Method and device for non-contact detection of pet vital signs

Abstract

The invention discloses a method and device for non-contact detection of pet vital signs, and belongs to the field of pet health monitoring, the device comprises a pad body and a camera unit, the pad body is provided with a waterproof outer layer, a piezoelectric sensor layer and a damping medium layer from top to bottom in sequence, the piezoelectric sensor layer is used for collecting the heart rate and breathing overall curve of a pet, the damping medium layer is used for isolating external noise signals, and the waterproof outer layer is used for protecting the piezoelectric sensor layer; according to the method, the overall curve of the heart rate and respiration of the pet is collected through the piezoelectric sensor, respiration signals and heartbeat signals are separated through filtering, and pet signs are monitored in an auxiliary mode in combination with visual data collected by the camera. Non-contact detection of pet vital signs is achieved, the damping medium layer effectively isolates external noise, the signal collection precision is improved, pet death and disease risks can be found in time in combination with auxiliary monitoring of visual data, and the pet vital sign detection device has the advantages of being accurate in detection and convenient and fast to use.

Description

Technical Field

[0001] This invention relates to the field of pet health monitoring technology, and in particular to a method and apparatus for non-contact detection of pet vital signs. Background Technology

[0002] With the increasing popularity of pet ownership, pet health is receiving more and more attention, and vital signs (heart rate, respiratory rate, etc.) are important indicators reflecting pet health. Traditional pet vital sign detection mostly uses contact methods, such as wearable devices or manual testing. These methods suffer from problems such as low pet cooperation, cumbersome testing processes, and inaccurate data due to the pet's emotional state. At the same time, most existing detection devices can only collect physiological signals individually and cannot combine visual information to assist in the judgment of pet mortality and disease risk, resulting in insufficient comprehensiveness and timeliness of monitoring.

[0003] In addition, noise and vibration from the external environment can also interfere with the accuracy of physiological signal acquisition, further affecting the reliability of the test results.

[0004] Therefore, a non-contact method and device for detecting pet vital signs is needed to solve the problems of low cooperation rate, susceptibility to noise interference in signal acquisition, and limited monitoring dimensions in existing pet vital sign detection methods. Summary of the Invention

[0005] This invention aims to solve the problems of low cooperation rate in existing pet vital sign detection methods, susceptibility to noise interference in signal acquisition, and limited monitoring dimensions. It provides a device and method for non-contact monitoring of pet vital signs, achieving accurate and non-contact detection of pet vital signs, and improving the comprehensiveness of monitoring by combining visual data.

[0006] Technical Solution: To solve the above-mentioned technical problems, according to one aspect of the present invention, more specifically, a non-contact monitoring device for pet vital signs, comprising a mat body and a camera unit, as detailed below: The pad body is provided with a waterproof outer layer, a piezoelectric sensor layer, and a shock-absorbing medium layer from top to bottom; The piezoelectric sensor layer is used to collect the pet's overall heart rate and respiratory curves, the shock-absorbing medium layer is used to isolate external noise signals, and the waterproof outer layer is used to protect the piezoelectric sensor layer. The camera unit includes an infrared camera and a visual camera; The infrared camera is used to monitor the pet's body temperature, and the visual camera is used to collect the pet's visual data to assist in monitoring the pet's death and disease information.

[0007] Furthermore, the damping medium layer is made of gel material to isolate external noise signals, with a damping coefficient between 0.1 and 0.3 and a density between 0.9 and 1.1 g / cm³. 3 The penetration (1 / 10mm) is 300~500.

[0008] Furthermore, the visual camera is equipped with a pre-trained pet target monitoring model and a disease recognition model trained based on the random forest algorithm.

[0009] Furthermore, the piezoelectric sensor layer is a flexible piezoelectric sensor array arranged in a serpentine pattern, which is suitable for pressure collection from pets of different sizes.

[0010] According to one aspect of the present invention, and more specifically a method for non-contact detection of pet vital signs, using a non-contact detection device for pet vital signs as described above, characterized by comprising the following steps: S1. Signal Acquisition: Collect the overall heart rate and respiratory curves of the pet on the mat through the piezoelectric sensor layer; S2. Signal Separation: Bandpass filters are used to separate respiratory and heartbeat signals. An equivalent cutoff frequency of 8-75 beats / minute is set for the respiratory signal, and an equivalent cutoff frequency of 83-240 beats / minute is set for the heartbeat signal. The respiratory rate and heartbeat rate are calculated by signal window analysis and fast Fourier transform, respectively. S3. Vital Signs Monitoring: Combines the separated physiological signals to monitor the pet's basic vital signs. At the same time, it uses visual data from the vision camera to determine the pet's risk of death through a pre-trained model, and uses a disease identification model trained by the random forest algorithm to determine the risk of diseases such as epilepsy and vomiting. Combines the body temperature data from the infrared camera to generate a comprehensive vital signs report.

