sensor, a loudspeaker and a display screen. The AI box is used for determining the position of a child through the acquired image and a human body detection model and performing safety judgment on the position; the image sensor at least comprises a camera. Through the arrangement of a corresponding system, the child monitoring function is started and closed based on gestures, parents can be liberated through a more intelligent control mode, image data and sound data do not leave equipment in the whole process, and the hidden danger of data leakage is completely eradicated; and laughter detection, cough sound statistics and typical disease cough detection are conveniently added in daily monitoring, so that daily monitoring of children is facilitated, and management of health of children can be assisted.

Description

Technical Field

[0001] This invention relates to the field of monitoring system technology, specifically to a children's room monitoring system and its application method. Background Technology

[0002] With societal development and an increasingly fast-paced lifestyle, it is becoming increasingly difficult for family members to accompany infants and toddlers around the clock. In this context, finding a method to ensure infant and toddler safety and provide real-time monitoring has become crucial. Camera-based infant and toddler monitoring systems were developed specifically to address this need. By monitoring the infant's living environment and physiological condition, they provide parents with real-time information, enabling them to detect and address potential safety hazards immediately.

[0003] However, most existing smart child monitoring products rely on sensor-collected information for simple information processing and logical judgment, making it difficult to identify and handle complex scenarios and meet the real-world child care needs. To address these issues, some products attempt to utilize deep learning algorithms for intelligent perception and decision-making, such as patent CN202210556558.2. However, deep learning algorithms are resource-intensive, and most products can only transmit the monitored images and sounds to a cloud server for inference before downloading the results back to the device. This means that monitoring functions fail entirely when the network is unstable. Even when the network is stable, there are concerns about data leakage.

[0004] Most child monitoring devices on the market use a combination of user devices and cloud services.

[0005] User equipment includes a high-definition camera, a sound sensor, an environmental sensor, and a wireless communication module. The high-definition camera provides real-time video streaming and image capture, and should support night vision for monitoring in low-light conditions. The sound sensor monitors the child's voice and movement to notify parents when the child cries. The environmental sensor monitors room temperature, humidity, and air quality to ensure a suitable environment in the child's room. The wireless communication module transmits data to a cloud server. In addition, there may be a display or mobile application allowing parents to easily access and control the monitoring system, receive cloud-based alarm information, etc.

[0006] The cloud service includes a cloud-based data processing platform with capabilities for real-time video stream processing, data analysis, and historical data storage and retrieval. Real-time video stream processing receives and stores live video streams transmitted from cameras, making them available for parents to view. Data analysis, through real-time analysis of sensor data, identifies children's behavioral patterns, such as sleeping, waking, and crying, and sends appropriate notifications based on the analysis results. Historical data storage and retrieval stores sensor data on a cloud server, allowing parents to access historical data when needed.

[0007] To protect the privacy of infants and families, data encryption and access control are employed during data upload and storage. Data encryption means that all data transmitted to the cloud server should be encrypted to prevent interception during transmission. Access control ensures that only authorized users can access video streams and sensor data. Applications should use secure login mechanisms, such as two-factor authentication.

[0008] The above solution is a popular technology for infant and toddler monitoring devices that can provide comprehensive infant and toddler monitoring services while protecting user privacy and data security to the greatest extent possible.

[0009] While cloud-based children's room monitoring systems have many advantages, such as easy access and stable data storage, they also have some potential drawbacks.

[0010] The activation process is not intelligent: With existing child monitoring systems, parents need to approach the screen and tap to activate the monitoring function or open the accompanying software. This adds to the burden for many parents who have just put their children to sleep and are already tired.

[0011] Privacy concerns: Any information transmitted over the internet is at risk of being illegally intercepted and used. Although many cloud service providers employ encryption technologies to protect data security, data breaches can still lead to the violation of children's and families' privacy.

