A non-contact animal respiratory rate detection method, system, device and medium based on rPPG

Abstract

A non-contact animal respiratory rate detection method, system, equipment and medium based on rPPG, the method comprising the following steps: fixing a camera on or around an animal to extract images of the animal's respiratory movements, uploading the images to an intelligent terminal, selecting a suitable ROI, extracting the BVP signal of the ROI, and processing the extracted signal using an rPPG algorithm to obtain the final respiratory rate; the intelligent terminal displays the respiratory rate, confidence level and final weighted respiratory rate calculated from the extracted signals of different ROIs; the system, equipment and medium perform animal respiratory rate detection based on the non-contact animal respiratory rate detection method based on rPPG; the present invention has the advantages of being convenient and fast, reducing animal emergency situations, being highly real-time and having a high accuracy rate; and being able to display the animal's respiratory rate in real time under non-contact conditions, thereby reducing the consumption of manpower and material resources in measuring the animal's respiratory rate and enhancing real-time performance and accuracy.

Description

Technical Field

[0001] The present invention relates to the field of animal health detection technology, and in particular to a non-contact animal respiratory rate detection method, system, equipment and medium based on rPPG. Background Art

[0002] As a key indicator of an animal's cardiac function and health, respiratory rate has significant application value in a variety of fields, including veterinary medicine, animal husbandry, animal behavior, and ecological conservation. The timeliness, accuracy, and reliability of respiratory rate monitoring directly impact disease prevention, diagnosis, treatment, and postoperative recovery. Traditional animal respiratory rate monitoring relies primarily on manual labor, requiring owners to manually measure, record, and plot respiratory rate curves daily, a significant workload. Furthermore, free-range animals, such as cattle and sheep, are unable to cooperate with respiratory rate monitoring, often requiring anesthesia.

[0003] To address this issue, contact sensors, such as electrocardiogram (ECG) monitors and respiration rate belts, have emerged. However, contact respiration rate monitoring often requires the device to be directly attached to the animal's body. These methods, which require direct or indirect contact with the animal's body surface, are not only cumbersome to operate but can also cause stress or discomfort to the animal. For example, when monitoring flocks of sheep or cattle, contact monitoring can be difficult to implement due to the animal's lack of cooperation.

[0004] Non-contact animal respiratory rate monitoring technologies primarily rely on external energy detection media. These technologies obtain respiratory rate information by detecting micro-movements or changes caused by physiological activities on or within the animal's body. Non-contact monitoring can be performed from a distance, eliminating the need for direct contact with the animal, reducing risks during the testing process. Furthermore, the technology is simple and quick to operate, making it suitable for animal respiratory rate monitoring in a variety of environments and conditions.

[0005] Patent application CN202210023274.7 uses a video processing method to perform frequency domain decomposition on the processed signal, and screens out the sub-components with the highest matching degree in the frequency domain based on the heart rate range unique to aquatic animals, and then outputs a cardiac waveform or calculates the heart rate value; it can only detect defects in aquatic animals.

[0006] Patent application CN202410051962.3 utilizes a dual-channel PPG method, employing anomaly detection and reconstruction techniques, and weighting heart rate values ​​based on anomaly scores to calculate the final heart rate. However, this method suffers from the drawback of high signal acquisition difficulty. In practical applications, obtaining high-quality dual-channel PPG data requires specialized acquisition equipment and a demanding acquisition environment. Summary of the Invention

[0007] In order to overcome the defects of the above-mentioned prior art, the purpose of the present invention is to provide a non-contact animal respiratory rate detection method, system, equipment and medium based on rPPG, which captures the movement of the animal's skin caused by breathing through a camera, and then selects the area with the most obvious undulation of the animal's abdomen as the ROI (region of interest), and extracts the signal of the GREEN channel in the RGB image as the BVP (blood volume pulse) signal through the rPPG (video-based remote heart rate detection) algorithm. The BVP signal is processed to obtain the animal's respiratory rate information and displayed on the interface, thereby realizing real-time remote monitoring of the animal's respiratory rate. There is no need for manual detection and recording of results, which saves manpower and makes the detection results more accurate, so as to solve the problems of traditional respiratory rate detection methods that are labor-intensive and may cause stress to animals.

