An image processing system for health monitoring during motion

Abstract

This invention relates to the field of artificial intelligence technology and provides an image processing system for health monitoring under motion conditions. The system includes: an image acquisition device for acquiring raw motion videos under specific motion conditions; and an image processing device for extracting and generating one or more corresponding motion video images from the raw motion videos based on the motion state of the subject under test. The extracted one or more corresponding motion video images are then subjected to noise reduction processing and used as input to a non-contact health monitoring device. This invention provides an image processing system for health monitoring under motion conditions. By acquiring raw motion videos of the subject under test under motion conditions using an image acquisition device, and then extracting raw motion videos under different motion conditions using an image processing device, the system integrates and generates motion video images, which are then input into a non-contact health monitoring algorithm model for calculation. This allows for more comprehensive monitoring of the subject's body.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an image processing system for health monitoring during motion. Background Technology

[0002] Health monitoring during exercise refers to the real-time monitoring and feedback of various health parameters of the test subject during exercise. This involves collecting video images of the test subject during exercise as input to a non-contact health detection AI algorithm. The algorithm then uses this mature algorithm to infer and analyze the test subject's physiological and psychological indicators. This method of health monitoring during exercise has broad application prospects and can be applied to multiple fields such as assisted physical activity and patient rehabilitation exercises. It has the function of real-time tracking of exercise status and using data to describe the exercise status. However, its application in everyday consumer goods is still rarely reported.

[0003] Chinese patent document CN112381011A discloses a non-contact heart rate measurement method, system, and device based on facial images. This invention combines CNN feature extraction and LSTM long short-term memory neural networks, and embeds a channel attention network to achieve non-contact measurement of human physiological parameters with low error rate and high efficiency. The Chinese patent document CN112381011A also discloses a non-contact heart rate measurement method based on facial video sequences. This method acquires video sequences containing human facial information and combines local texture features and skin color models of the images to detect and track the human facial region in real time to achieve non-contact measurement of human physiological parameters.

[0004] The above methods are mainly for non-contact health monitoring of stationary bodies. When the subject is in motion, the position of the image acquisition often changes, and the image and video are unstable, making monitoring impossible or the monitoring results inaccurate. How to conduct health monitoring during movement is an urgent problem to be solved. Summary of the Invention

[0005] To address the shortcomings of the existing technology, the present invention provides an image processing system for health monitoring during motion, comprising:

[0006] An image acquisition device used to acquire raw motion video under specific motion conditions;

[0007] An image processing device is used to extract and generate one or more corresponding motion video images from the original motion video based on the motion state of the object to be tested; and to input the extracted one or more corresponding motion video images, after noise reduction processing, into a non-contact health monitoring device.

[0008] Furthermore, the specific motion state includes motion speed, motion rate, or body posture.

[0009] Furthermore, it also includes an image preprocessing device for presetting a motion speed range value or a motion rate range value; and for acquiring only the original motion video within the time period that conforms to the motion speed range value or the motion rate range value as motion video images.

[0010] Furthermore, it also includes an image preprocessing device for calculating the image frame size and preseting a portrait-to-frame ratio; and for capturing only the original motion video within the preset portrait-to-frame ratio as motion video images.

[0011] Furthermore, it also includes an image preprocessing device for:

[0012] Set horizontal reference lines for specific parts of the test subject based on the subject's height;

[0013] Collect, extract, and integrate original motion videos of specific body parts of the subject being tested being above the horizontal reference line during motion, and generate the first motion video;

[0014] Collect, extract, and integrate original motion videos of specific body parts of the test subject being below the horizontal reference line during motion, and generate a second motion video;

[0015] The first motion video and / or the second motion video are used as the input source for the non-contact health monitoring device.

[0016] Furthermore, if the first motion video and the second motion video are used as the input sources for the non-contact health monitoring device, the non-contact health monitoring device outputs the health data obtained from the test corresponding to the first motion video and the health data obtained from the test corresponding to the second motion video, and finally performs average processing on the two data.

[0017] Furthermore, it also includes an image preprocessing device for calculating the frequency of the body movement of the test object based on the motion state of the test object; and setting an image acquisition interval time point according to the frequency, and acquiring motion video at the image acquisition interval time point.

[0018] Furthermore, if the lighting conditions are insufficient when acquiring images, the image preprocessing device is also used for:

[0019] The acquired motion video images are used as input sources to a non-contact health monitoring algorithm model to calculate health data indicators.

[0020] The detection is performed using contact medical devices, and the data obtained from the detection is compared with the data calculated by the algorithm.

[0021] Based on the comparison results, adjust the illumination and focusing parameters of the image acquisition device and perform another comparison;

[0022] Repeat the above steps until the data comparison error is within a certain range, then lock the supplementary lighting parameters of the image acquisition device to acquire motion video.