[0012] Network dependency: Cloud-based monitoring systems require a stable network connection to function properly. Network problems or slow speeds may affect real-time video transmission and data uploads.

[0013] Data ownership: Using cloud services means your data is stored on the service provider's servers. While most service providers promise to protect user data security and privacy, users' control over their data is diminished to some extent.

[0014] Cost: Many cloud service providers charge for storage space and data transfer volume. If you require long-term video recording and storage, you may need to pay additional fees.

[0015] System Availability: The stability and availability of cloud services depend on the service provider. If the service provider's servers fail, the monitoring system may not function properly.

[0016] Therefore, a new solution is needed to address the above problems. Summary of the Invention

[0017] The purpose of this invention is to provide a children's room monitoring system and its application method.

[0018] To achieve the above objectives, the present invention provides the following technical solution: a children's room monitoring system, comprising at least an AI box, an image sensor, a sound sensor, an auxiliary sensor, a speaker, and a display screen;

[0019] The AI ​​box is used to determine the child's location using acquired images and a human detection model, and to assess the safety of the location.

[0020] The image sensor includes at least a camera, and the image sensor is used to obtain a clear image with details in combination with the structure in the AI ​​box in the absence of infrared fill light;

[0021] The sound sensor is used to capture sound;

[0022] The auxiliary sensor is used to supplement functions according to customer needs;

[0023] The speaker is used to play warning messages or audio preset by the user.

[0024] The display screen is for customer control purposes.

[0025] Preferably, the AI ​​box includes at least an ISP chip, a CPU, a storage unit, a sound classification model, an image detection model, and an edge computing device. The ISP chip is used to improve image clarity, the CPU is used for task scheduling, the edge computing device is used to provide computing power for model inference, the storage unit is used to store the model and the image, the storage unit is used to cooperate in implementing storage applications, and the sound classification model is used to preprocess the sound file into an MFCC image during the sound classification process before training the classification model.

[0026] Preferably, the computing power of the edge computing device is greater than 1T.

[0027] Preferably, the CPU is a rk3588 processing chip.

[0028] Preferably, the auxiliary sensors include at least a temperature sensor and a humidity sensor.

[0029] Preferably, the sound classification model uses a lightweight classification network such as MobileNetV3.

[0030] Preferably, the image detection model uses a lightweight network from the YOLO series.

[0031] A method for applying a children's room monitoring system, for the above-mentioned purposes, includes at least the following steps:

[0032] S1: The user places the devices included in the system at a suitable angle to ensure that the camera can observe the area where the child is frequently present;

[0033] S2: A gesture is predefined in the AI ​​box, which corresponds to the activation of the monitoring function. The predefined gesture and the child are detected by the detection model deployed in the AI ​​box. At the same time, a sound classification model is deployed in the AI ​​box to classify crying, mild cough, cough of typical disease type and laughter.

[0034] S3: The AI ​​box acquires images from the image sensor and sends them to the detection model. When a predefined gesture is detected, the child monitoring function is activated and the parents are prompted. The prompt includes at least one of the following: a sound is played and the screen flashes to inform the parents that monitoring has been activated.

[0035] S4: After the child monitoring function is activated, the AI ​​box acquires images from the image sensor and sends them to the detection model. When it detects a drastic change in the child's position or detects crying, it automatically plays the voice or music recorded by the guardian to soothe the child. If a predefined gesture is detected again, the child monitoring function will be deactivated.

[0036] S5: During device operation, the AI ​​box acquires sound from the sound sensor and sends it to the sound classification model. When the sound classification result is laughter, it automatically saves the corresponding sound and image as a memory recording. When the detection result is crying and the child monitoring function is enabled, it automatically plays the voice or music recorded by the guardian to soothe the child. When a typical disease-related cough is detected, it plays a medical reminder. When a mild cough is detected, it records the time of the cough to count the frequency of coughs.