[0008] In order to achieve the above object, the technical solution adopted by the present invention is:

[0009] A non-contact animal respiratory rate detection method based on rPPG, in which a camera is fixed on or around the animal to extract images of the animal's respiratory movements and upload them to an intelligent terminal. A suitable ROI is selected, and then the BVP signal of the ROI is extracted. The extracted signal is processed using the rPPG algorithm to obtain the final respiratory rate; the intelligent terminal displays the respiratory rate, confidence level and final weighted respiratory rate calculated from the extracted signals of different ROIs.

[0010] A non-contact animal respiratory rate detection method based on rPPG specifically comprises the following steps:

[0011] Step 1: Fix the camera on or around the animal so that it can accurately capture images of the animal's breathing movements;

[0012] Step 2: The image captured by the camera in step 1 is transmitted to the smart terminal, the region is located, and the ROI image data is selected;

[0013] Step 3: The respiratory rate, confidence level, and weighted sum of the final respiratory rate for each ROI are calculated using the rPPG algorithm;

[0014] Step 4: Display the image captured by the camera in step 1 and the respiratory rate, confidence level, and weighted sum of the final respiratory rate of each ROI calculated in step 3 on the interface.

[0015] The specific steps of the rPPG algorithm described in step 3 are as follows:

[0016] 3.1. Read the image data of the ROI selected in step 2; obtain the RGB channel histogram of each ROI; subtract the signals of the green (GREEN) channel and the red (RED) channel in the RGB channel histogram, and obtain the absolute value of the difference between the two as the BVP signal; filter out the deviation data in the BVP signal that is greater than the preset deviation, and then filter it using a bandpass filter to obtain a filtered signal; perform a fast Fourier transform (FFT) on the obtained filtered signal to obtain the filtered signal spectrum, and then perform peak detection to obtain the instantaneous respiratory rate of each ROI;

[0017] 3.2. Use the formula BPM(n) = α·BPM(n-1) + 1-αBPMSig(n) to weight the instantaneous respiratory rate obtained in step 3.1. Set α = 0.95 to ensure that the instantaneous respiratory rate results of each ROI do not jump;

[0018] 3.3, based on the peak value of the bvp signal of each ROI extracted in step 3.1, the filtered signal is compared with the peak value of the spectrum, and the mean square error (MSE) of the spectrum is calculated using the function to express the degree of spectrum leakage, which is recorded as leakage, and the confidence is used. i =1 / 1+leakage to calculate the BPM confidence of each ROI; based on the confidence and the instantaneous respiratory rate calculated in step 3.2 without jumps, use the following formula to calculate the final respiratory rate result by weighted summation:

[0019]

[0020] A non-contact animal respiratory rate detection system based on rPPG, comprising:

[0021] The image acquisition module is used in step 1 to capture images with a camera to collect images;

[0022] The ROI confirmation module is used in step 2 to locate the ROI area of ​​the collected images;

[0023] The signal processing module is used in step 3 to extract and process the signal of the ROI determined in step 2 to calculate the respiratory rate.

[0024] A non-contact animal respiratory rate detection device based on rPPG, comprising:

[0025] memory for storing computer programs;

[0026] A processor is used to implement the rPPG-based non-contact animal respiratory rate detection method described in steps 1 to 4 when executing the computer program.

[0027] A computer-readable storage medium stores a computer program. When executed by a processor, the computer program can perform rPPG-based non-contact animal respiratory rate detection based on the rPPG-based non-contact animal respiratory rate detection method described in steps 1 to 4.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. The present invention fixes a camera on or around the animal, extracts images of the animal's respiratory movement, uploads them to an intelligent terminal, extracts the BVP signal of the ROI, and processes it using the rPPG algorithm. This can achieve automatic detection and remote real-time detection of the animal's respiratory rate, eliminating the need for manual recording of the respiratory rate status of the detected animal, saving a lot of manpower and reducing labor.

[0030] 2. This method extracts the BVP signal from each ROI, filters it, and Fourier transforms it, then performs peak detection to obtain the instantaneous respiratory rate for each ROI. This weighted update of the instantaneous respiratory rate increases detection accuracy. Based on the confidence level and the obtained instantaneous respiratory rate that does not jump, a weighted summation is performed to calculate the final respiratory rate result, resulting in efficient processing and high accuracy.

