Pet car intelligent management system integrating positioning and health data acquisition

Abstract

The invention belongs to the technical field of pet nursing, and provides a pet car intelligent management system integrating positioning and health data acquisition, and the system comprises an image monitoring module, a target marking module, an image processing module, a feature analysis module and a danger early warning module. The system utilizes a visual intelligent technology to accurately identify the current psychological state of a pet, monitors the physiological change of the pet in real time through visual detection, and can reduce irreversible psychological injury caused by a long-time high-voltage state in combination with control logic linkage of the pet car, so that the intelligence and the safety of the pet car in an automatic tracking mode are remarkably improved, and the pet car is convenient to use. According to the system, emotion-friendly accurate control over the pet can be achieved, and the comfort level and compliance of the pet in the interaction process can be improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of pet care, and particularly relates to a pet vehicle intelligent management system fusing positioning and health data collection. BACKGROUND

[0002] With the deepening of people's understanding of pets, the social status of pets is also constantly improving. Positioning and tracking of pets and analysis of the state of pets are very important to ensure the safety and health of pets. The combination of industrial vision and visual intelligence technology makes pet positioning more accurate. Positioning and tracking enable pet owners to know the location of pets in real time and prevent pets from getting lost. At the same time, visual detection technology is used to monitor the state of pets, which not only helps to evaluate the physical health of pets, such as exercise amount and rest pattern, but also provides important clues for the emotional state of pets, such as signs of anxiety or depression. Such monitoring can guide pet owners to adjust the daily activities and environment of pets to promote their physical and mental health. Further, continuous and comprehensive positioning and tracking of pets are crucial because they not only ensure the safety of pets and prevent pets from getting lost or encountering accidental injuries, but also help pet owners to know the location and activity state of pets in real time, so as to better manage and care for pets. Using AI vision technology for accurate analysis, the pet vehicle can adjust the tracking strategy in a timely manner according to the state of the pet to ensure that the pet receives reasonable attention and care. In a Chinese invention patent named "Pet tracking method, device, equipment and medium combined with UWB positioning and vision" with an application date of March 20, 2024 and a publication number of CN118283533B, according to pet vehicle position information, pet position information, small car motion parameters and pet target detection results, the pet vehicle is controlled to track the pet in real time, solving the problem that the existing technology cannot accurately and continuously position and track the pet. This method is suitable for indoor and outdoor environments and can realize real-time positioning and tracking of pets. However, when the pet is in a new environment, it is easy to produce stress reactions due to external stimuli. Visual measurement technology can further accurately evaluate the behavior and physiological reactions of pets. In a stress state, pets are abnormally sensitive to changes in the surrounding environment. At this time, if the pet vehicle continuously tracks the pet, it is easy to cause continuous contraction of the abdominal visceral blood vessels of the pet, accelerate the heart rate, and even cause sudden death of the pet due to continuous severe stress. SUMMARY

[0003] The purpose of the present application is to provide a pet vehicle intelligent management system fusing positioning and health data collection to solve one or more technical problems existing in the prior art and at least provide a beneficial choice or create conditions.

[0004] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a pet vehicle intelligent management system fusing positioning and health data collection is provided, which comprises the following modules:

[0005] an image monitoring module, configured to monitor activity track information of the pet in real time and acquire multiple frames of real-time images of the pet;

[0006] a target marking module, configured to perform target detection and tracking on the multiple frames of real-time images to obtain a pet target image sequence;

[0007] an image processing module, configured to pre-process the pet target image sequence to obtain pre-processed images;

[0008] a feature analysis module, configured to acquire feature information of the pet in the pre-processed images and calculate a risk coefficient of the feature information;

[0009] a risk warning module, configured to make a behavior judgment on the pet according to the risk coefficient of the feature information and issue a care reminder;

[0010] Further, the Internet of Things terminal module further comprises a sensor for monitoring data.

[0011] Further, the Internet of Things terminal module is connected with the image monitoring module, the target marking module, the image processing module, the feature analysis module and the risk warning module.