[0037] Compared with the prior art, the beneficial effects of the present invention are:

[0038] This invention enables parents to activate and deactivate child monitoring functions based on gestures through corresponding system settings. This more intelligent control method frees up parents, ensuring that image and sound data remain on the device throughout the process, completely eliminating the risk of data leakage. It also facilitates the addition of laughter detection, cough statistics, and typical disease-related cough detection in daily monitoring, making it convenient for daily monitoring of children while also assisting in the management of children's health. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the entire system of the present invention. Detailed Implementation

[0041] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0042] Please see Figure 1 A children's room monitoring system includes at least an AI box, an image sensor, a sound sensor, an auxiliary sensor, a speaker, and a display screen;

[0043] The AI ​​box is used to determine the child's location by acquiring images and human detection models, and to make a safety assessment of the location;

[0044] The image sensor includes at least a camera. The image sensor is used to acquire clear images with details in the form of a camera and an ISP chip in the AI ​​box, in the absence of infrared illumination.

[0045] Sound sensors are used to capture sound;

[0046] Auxiliary sensors are used to supplement functions according to customer needs;

[0047] The speaker is used to play warning messages or user-defined audio.

[0048] The display screen is for customer control purposes.

[0049] The AI ​​box includes at least an ISP chip, CPU, storage unit, sound classification model, image detection model, and edge computing device. The ISP chip is used to improve image clarity, the CPU is used for task scheduling, the edge computing device is used to provide model inference, the storage unit is used to store the model and the image computing power, the storage unit is used to cooperate in realizing storage applications, and the sound classification model is used to preprocess the sound file into an MFCC image during the sound classification process, and then train the classification model.

[0050] Edge computing devices have a computing power greater than 1T;

[0051] The CPU is a rk3588 processing chip;

[0052] The auxiliary sensors include at least a temperature sensor and a humidity sensor;

[0053] The sound classification model uses lightweight classification networks such as MobileNetV3;

[0054] The image detection model uses lightweight YOLO networks.

[0055] Addressing the shortcomings mentioned in the background technology, a children's room monitoring system is proposed. Specifically, this system is an AI-based, leak-proof, automatic children's room monitoring system. This system includes hardware such as a high-performance AI box, image sensors, sound sensors, and speakers, as well as software such as a human detection model, a sound analysis model, and data processing logic. The image sensor acquires real-time images of the children's room, and the sound sensor acquires real-time sound data. The AI ​​box determines the child's location based on the acquired images and the human detection model, and performs a safety assessment of that location. If the child exceeds the safe location set by the guardian, a preset alert is issued. The system detects the child's state using sound data and a sound classification model. When crying is detected, a custom sound can be played to soothe the child. When continuous barking-like coughing is detected, a medical reminder is issued. The integrated monitoring process requires no cloud server or internet connection. Image and sound data remain within the AI ​​box, fundamentally eliminating the risk of privacy leaks.

[0056] Example 2:

[0057] A method for applying a children's room monitoring system, used in the above embodiments, includes at least the following steps:

[0058] S1: The user places the devices included in the system at a suitable angle to ensure that the camera can observe the area where the child is frequently present;

[0059] S2: A gesture is predefined in the AI ​​box, which corresponds to the activation of the monitoring function. The predefined gesture and the child are detected by the detection model deployed in the AI ​​box. At the same time, a sound classification model is deployed in the AI ​​box to classify crying, mild cough, cough of typical disease type and laughter.

[0060] S3: The AI ​​box acquires images from the image sensor and sends them to the detection model. When a predefined gesture is detected, the child monitoring function is activated and the parents are prompted. The prompt includes at least one of the following: a sound is played and the screen flashes to inform the parents that monitoring has been activated.

[0061] Whether to play sound or flicker on the screen can be set in the display settings, and the sound volume and flicker brightness can also be adjusted.

[0062] S4: After the child monitoring function is activated, the AI ​​box acquires images from the image sensor and sends them to the detection model. When it detects a drastic change in the child's position or detects crying, it automatically plays the voice or music recorded by the guardian to soothe the child. If a predefined gesture is detected again, the child monitoring function will be deactivated.