[0023] Furthermore, supplementary lighting devices are also provided on both sides of the image acquisition device. Each supplementary lighting device contains multiple independent supplementary lights with equal spacing, and the brightness of each supplementary light is adjustable. The supplementary lighting device determines that the light on a certain side is weaker by symmetrically comparing a specific part of the video image acquired by the image acquisition device. Then, it activates the supplementary lighting device on that side and dynamically and independently adjusts the brightness of each supplementary light on that side according to the symmetrical signal-to-noise ratio of a specific part of the video image acquired by the image acquisition device.

[0024] Furthermore, the system contains multiple image acquisition devices. The corresponding image acquisition devices are activated to acquire video images according to the different motion states of the test object. The acquired multiple video images are integrated to generate a motion video image as an input source to the non-contact health monitoring device.

[0025] Based on the above, compared with the prior art, the image processing system for health monitoring under motion conditions provided by the present invention acquires the original motion video of the test subject under motion conditions through an image acquisition device, and then the image processing device extracts the original motion video under different motion conditions, integrates and generates motion video images, and inputs them into a non-contact health monitoring algorithm model for calculation, which can more comprehensively monitor the body of the test subject.

[0026] Other features and beneficial effects of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects of the invention and other beneficial effects may be realized and obtained by means of the structures particularly pointed out in the description and claims. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. The technical features designed in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] In the description of this invention, it should be noted that all terms used in this invention (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains, and should not be construed as limiting the invention; it should be further understood that the terms used in this invention should be understood to have the same meaning as those in the context of this specification and in the relevant field, and should not be understood in an idealized or overly formal sense, except as expressly defined in this invention.

[0029] Example 1

[0030] An image processing system for health monitoring during motion includes:

[0031] An image acquisition device used to acquire raw motion video under specific motion conditions;

[0032] An image processing device is used to extract and generate one or more corresponding motion video images from the original motion video based on the motion state of the object to be tested; and to input the extracted one or more corresponding motion video images, after noise reduction processing, into a non-contact health monitoring device.

[0033] Specific motion states include motion speed, motion rate, or body posture.

[0034] Body posture refers to the state of the body and its parts at different stages of movement; movement speed refers to the displacement distance of the body or a part of the body per unit time, which is more applicable to exercises that require displacement changes, such as walking and running, in health monitoring; movement rate refers to the number of times a movement action is repeated per unit time, also known as movement frequency, which is more suitable for exercises that do not require displacement changes, such as skipping rope, lifting dumbbells, and lifting barbells, in health monitoring.

[0035] The image acquisition device acquires raw motion videos under specific motion states. Then, the image processing device extracts motion video images from the raw motion videos according to different motion states. For example, when the speed or rate is high, one motion video is extracted, and when the speed or rate is low, another video is extracted. One or more different motion video images can also be extracted according to different body postures during the movement. Then, the one or more corresponding motion video images generated are denoised and input into a non-contact health monitoring device for calculation to generate physiological health data of the test subject under motion state.

[0036] In this embodiment, an image acquisition device acquires raw motion videos of the test subject in motion. Subsequently, an image processing device extracts raw motion videos from different motion states, integrates them to generate motion video images, and inputs them into a non-contact health monitoring device for calculation. This allows for more comprehensive monitoring of the test subject's body.

[0037] Example 2

[0038] Different exercise states correspond to different health parameters, and health parameters obtained under specific exercise states can more accurately reflect the body's health status. For example, running at a high pace requires good cardiopulmonary function, so cardiopulmonary function can be better monitored at this time. In order to eliminate the interference of different exercise states on the monitoring results, this embodiment two is a further improvement on the first embodiment. Specifically, the image processing system for health monitoring under exercise states also includes:

[0039] An image preprocessing device is used to preset a range of motion speed values ​​or a range of motion rate values; and to collect only the original motion video within the time period that conforms to the range of motion speed values ​​or the range of motion rate values ​​as motion video images.

[0040] This embodiment focuses on running, and the exercise state refers to the running speed. However, it should be noted that the exercise is not limited to running. It can also be other aerobic exercises such as rope skipping and aerobics, as well as other types of anaerobic exercises. When the exercise is rope skipping, the exercise state refers to the rope skipping rate of the test subject per unit time. When the exercise is anaerobic exercise such as lifting dumbbells or barbells, it refers to the number of repetitions of a specific body posture of the test subject. The corresponding exercise state is selected for video recording according to the type of exercise.

[0041] The image acquisition device acquires motion videos (running videos) of the test subject. In this embodiment, speeds less than 9 km / h are set as low-pace running range values, speeds between 9 km / h and 12 km / h are set as medium-pace running range values, and speeds greater than 12 km / h are set as high-pace running range values. The image preprocessing device can be set to only acquire motion videos that meet the high-pace running range values, i.e., running motion video images with a pace greater than 12 km / h. The corresponding images are extracted from the original motion videos and input into the image processing device to merge and generate motion video images. These images are then input into a non-contact health monitoring device for detection to obtain the physiological health data of the test subject under exercise conditions. The image preprocessing device can also be set to only acquire original motion videos corresponding to the low-pace running range values ​​or the medium-pace running range values; the specific settings depend on the data to be acquired.