[0012] Further, the data monitoring module comprises a user data acquisition unit, a satellite positioning unit and a monitoring unit.

[0013] The user data acquisition unit is configured to acquire data information of the pet owner.

[0014] The satellite positioning unit is configured to acquire activity location information of the pet.

[0015] The monitoring unit is configured to monitor pet activities in real time through a pre-installed monitoring device.

[0016] An output end of the monitoring unit is connected with input ends of the user data acquisition unit and the satellite positioning unit.

[0017] Further, the Internet of Things terminal module comprises a state detection unit, a data transmission unit and a network control unit.

[0018] The state detection unit is configured to acquire position and speed information of the pet through a sensor.

[0019] The data transmission unit is configured to receive and transmit data according to an execution command.

[0020] The network control unit is configured to ensure data transmission in a state of normal network connection.

[0021] The output end of the control unit is connected with the input end of the pressure detection unit and the data transmission unit.

[0022] A pet car intelligent management method fusing positioning and health data collection, characterized by comprising the following steps:

[0023] S1, real-time monitoring of the activity track information of the pet and obtaining multiple real-time images of the pet;

[0024] S2, target detection and tracking of the multiple real-time images to obtain a pet target image sequence;

[0025] S3, preprocessing of the pet target image sequence to obtain preprocessed images;

[0026] S4, obtaining feature information of the pet in the preprocessed images and calculating a risk coefficient of the feature information;

[0027] S5, behavior judgment of the pet according to the risk coefficient of the feature information and issuing a care reminder.

[0028] Further, in S1, the specific method of real-time monitoring of the activity track information of the pet and obtaining multiple real-time images of the pet is that a camera preset in a pet activity scene is used to shoot real-time activity videos of the pet, the real-time activity videos are decomposed into multiple real-time images, and the multiple real-time images include position information of the pet, the position information including pet position information and pet car position information.

[0029] Further, in S2, the specific method of target detection and tracking of the multiple real-time images to obtain a pet target image sequence is that a pre-trained pet target detection model is used, specifically a basic pet detection model, the detection model extracts features and classifies images to output a bounding box, a category and a confidence score containing the pet, a preset target tracking algorithm is used to track the pet target in each real-time image, and the pet target image sequence and a pet target box position sequence are output as the target image sequence.

[0030] Further, the pet detection model is any one of YOLO8S (You Only Look Once), SSD (Single Shot MultiBox Detector) or Faster R-CNN.

[0031] Preferably, the pet detection model is YOLO8S.

[0032] The YoloV8s model is characterized by its ability to directly output the location and category information of the target in one forward propagation without the need for multiple scans of the image, resulting in high computational performance and faster detection speed for pet targets. The detection results will include the bounding box coordinates of the pet target, the category label, and its confidence in the image. This pet target detection process enables the system to identify and locate pets in real-time, providing key information to pet owners. This is a critical function for pet care systems, offering users a convenient and practical monitoring and tracking tool to ensure the health and safety of pets.

[0033] Further, in S3, the specific method of preprocessing the pet target image sequence to obtain the preprocessed image is: after the target image sequence is subjected to image filtering processing to remove Gaussian noise by Gaussian filtering and salt and pepper noise by median filtering, the image sequence after image filtering processing is subjected to de-duplication processing to remove duplicate images, and the preprocessed image is output.

[0034] Further, in S4, the feature information of the pet in the preprocessed image is obtained and the danger coefficient of the feature information is calculated. The position information in the preprocessed image is obtained as the feature information, the pet position information is denoted as P(i), the pet car position information is denoted as Q(i), the straight-line distance between the pet position information and the pet car position information is calculated as D(i), and if D(i) exceeds a preset threshold Th1, the direction and distance that the car needs to travel are calculated based on the pet car position information and the pet position information, so as to determine the motion parameters of the car, such as speed, direction of travel, and turning angle. In this way, the pet car can automatically navigate and move along the optimal path to approach the pet's position.