[0011] The beneficial effects of the non-contact detection method and device for pet vital signs of the present invention are as follows: (1) The present invention adopts a non-contact detection method, which collects signals through the piezoelectric sensor on the mat. There is no need for pets to wear equipment, which improves the pets' cooperation and makes the detection process more convenient. The damping medium layer effectively isolates external noise interference, and combined with a bandpass filter with a specific cutoff frequency to separate signals, it improves the accuracy of heart rate and respiratory signal acquisition. By combining infrared camera for body temperature monitoring and visual camera for behavior monitoring, a multi-dimensional fusion of physiological signals with visual and body temperature data is achieved, enabling timely detection of pet mortality and disease risks, and improving the comprehensiveness and timeliness of monitoring. Attached Figure Description

[0012] The present invention will now be described in further detail with reference to the accompanying drawings and specific implementation methods.

[0013] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a schematic diagram of the structure of the mat body in this invention.

[0014] In the diagram: 1. Mat body; 11. Shock-absorbing medium layer; 12. Piezoelectric sensor layer; 13. Waterproof outer layer; 2. Camera unit. Detailed Implementation

[0015] The present invention will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the present application can be combined with each other.

[0016] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] Reference Figures 1-2 The present invention discloses a method and apparatus for non-contact detection of vital signs in pets, the implementation of which is as follows: 1. Device Preparation Create a device for non-contact monitoring of pet vital signs: The waterproof outer layer 13 of the mat body 1 is made of TPU flexible waterproof material with a thickness of 0.5mm; The piezoelectric sensor layer 12 uses a PVDF piezoelectric thin film sensor array, which is laid out in a serpentine pattern with a sensor spacing of 2cm, uniformly covering the area below the waterproof outer layer 13, and is suitable for pressure collection of pets of different sizes. The damping medium layer 11 is made of gel material, with its damping coefficient adjusted to 0.2 and its density controlled between 0.9 and 1.1 g / cm³. 3 The penetration (1 / 10mm) is 300~500, the thickness is 1cm, and it is laid under the piezoelectric sensor layer 12; A camera unit 2 is installed on one side of the mat body 1. The camera unit 2 includes an infrared camera (1080P resolution, temperature measurement range 0-50℃) and a visual camera (2K resolution, 30fps frame rate). The camera's shooting angle covers the entire mat area.

[0018] 2. Implementation of testing methods S1. Signal Acquisition: The pet is placed on the prepared mat. The piezoelectric sensor layer 12 senses the slight pressure changes of the pet's body, converts them into electrical signals, and then acquires the overall curves of heart rate and respiration. The sampling frequency is set to 100Hz.

[0019] S2. Respiratory Signal Separation and Analysis: A bandpass filter is used to filter the overall curve, with an equivalent cutoff frequency of 8-75 breaths / minute to separate the respiratory signal. The respiratory signal is segmented with a 10-second signal window and a 5-second rolling step. The peaks and troughs in each signal window are identified using a peak-trough-peak detection algorithm. The time interval between peaks and troughs is calculated, and the average value is taken and the reciprocal is used to obtain the respiratory rate. If the signal amplitude in a certain signal window is less than the standard deviation of the signal within 1 minute, the signal window is considered invalid and the corresponding data is discarded.

[0020] S3. Heartbeat Signal Separation and Analysis: A bandpass filter is used to filter the overall curve, and the equivalent cutoff frequency is set to 83~240 beats / minute to separate the heartbeat signal. Similarly, the signal is divided into segments with a 10-second signal window and a 5-second rolling step size. A fast Fourier transform is performed on the signal in each signal window, and the frequency corresponding to the peak value of the transformed spectrum is taken as the heartbeat frequency in that signal window.

[0021] S4, Auxiliary Monitoring: Death information monitoring: The visual camera captures images of the pet and tracks the pet's position through a pre-trained pet target detection model. If the pet's movement pixels do not exceed 20 within 1 minute and breathing signals cannot be detected, a pet death notification message is triggered. Disease information monitoring: Input the pet behavior video dataset of epilepsy and vomiting into the random forest algorithm for model training. After training, input the pet behavior data collected by the visual camera into the model. When the confidence of the model recognition is higher than 90%, attach the disease status information of epilepsy or vomiting to the vital signs report. Infrared cameras collect pets' body temperature data, and the data on body temperature, respiratory rate, and heart rate are integrated to generate a comprehensive report on the pet's vital signs.