[0031] 3. The present invention is connected to an intelligent monitoring system on a remote network and uploads respiratory rate status information to the remote system. Relevant personnel can understand the historical respiratory rate status of the monitored animals at any time through the network, and can use the expert system to analyze the different respiratory rate changes of the monitored objects to achieve more accurate health monitoring.

[0032] In summary, the respiratory rate detection system of the present invention does not require direct contact with animals, reduces animal stress, and increases detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a scatter plot comparing the respiratory rate tested by the present invention and the respiratory rate observed manually.

[0034] Figure 2 This is a flow chart of the animal respiratory rate detection method of the present invention.

[0035] Figure 3 Flowchart of the rPPG algorithm for calculating animal respiratory rate according to the present invention. DETAILED DESCRIPTION

[0036] See also Figure 2 、 Figure 3 A non-contact animal respiratory rate detection method based on rPPG specifically comprises the following steps:

[0037] Step 1: Fix the camera on or around the animal so that it can accurately capture images of the animal's breathing movements;

[0038] Step 2: The image captured by the camera in step 1 is transmitted to the smart terminal, the region is located, and the ROI image data is selected;

[0039] Step 3: The respiratory rate, confidence level, and weighted sum of the final respiratory rate for each ROI are calculated using the rPPG algorithm;

[0040] 3.1. Read the image data of the ROI selected in step 2; obtain the RGB channel histogram of each ROI; subtract the signals of the green (GREEN) channel and the red (RED) channel in the RGB channel histogram, and obtain the absolute value of the difference between the two as the BVP signal; filter out the deviation data in the BVP signal that is greater than the preset deviation, and then filter it using a bandpass filter to obtain a filtered signal; perform a fast Fourier transform (FFT) on the obtained filtered signal to obtain the filtered signal spectrum, and then perform peak detection to obtain the instantaneous respiratory rate of each ROI;

[0041] 3.2. Use the formula BPM(n) = α·BPM(n-1) + 1-αBPMSig(n) to weight the instantaneous respiratory rate obtained in step 3.1. Set α = 0.95 to ensure that the instantaneous respiratory rate results of each ROI do not jump;

[0042] 3.3, based on the peak value of the bvp signal of each ROI extracted in step 3.1, the filtered signal is compared with the peak value of the spectrum, and the mean square error (MSE) of the spectrum is calculated using the function to express the degree of spectrum leakage, which is recorded as leakage, and the confidence is used. i =1 / 1+leakage to calculate the BPM confidence of each ROI; based on the confidence and the instantaneous respiratory rate calculated in step 3.2 without jumps, use the following formula to calculate the final respiratory rate result by weighted summation:

[0043]

[0044] Step 4: Display the image captured by the camera in step 1 and the respiratory rate, confidence level, and weighted sum of the final respiratory rate of each ROI calculated in step 3 on the interface.

[0045] A non-contact animal respiratory rate detection system based on rPPG, comprising:

[0046] The image acquisition module is used in step 1 to capture images with a camera to collect images;

[0047] The ROI confirmation module is used in step 2 to locate the ROI area of ​​the collected images;

[0048] The signal processing module is used in step 3 to extract and process the signal of the ROI determined in step 2 to calculate the respiratory rate.

[0049] A non-contact animal respiratory rate detection device based on rPPG, comprising:

[0050] memory for storing computer programs;

[0051] A processor is used to implement the rPPG-based non-contact animal respiratory rate detection method described in steps 1 to 4 when executing the computer program.

[0052] A computer-readable storage medium stores a computer program. When executed by a processor, the computer program can perform rPPG-based non-contact animal respiratory rate detection based on the rPPG-based non-contact animal respiratory rate detection method described in steps 1 to 4.

[0053] Simulation experiment

[0054] Simulation conditions and equipment

[0055] 1. Environmental conditions

[0056] The simulation environment is a general breeding environment with a temperature of 18-25°C and a humidity of 40%-60% RH. Healthy, quiet, undisturbed sheep that are not eating or ruminating are selected. Before the simulation experiment, the sheep are placed in the experimental environment for more than 30 minutes to adapt to their physiological state and ensure that the collected respiratory rate data is representative and stable.

[0057] 2. Simulation equipment

[0058] 1. High-definition camera, resolution 1080p, frame rate 50fps, bracket mountable.