[0035] The time when the pet car starts to move is denoted as ST, and the time when the pet car stops moving is denoted as ED. The electrocardiogram waveform of the pet from ST to ED is obtained through the electrocardiogram sensor worn in the pet's collar, the signal strength of the electrocardiogram waveform is denoted as E(i), the entire electrocardiogram waveform is traversed, the entire electrocardiogram waveform is subjected to high-pass and low-pass filtering to remove electromyographic interference and baseline drift, and then the electrocardiogram waveform is subjected to wavelet transform to remove small signal interference, so as to ensure that the processed electrocardiogram waveform is smooth and free of burrs. The R wave is marked using the Pan-Tompkins algorithm, the peak of all R waves is denoted as PK(j), the time interval between PK(j) and PK(j+1) is denoted as T(j), j is the serial number of the time interval, all time intervals T(j) are traversed, and the formula calculate the risk coefficient, M is the number of time intervals, if RK(j+1)-RK(j)>RK(j)-RK(j-1) and T(j)>T(j+1) are satisfied, or RK(j+1)-RK(j)<RK(j)-RK(j-1) and T(j)<T(j+1) are satisfied, the time corresponding to PK(j) is recorded as the stress moment, and the pet car is controlled to stop moving;

[0036] Many current pet monitoring technologies are basically based on static behavior pattern recognition, and the stress state is judged by setting fixed parameter thresholds. When the parameters do not reach the threshold all the time, the judgment system often does not respond, resulting in delay and misjudgment in obtaining the stress moment. However, the above method provides more sensitive and real-time stress detection by dynamically calculating the risk coefficient and the trend of the electrocardiogram waveform, and can capture the subtle fluctuations in heart rate changes in a short time, thereby more accurately judging the stress moment of the pet.

[0037] When the pet is out, if the pet car continuously tracks the movement of the pet, it may cause the pet to have a stress reaction. Since the tracking behavior of the pet car is uninterrupted and close to the activity track of the pet, the pet will feel oppressed or threatened, thereby triggering anxiety or panic emotions. This change in emotion will activate the sympathetic nervous system, thereby causing the heart rate to accelerate. On the electrocardiogram waveform, it is usually manifested as a sudden shortening of the RR wave peak, indicating an abnormal increase in heart rate, and the heart rate is irregular. The above method marks the time point at which the stress reaction occurs to accurately control the tracking behavior of the pet car. When the pet has a stress reaction, the tracking is stopped in time, which can effectively alleviate the anxiety of the pet and help prevent the accumulation of psychological problems, thereby reducing the long-term negative impact on the physical and mental health of the pet. However, when the pet is in an excited state, similar RR wave peak mutation phenomena will occur on the electrocardiogram waveform in a short time. The above method cannot effectively distinguish between the anxiety and excitement states of the pet by observing the instantaneous heart rate change of the pet, thereby causing the pet car to lose track and the pet to be lost. To solve the above problems, the present application proposes the following method to distinguish between the excited state and the stress state of the pet by calculating the heart rate change in the duration, so as to accurately control the movement of the pet car:

[0038] All time intervals T(j) are traversed, and the difference between T(j+1) and T(j) is calculated as TO(j). All difference values TO(j) are traversed in turn. If TO(j+1)>TO(j) and TO(j)<TO(j-1), T(j) is marked as the fluctuation time start point. The remaining R wave peaks are searched from the R wave peak value corresponding to T(j) along the electrocardiogram waveform, and the difference between the remaining R wave peaks and the R wave peak corresponding to the time start point is calculated as DIF(k). k represents the serial number of the R wave peak searched from the R wave peak value corresponding to T(j) along the electrocardiogram waveform. The formula is calculating a risk coefficient, wherein N is the number of R-wave peaks retrieved backward from the R-wave peak value corresponding to T(j) along the electrocardiogram waveform, DIF_max represents the maximum difference between the R-wave peak retrieved backward and the R-wave peak corresponding to T(j), DIF_min represents the minimum difference between the R-wave peak retrieved backward and the R-wave peak corresponding to T(j), if RK(j) is less than zero, it is considered that the pet is in an excited state, the pet vehicle is controlled to continue tracking to prevent the pet from being lost, otherwise, it is considered that the pet is in a stress state, the pet vehicle is controlled to stop tracking to prevent the pet from being stressed.