[0022] In summary, the device of the present invention has a simple structure, is easy to mass-produce, and can be applied to scenarios such as pet hospitals, pet boarding centers, and family pet breeding; the detection method is convenient to operate, requires no professional personnel to operate, and can realize real-time and automatic monitoring of pet vital signs, which has significant industrial application value and market prospects.

[0023] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A non-contact device for monitoring the vital signs of pets, comprising a mat body (1) and a camera unit (2), characterized in that: The pad body (1) is provided with a waterproof outer layer (13), a piezoelectric sensor layer (12), and a shock-absorbing medium layer (11) from top to bottom. The piezoelectric sensor layer (12) is used to collect the pet's heart rate and overall respiratory curve, the shock-absorbing medium layer (11) is used to block external noise signals, and the waterproof outer layer (13) is used to protect the piezoelectric sensor layer (12). The camera unit (2) includes an infrared camera and a visual camera; The infrared camera is used to monitor the pet's body temperature, and the visual camera is used to collect the pet's visual data to assist in monitoring the pet's death and disease information.

2. The device for non-contact monitoring of pet vital signs according to claim 1, characterized in that, The damping medium layer (11) is made of gel material with a damping coefficient between 0.1 and 0.3 and a density between 0.9 and 1.1 g / cm³. 3 The penetration (1 / 10mm) is 300~500.

3. The device for non-contact monitoring of pet vital signs according to claim 1, characterized in that, The visual camera is equipped with a pre-trained pet target monitoring model and a disease recognition model trained based on the random forest algorithm.

4. The device for non-contact monitoring of pet vital signs according to claim 1, characterized in that, The piezoelectric sensor layer (12) is a flexible piezoelectric sensor array arranged in a serpentine pattern, which is suitable for pressure collection of pets of different sizes.

5. A method for non-contact detection of pet vital signs, using the device for non-contact detection of pet vital signs as described in claim 4, characterized in that, Includes the following steps: S1, Signal Acquisition: The overall curve of the pet's heart rate and respiration on the mat body (1) is detected by the piezoelectric sensor layer (12) in the device; S2. Signal separation: The overall curve is separated by filtering to separate the respiratory signal containing respiratory features from the heartbeat signal containing heart rate features. S3. Vital Signs Monitoring: Based on the separated signals, the device monitors the pet's vital signs and combines visual data collected by the visual camera in the device to assist in monitoring the pet's vital signs.

6. The method for non-contact detection of pet vital signs according to claim 5, characterized in that, In step S2, when separating the respiratory signal, a bandpass filter is used, and the equivalent cutoff frequency of the bandpass filter is between 8 and 75 times per minute. The separated respiratory signal is analyzed with a signal window of 10 seconds and a rolling step of 5 seconds to find the peaks and troughs within the signal window. The time of peak-trough-peak is regarded as one breath, and the reciprocal of the average duration of the breath within the signal window is taken as the respiratory frequency. If the signal amplitude within the signal window is less than the standard deviation within 1 minute, the signal window is determined to be an invalid signal window.

7. The method for non-contact detection of pet vital signs according to claim 5, characterized in that, In step S2, when separating the heartbeat signal, a bandpass filter is used, and the equivalent cutoff frequency of the bandpass filter is between 83 and 240 beats per minute. The separated heartbeat signal is analyzed with a signal window of 10 seconds and a rolling step of 5 seconds. The signal within the signal window is subjected to a fast Fourier transform, and the peak value of the spectrum is taken as the corresponding heartbeat frequency within the signal window.

8. The method for non-contact detection of pet vital signs according to claim 5, characterized in that, The specific process of combining visual camera data to assist in monitoring death information in step S3 is as follows: use a pre-trained dataset to monitor pet targets. If the pet target moves no more than 20 pixels within 1 minute and the breathing signal cannot be detected, then issue a prompt message that the pet has died.

9. The method for non-contact detection of pet vital signs according to claim 5, characterized in that, The specific process of combining visual camera data to assist in monitoring disease information in step S3 is as follows: using epilepsy and vomiting datasets, machine learning is trained using the random forest method. When the confidence level of the model recognition is higher than 90%, the corresponding disease status information is attached while reporting vital signs information.

10. The method for non-contact detection of pet vital signs according to claim 5, characterized in that, In step S3, the pet's body temperature data is collected by the infrared camera in the device, and the body temperature data is integrated with heart rate and respiratory rate data to generate a pet vital signs report.

Images