[0063] S5: During device operation, the AI ​​box acquires sound from the sound sensor and sends it to the sound classification model. When the sound classification result is laughter, it automatically saves the corresponding sound and image as a memory recording. When the detection result is crying and the child monitoring function is enabled, it automatically plays the voice or music recorded by the guardian to soothe the child. When a typical disease-related cough is detected, it plays a medical reminder. When a mild cough is detected, it records the time of the cough to count the frequency of coughs.

[0064] The monitor displays the monitoring logs, including videos of fond memories saved from laughter and the frequency of coughs.

[0065] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A children's room monitoring system, characterized in that: It includes at least an AI box, image sensor, sound sensor, auxiliary sensor, speaker, and display screen; The AI ​​box is used to determine the child's location using acquired images and a human detection model, and to assess the safety of the location. The image sensor includes at least a camera, and the image sensor is used to obtain a clear image with details in combination with the structure in the AI ​​box in the absence of infrared fill light; The sound sensor is used to capture sound; The auxiliary sensor is used to supplement functions according to customer needs; The speaker is used to play warning messages or audio preset by the user. The display screen is for customer control purposes.

2. The children's room monitoring system according to claim 1, characterized in that: The AI ​​box includes at least an ISP chip, a CPU, a storage unit, a sound classification model, an image detection model, and an edge computing device. The ISP chip is used to improve image clarity, the CPU is used for task scheduling, the edge computing device is used to provide computing power for model inference, the storage unit is used to store the model and the image, the storage unit is used to cooperate in implementing storage applications, and the sound classification model is used to preprocess the sound file into an MFCC image during the sound classification process before training the classification model.

3. A children's room monitoring system according to claim 2, characterized in that: The computing power of the edge computing device is greater than 1T.

4. A children's room monitoring system according to claim 2, characterized in that: The CPU is a rk3588 processing chip.

5. A children's room monitoring system according to claim 1, characterized in that: The auxiliary sensors include at least a temperature sensor and a humidity sensor.

6. A children's room monitoring system according to claim 2, characterized in that: The sound classification model uses lightweight classification networks such as MobileNetV3.

7. A children's room monitoring system according to claim 2, characterized in that: The image detection model uses a lightweight YOLO network.

8. A method for applying a children's room monitoring system, used in any one of the children's room monitoring systems according to claims 1-7, characterized in that: At least the following steps are included: S1: The user places the devices included in the system at a suitable angle to ensure that the camera can observe the area where the child is frequently present; S2: A gesture is predefined in the AI ​​box, which corresponds to the activation of the monitoring function. The predefined gesture and the child are detected by the detection model deployed in the AI ​​box. At the same time, a sound classification model is deployed in the AI ​​box to classify crying, mild cough, cough of typical disease type and laughter. S3: The AI ​​box acquires images from the image sensor and sends them to the detection model. When a predefined gesture is detected, the child monitoring function is activated and the parents are prompted. The prompt includes at least one of the following: a sound is played and the screen flashes to inform the parents that monitoring has been activated. S4: After the child monitoring function is activated, the AI ​​box acquires images from the image sensor and sends them to the detection model. When it detects a drastic change in the child's position or detects crying, it automatically plays the voice or music recorded by the guardian to soothe the child. If a predefined gesture is detected again, the child monitoring function will be deactivated. S5: During device operation, the AI ​​box acquires sound from the sound sensor and sends it to the sound classification model. When the sound classification result is laughter, it automatically saves the corresponding sound and image as a memory recording. When the detection result is crying and the child monitoring function is enabled, it automatically plays the voice or music recorded by the guardian to soothe the child. When a typical disease-related cough is detected, it plays a medical reminder. When a mild cough is detected, it records the time of the cough to count the frequency of coughs.

Images