[0039] The beneficial effect of the above steps is that by analyzing the change rule of the RR interval (R-R Interval) in the electrocardiogram waveform within the duration, the excited state and the stress state of the pet are effectively distinguished. In the excited state, the autonomic nervous system of the pet mainly shows temporary activation of the sympathetic nerve, acceleration of the heart rate, and shortening of the RR interval, but then shows a stable trend. In the above formula, partly tends to the left Therefore, RK(j) is less than zero in the excited state, and in the stress state, due to the disorderly regulation of the sympathetic-parasympathetic nervous system, the RR interval fluctuates greatly, and abnormal electrophysiological phenomena such as T-wave inversion, ST segment elevation or depression appear in the electrocardiogram waveform. In the above formula, partly obviously greater than This is because the false R wave (including the appearance of the T wave, the superposition of the P wave, or the strong muscle electrical interference due to the stress phenomenon) caused by the unstable peak value causes the difference between the local adjacent R wave peaks to be much greater than the normal range. Therefore, RK(j) is greater than zero in the stress state. The above method avoids the misidentification of the pet state in the instantaneous detection through the RR dynamic monitoring algorithm constructed within the duration, reduces the risk of the pet getting lost, and also avoids the psychological problems of the pet caused by the continuous tracking of the pet vehicle when the pet is in a stress state.

[0040] Further, in S5, the specific method of making a behavior judgment on the pet and issuing a care reminder according to the risk coefficient of the feature information is: according to the calculated risk coefficient RK(j) of the feature information, if RK(j) is less than zero, it is considered that the pet is in a stress state, the pet vehicle is controlled to stop tracking to prevent the pet from being stressed, and a care alarm is issued to remind the owner to take good care of the pet.

[0041] Beneficial effects: The present invention continuously collects the pet's ECG data while the pet car is tracking the pet, and calculates the pet's risk coefficient based on the RR interval dynamics system multidimensional filtering algorithm and time series fluctuation evaluation model, effectively suppressing local ECG abnormalities caused by non-continuous behaviors such as short-term excitement and exercise of the pet, and avoiding misjudgment of overall emotional judgment due to instantaneous emotional fluctuations. Through trend analysis of ECG waveforms and emotional state modeling, the system can accurately identify the pet's current psychological state (such as excitement, stress), and link with the control logic of the pet car. When the system identifies that the pet has a high stress risk or is emotionally unstable, it can actively adjust or suspend tracking behavior, thereby avoiding continuous exercise inducing or aggravating the pet's psychological burden, reducing irreversible psychological damage caused by long-term high-pressure state, significantly improving the intelligence and safety of the pet car in automatic tracking mode, and realizing emotionally friendly and precise control of the pet, which helps to improve the pet's comfort and compliance during the interaction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Shown is a flow chart of the pet stroller intelligent management system that integrates positioning and health data collection. DETAILED DESCRIPTION

[0043] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0044] Example 1:

[0045] Figure 1 Shown is a flow chart of the pet stroller intelligent management system that integrates positioning and health data collection.

[0046] Reference Figure 1 The present invention proposes an intelligent management system for pet strollers that integrates positioning and health data collection. The system includes the following modules:

[0047] Image monitoring module, used to monitor the pet's activity trajectory information in real time and obtain multiple frames of real-time images of the pet;

[0048] The target marking module is used to detect and track targets in multiple frames of real-time images to obtain a pet target image sequence;

[0049] An image processing module, configured to preprocess the pet target image sequence to obtain a preprocessed image;

[0050] A feature analysis module is used to obtain feature information of the pet in the pre-processed image and calculate the risk coefficient of the feature information;

[0051] The danger warning module is used to judge the pet's behavior based on the risk factor of the characteristic information and issue a care reminder;

[0052] Furthermore, the Internet of Things terminal module monitors the pet's electrocardiogram waveform through the PetPace sensor.