[0059] 2. High-performance computer with professional data processing software, Python, etc.

[0060] Simulation content

[0061] The camera is aimed at the sheep's abdomen to collect data. Next, the ROI is determined and the respiratory rate is calculated using a non-contact animal respiration rate detection algorithm based on rPPG. The calculated respiration rate is then compared with the actual respiration rate, which is the average of three minutes of manual observation.

[0062] The simulation results are as follows:

[0063]

[0064]

[0065] Draw a scatter plot, such as Figure 1 As shown, the horizontal axis represents the heart rate detected by the present invention, and the vertical axis represents the manually observed respiratory rate (averaged after three minutes of observation). The calculated R-squared value (used to measure how well the predicted value fits the true value, with a value of 1 considered a perfect fit and a value of 0.7 or greater generally considered to indicate a good fit) is 0.841, demonstrating that the present invention not only enables contactless animal heart rate detection, but also demonstrates a very high accuracy in respiratory rate prediction, resulting in a good fit.

Claims

1. A non-contact animal respiratory rate detection method based on rPPG, characterized in that: The camera is fixed on or around the animal to capture images of the animal's respiratory movements and upload them to the smart terminal. The appropriate ROI is selected and the BVP signal of the ROI is extracted. The extracted signal is processed using the rPPG algorithm to obtain the final respiratory rate. The intelligent terminal displays the respiratory rate, confidence level and final weighted respiratory rate calculated from the extracted signals of different ROIs; Step 1: Fix the camera on or around the animal so that it can accurately capture images of the animal's breathing movements; Step 2: The image captured by the camera in step 1 is transmitted to the smart terminal, the region is located, and the ROI image data is selected; Step 3: The respiratory rate, confidence level, and weighted sum of the final respiratory rate for each ROI are calculated using the rPPG algorithm; 3.

1. Read the image data of the ROI selected in step 2; obtain the RGB channel histogram of each ROI; subtract the signals of the green (GREEN) channel and the red (RED) channel in the RGB channel histogram, and obtain the absolute value of the difference between the two as the BVP signal; filter out the deviation data in the BVP signal that is greater than the preset deviation, and then filter it using a bandpass filter to obtain a filtered signal; perform a fast Fourier transform (FFT) on the obtained filtered signal to obtain the filtered signal spectrum, and then perform peak detection to obtain the instantaneous respiratory rate of each ROI; 3.

2. Use the formula BMP(n) = α·BMP(n-1) + (1-α)BMP(Sig(n)) to weight the instantaneous respiratory rate obtained in step 3.

1. Set α = 0.95 to ensure that the instantaneous respiratory rate results of each ROI do not jump. BPM is the respiratory rate; n is the number of BMPs. 3.3, based on the peak value of the bvp signal of each ROI extracted in step 3.1, the filtered signal is compared with the peak value of the spectrum, and the mean square error (MSE) of the spectrum is calculated using the function to express the degree of spectrum leakage, which is recorded as leakage, and the confidence is used. i = 1 / (1 + leakage) to calculate the BPM confidence of each ROI, where i is the ROI region identifier; based on the confidence and the instantaneous respiratory rate that does not jump calculated in step 3.2, the final respiratory rate result is calculated by weighted summation using the following formula: Step 4: Display the image captured by the camera in step 1 and the respiratory rate, confidence level, and weighted sum of the final respiratory rate of each ROI calculated in step 3 on the interface.

2. The detection system of the non-contact animal respiratory rate detection method based on rPPG according to claim 1, characterized in that: include: The image acquisition module is used in step 1 to capture images with a camera to collect images; The ROI confirmation module is used in step 2 to locate the ROI area of ​​the collected images; The signal processing module is used in step 3 to extract and process the signal of the ROI determined in step 2 to calculate the respiratory rate.

3. The detection device of the non-contact animal respiratory rate detection method based on rPPG according to claim 1, characterized in that: include: memory for storing computer programs; A processor is used to implement the rPPG-based non-contact animal respiratory rate detection method described in steps 1 to 4 when executing the computer program.

4. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it can perform rPPG-based non-contact animal respiration rate detection according to steps 1 to 4 of the rPPG-based non-contact animal respiration rate detection method according to claim 1.

Images