[0053] Furthermore, the IoT terminal module monitors the pet's electrocardiogram waveform via the Tractive GPS 4 sensor.

[0054] Furthermore, the Internet of Things terminal module is connected to the image monitoring module, the target marking module, the image processing module, the feature analysis module, and the danger warning module.

[0055] Furthermore, the data monitoring module includes a user data acquisition unit, a satellite positioning unit and a monitoring unit;

[0056] The user data acquisition unit is used to acquire data information of the pet owner;

[0057] The satellite positioning unit is used to obtain the activity location information of the pet;

[0058] The monitoring unit is used to monitor the pet's activities in real time through the monitoring device on the pre-worn Whistle GO Explore smart collar;

[0059] The output end of the monitoring unit is connected to the input ends of the user data acquisition unit and the satellite positioning unit.

[0060] Furthermore, the Internet of Things terminal module includes a status detection unit, a data transmission unit and a network control unit;

[0061] The state detection unit is used to obtain the position and speed information of the pet through a sensor;

[0062] The data transmission unit is used to receive and transmit data according to the execution command;

[0063] The network control unit is used to ensure that data is transmitted under normal network connection status;

[0064] The output end of the control unit is connected to the input ends of the pressure detection unit and the data transmission unit.

[0065] A pet stroller intelligent management method integrating positioning and health data collection, characterized by comprising the following steps:

[0066] S1, real-time monitoring of the pet's activity trajectory information and acquisition of multiple frames of real-time images of the pet;

[0067] S2, target detection and tracking are performed on the multiple frames of real-time images to obtain a pet target image sequence;

[0068] S3, preprocessing is performed on the pet target image sequence to obtain a preprocessed image;

[0069] S4, feature information of the pet in the preprocessed image is obtained and a dangerous coefficient of the feature information is calculated;

[0070] S5, behavior judgment is made on the pet according to the dangerous coefficient of the feature information and a care reminder is issued.

[0071] Further, in S1, the specific method for monitoring the activity track information of the pet in real time and obtaining multiple frames of real-time images of the pet is as follows: a camera preset in the pet activity scene is used to shoot real-time activity videos of the pet, the real-time activity videos are decomposed into multiple frames of real-time images, and the multiple frames of real-time images include position information of the pet, the position information including pet position information and pet car position information.

[0072] Further, in S2, the specific method for obtaining a pet target image sequence by performing target detection and tracking on multiple frames of real-time images is as follows: a pre-trained pet target detection model is used, specifically a basic pet detection model, the detection model extracts features from images and classifies them, outputs a bounding box containing a pet, a category, and a confidence score, a preset target tracking algorithm is used to track the pet target in each frame of real-time image, and the pet target image sequence and pet target box position sequence are output as the target image sequence.

[0073] Preferably, the pet detection model is YOLO8S.

[0074] The YoloV8s model is characterized in that it can directly output the position and category information of the target in one forward propagation without the need for multiple scans of the image, which brings high computational performance, making the pet target detection faster, and the detection result will include the bounding box coordinates, category label, and confidence of the pet target in the image. This pet target detection process enables the system to identify and locate the pet in the scene in real time, providing key information for the pet owner, which is a key function for the pet care system, providing a convenient and practical monitoring and tracking tool for users to ensure the health and safety of pets.

[0075] Further, in S3, the specific method of preprocessing the pet target image sequence to obtain the preprocessed image is: the target image sequence is subjected to image filtering processing to obtain the preprocessed image, wherein the image filtering processing includes Gaussian filtering to remove Gaussian noise and median filtering to remove salt and pepper noise; the image sequence subjected to the image filtering processing is subjected to de-duplication processing to remove duplicate images; and the preprocessed image is output.

[0076] Further, in S4, the feature information of the pet in the preprocessed image is acquired and a danger coefficient of the feature information is calculated, the position information in the preprocessed image is acquired as the feature information, the pet position information is denoted as P(i), the pet car position information is denoted as Q(i), the straight-line distance between the pet position information and the pet car position information is calculated as D(i), if D(i) exceeds a preset threshold Th1, the direction and distance that the car needs to travel are calculated according to the pet car position information and the pet position information, so as to determine the motion parameters of the car, such as speed, travel direction and steering angle, so that the pet car can automatically navigate and move along the optimal path to approach the position of the pet.

[0077] The time when the pet car starts to move is denoted as ST, and the time when the pet car stops moving is denoted as ED. The electrocardiogram waveform of the pet from ST to ED is acquired by the electrocardiogram sensor worn in the pet's collar, the signal strength of the electrocardiogram waveform is denoted as E(i), the entire electrocardiogram waveform is traversed, the entire electrocardiogram waveform is subjected to high-pass and low-pass filtering to remove electromyographic interference and baseline drift, and then subjected to wavelet transform to remove small signal interference, so as to ensure that the processed electrocardiogram waveform is smooth and free of burrs, the R wave is marked using the Pan-Tompkins algorithm, the peak of all R waves is denoted as PK(j), and the time interval between PK(j) and PK(j+1) is denoted as T(j), j is the serial number of the time interval, all time intervals T(j) are traversed, and the danger coefficient is calculated according to the formula M is the number of time intervals, if RK(j+1)-RK(j)>RK(j)-RK(j-1) and T(j)>T(j+1), or RK(j+1)-RK(j)<RK(j)-RK(j-1) and T(j)<T(j+1), the time corresponding to PK(j) is recorded as the stress time, and the pet car is controlled to stop moving;

[0078] Further, in S5, the specific method of making a behavior judgment on the pet according to the danger coefficient of the feature information and issuing a care reminder is: according to the calculated danger coefficient RK(j) of the feature information, if RK(j) is less than zero, it is considered that the pet is in a stress state, the pet car is controlled to stop tracking, to prevent the pet from being stressed, and a care alarm is issued to remind the owner to take good care of the pet.

[0079] Embodiment 2

[0080] The embodiment 2 is to replace the method of obtaining the feature information of the pet in the preprocessed image and calculating the risk coefficient of the feature information based on the embodiment 1, and specifically is:

[0081] All time intervals T(j) are traversed, and the difference between T(j+1) and T(j) is calculated as TO(j). All the differences TO(j) are traversed in turn. If TO(j+1)>TO(j) and TO(j)<TO(j-1), the T(j) is marked as the fluctuation time starting point. The remaining R wave peaks are searched along the electrocardiogram waveform from the R wave peak value corresponding to T(j) to the rear, the difference between the remaining R wave peaks and the R wave peak corresponding to the time starting point is calculated as DIF(k), k represents the serial number of the R wave peak searched along the electrocardiogram waveform from the R wave peak value corresponding to T(j) to the rear, and the risk coefficient RK(j) is calculated by the formula The risk coefficient RK(j) is calculated, where N is the number of R wave peaks searched along the electrocardiogram waveform from the R wave peak value corresponding to T(j) to the rear, DIF_max represents the maximum difference between the R wave peak searched to the rear and the R wave peak corresponding to T(j), and DIF_min represents the minimum difference between the R wave peak searched to the rear and the R wave peak corresponding to T(j). If RK(j) is less than zero, it is considered that the pet is in an excited state, and the pet car is controlled to continue tracking to prevent the pet from being lost. Otherwise, it is considered that the pet is in a stress state, and the pet car is controlled to stop tracking to prevent the pet from being stressed.

[0082] The pet car intelligent management system integrating positioning and health data acquisition can run in desktop small computers, notebooks, palmtop computers and cloud servers and other computing devices. The pet car intelligent management system integrating positioning and health data acquisition can run a system that can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the example is only an example of the pet car intelligent management system integrating positioning and health data acquisition, and does not constitute a limitation on the pet car intelligent management system integrating positioning and health data acquisition, and can include more or fewer components, or combine certain components, or different components, for example, the pet car intelligent management system integrating positioning and health data acquisition can also include input and output devices, network access devices, buses, etc.

[0083] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic components, discrete hardware components, etc. The general-purpose processor can be a microprocessor or can also be any conventional processor. The processor is a control center of the pet car intelligent management system running system of the fusion positioning and health data acquisition, and connects each part of the whole pet car intelligent management system running system of the fusion positioning and health data acquisition through various interfaces and lines.

[0084] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the pet car intelligent management system of the fusion positioning and health data acquisition by running or executing the computer program and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc. The data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0085] Although the description of the present application has been quite detailed and particularly described with respect to several embodiments, it is not intended to be limited to any of these details or embodiments or any special embodiment, so as to effectively cover the intended scope of the present application. In addition, the present application is described above in the embodiments that the inventors can foresee, and the purpose is to provide a useful description, and those non-essential modifications to the present application that have not yet been foreseen can still represent equivalent modifications of the present application.

Claims

1. The intelligent management system for pet cars that integrates positioning and health data collection is characterized by: The system includes the following modules: Image monitoring module, used to monitor the pet's activity trajectory information in real time and obtain multiple frames of real-time images of the pet; The target marking module is used to detect and track targets in multiple frames of real-time images to obtain a pet target image sequence; An image processing module, configured to preprocess the pet target image sequence to obtain a preprocessed image; A feature analysis module is used to obtain feature information of the pet in the pre-processed image and calculate the risk coefficient of the feature information; The danger warning module is used to judge the pet's behavior based on the risk factor of the characteristic information and issue a care reminder; The IoT terminal module monitors data by controlling sensors; The Internet of Things terminal module is connected to the image monitoring module, the target marking module, the image processing module, the feature analysis module, and the danger warning module.

2. The pet stroller intelligent management system integrating positioning and health data collection according to claim 1 is characterized in that: The data monitoring module includes a user data acquisition unit, a satellite positioning unit and a monitoring unit; The user data acquisition unit is used to acquire data information of the pet owner; The satellite positioning unit is used to obtain the activity location information of the pet; The monitoring unit is used to monitor the pet's activities in real time through a pre-installed monitoring device; The output end of the monitoring unit is connected to the input ends of the user data acquisition unit and the satellite positioning unit.

3. The pet stroller intelligent management system integrating positioning and health data collection according to claim 1 is characterized in that: The Internet of Things terminal module includes a status detection unit, a data transmission unit and a network control unit; The state detection unit is used to obtain the position and speed information of the pet through a sensor; The data transmission unit is used to receive and transmit data according to the execution command; The network control unit is used to ensure that data is transmitted under normal network connection status; The output end of the control unit is connected to the input ends of the pressure detection unit and the data transmission unit.

4. A pet stroller intelligent management method integrating positioning and health data collection, applied to the pet stroller intelligent management system integrating positioning and health data collection as claimed in any one of claims 1 to 3, characterized in that: The following steps are involved: S1, real-time monitoring of the pet's activity trajectory information and acquisition of multiple frames of real-time images of the pet; S2, performing target detection and tracking on multiple frames of real-time images to obtain a pet target image sequence; S3, preprocessing the pet target image sequence to obtain a preprocessed image; S4, obtaining feature information of the pet in the preprocessed image and calculating a risk coefficient of the feature information; S5, makes behavioral judgments on the pet based on the risk factor of the characteristic information and issues a care reminder.

5. The pet stroller intelligent management method integrating positioning and health data collection according to claim 4 is characterized in that: In S1, the method for real-time monitoring of the pet's activity trajectory information and obtaining multiple frames of real-time images of the pet is: using a camera preset in the pet's activity scene to shoot the pet's real-time activity video, and decomposing the real-time activity video into multiple frames of real-time images, wherein the multiple frames of real-time images include the pet's location information, and the location information includes the pet's location information and the pet car's location information.

6. The pet stroller intelligent management method integrating positioning and health data collection according to claim 5 is characterized in that: In S2, the method for performing object detection and tracking on multi-frame real-time images to obtain a pet object image sequence is as follows: Use a pre-trained pet object detection model, specifically a base pet detection model. The detection model extracts features and classifies the image, and outputs a bounding box, category, and confidence score containing the pet. Using a preset object tracking algorithm, track the pet objects in each frame of the real-time image, and output the pet object image sequence and the pet object box position sequence as the target image sequence.

7. The pet stroller intelligent management method integrating positioning and health data collection according to claim 6 is characterized in that: In S3, the method for preprocessing the pet object image sequence to obtain the preprocessed image is as follows: The target image sequence is processed by image filtering to obtain the preprocessed image. Image filtering includes Gaussian filtering to remove Gaussian noise and median filtering to remove salt-and-pepper noise. The image sequence after image filtering is de-duplicated to remove duplicate images, and the preprocessed image is output.

8. The pet stroller intelligent management method integrating positioning and health data collection according to claim 7 is characterized in that: In S4, the method for obtaining the feature information of the pet in the preprocessed image and calculating the risk coefficient of the feature information is as follows: Obtain the position information in the processed image as the feature information. Denote the pet position information as P(i) and the pet car position information as Q(i). Calculate the straight-line distance between the pet position information and the pet car position information as D(i). If D(i) exceeds the preset threshold Th1, based on the pet car position information and the pet position information, calculate the direction and distance that the car needs to travel to determine the motion parameters of the car, such as speed, traveling direction, and steering angle. In this way, the pet car can automatically navigate and move along the optimal path to approach the position of the pet. Denote the start time of the pet car's movement as ST and the stop time of the pet car's movement as ED. Obtain the electrocardiogram waveform of the pet from ST to ED through an electrocardiogram sensor worn on the pet collar. Denote the signal intensity of the electrocardiogram waveform as E(i). Traverse the entire electrocardiogram waveform. Remove electromyogram interference and baseline drift from the entire electrocardiogram waveform through high-pass and low-pass filters, and then remove small signal interference through wavelet transform to ensure that the processed electrocardiogram waveform is smooth and free of burrs. Use the Pan-Tompkins algorithm to mark the R waves. Denote the peaks of all R waves as PK(j). Denote the time interval between PK(j) and PK(j + 1) as T(j). Traverse all time intervals T(j). Calculate the risk coefficient according to the formula. If RK(j + 1) - RK(j) > RK(j) - RK(j - 1) and T(j) > T(j + 1), or RK(j + 1) - RK(j) < RK(j) - RK(j - 1) and T(j) < T(j + 1), record the time corresponding to PK(j) as the stress moment and control the pet car to stop moving.

9. The pet stroller intelligent management method integrating positioning and health data collection according to claim 8, characterized in that: Replace the method of obtaining the feature information of the pet in the preprocessed image and calculating the risk coefficient of the feature information with the following: Traverse all time intervals T(j), calculate the difference between T(j + 1) and T(j) as TO(j) in sequence, traverse all the differences TO(j) in sequence. If TO(j + 1) > TO(j) and TO(j) < TO(j - 1), then mark T(j) as the starting point of the fluctuation time. Retrieve the remaining R-wave peaks backward along the electrocardiogram waveform from the R-wave peak value corresponding to T(j), and calculate the difference between the remaining R-wave peaks and the R-wave peak corresponding to the time starting point as DIF(k), where k represents the serial number of the R-wave peak retrieved backward along the electrocardiogram waveform from the R-wave peak value corresponding to T(j). Calculate the risk coefficient RK(j) through the formula. If RK(j) is less than zero, it is considered that the pet is in an excited state at present, and control the pet car to continue tracking to prevent the pet from getting lost. Otherwise, it is considered that the pet is in a stress state, and control the pet car to stop tracking to prevent the pet from being stressed.

10. The pet stroller intelligent management method integrating positioning and health data collection according to claim 9, characterized in that: In S5, the method of making a behavioral judgment on the pet according to the risk coefficient of the feature information and issuing a care reminder is as follows: Calculate the risk coefficient RK(j) according to the feature information. If RK(j) is less than zero, it is considered that the pet is in a stress state, control the pet car to stop tracking to prevent the pet from being stressed, and issue a care alarm to remind the owner to take good care of the pet.

Images