[0042] In this embodiment, the image preprocessing device acquires the original motion video of the test subject and extracts and integrates the motion video images with preset motion speed range values ​​or motion rate range values ​​to detect the physiological health data of the test subject. The data obtained by detecting the same motion state will not differ too much, and the data will not affect each other, making the calculated physiological health data more accurate.

[0043] To verify the effectiveness of this embodiment, the present invention also provides a comparison data table of the image processing system with image preprocessing device of the present invention and the invention patent method for monitoring exercise heart rate as described in the table below. The comparison data is obtained by simultaneously monitoring the same test subject in different running states. Among them, the heart rate data obtained by monitoring the same test subject in the same exercise state using a health monitoring contact wristband is used as a control. Three sets of comparison results were obtained for three test subjects.

[0044] The comparative example used in this embodiment is the exercise heart rate monitored by the invention patent method with publication number CN112381011A (hereinafter referred to as the comparative prior art).

[0045] Table 1 shows the heart rate data of the test subjects during low-pace running.

[0046] Table 1

[0047]

[0048] Table 2 shows the heart rate data of the subjects during high-pace running.

[0049] Table 2

[0050]

[0051] Tables 1 and 2 show that the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the traditional non-contact monitoring method without this system differs from the heart rate monitoring results measured by the contact wristband by approximately ±10. Therefore, it can be seen that the image processing system for health monitoring under exercise provided in this application can greatly improve the accuracy of health monitoring results under exercise.

[0052] Example 3

[0053] During the movement of the test subject, the distance between the test subject and the video acquisition device will change. For example, when applied to outdoor long-distance running, the distance between the test subject and the screen will change from far to near. When the test subject is too far from the acquisition device, the acquired image will be blurry and cannot yield an accurate result. To solve the problem of inaccurate detection results caused by the test subject being too far from the video image acquisition device during the video image acquisition process, this invention provides an embodiment three based on embodiment one. Specifically, the image processing system for health monitoring in motion state further includes:

[0054] An image preprocessing device is used to calculate the image frame size and preset a portrait frame ratio; only raw motion video within the preset portrait frame ratio is captured as motion video images.

[0055] In practical implementation, such as in an outdoor long-distance running project, the test subject runs towards the video image acquisition device from a distance. At this time, the size of the image captured by the image acquisition device is fixed. Therefore, the proportion of the test subject's image in the video image will gradually change as the running progresses. In order to ensure the accuracy of the results, the image preprocessing device calculates the image size and sets a preset human-image ratio. In this embodiment, the preset human-image ratio is 1:10, that is, the area of ​​the human image in the video image is one-tenth. After setting the preset human-image ratio, the image preprocessing device only extracts the original motion video with a ratio greater than the preset human-image ratio as the motion video image. At this time, the test results are no longer affected by the distance between the test subject and the video image acquisition device.

[0056] It should be noted that the preset portrait aspect ratio is not limited to 1:10. It can be determined according to the pixel size of the video image acquisition device. If the video image acquisition device acquires a high pixel size, the preset portrait aspect ratio can be set relatively smaller. If the video image acquisition device acquires a low pixel size, the preset portrait aspect ratio can be set larger.

[0057] In this embodiment, the image preprocessing device calculates the image size and sets a preset human-image ratio so that the detection results are no longer affected by the distance between the test object and the video image acquisition device during the video image acquisition process.

[0058] To verify the effectiveness of this embodiment, the present invention also provides a comparison data table of the image processing system with image preprocessing device of the present invention and the invention patent method for monitoring exercise heart rate as described in the table below. The comparison data is obtained by simultaneously monitoring the same test subject in different running states. Among them, the heart rate data obtained by monitoring the same test subject in the same exercise state using a health monitoring contact wristband is used as a control. Three sets of comparison results were obtained for three test subjects.

[0059] The comparative example used in this embodiment is the exercise heart rate monitored by the invention patent method with publication number CN112381011A (hereinafter referred to as the comparative prior art).

[0060] Table 3 shows the heart rate data of the test subjects during outdoor running.

[0061] Table 3

[0062]

[0063] As can be seen from Table 3, the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the traditional non-contact monitoring method without this system differs from the heart rate monitoring results measured by the contact wristband by approximately ±10. Therefore, it can be seen that the image processing system for health monitoring under exercise provided in this application can greatly improve the accuracy of health monitoring results under exercise.

[0064] Example 4

[0065] During activities like rope skipping or burpees, people jump up and down. In the captured motion video, the human body is at different heights in the video image. Moreover, the position of the test subject in the video image changes in real time during the jumping process. Existing non-contact monitoring algorithms mostly use the face as the main detection object to calculate physiological health data. Therefore, it is necessary to locate the face in real time during the monitoring process. The position of the face changes little when it is stationary, but the real-time changes in the relative position of the face during movement increase the algorithm's calculation time, resulting in inaccurate calculation results.

[0066] To address the technical problem of increased monitoring time and inaccurate monitoring results due to real-time changes in the human body's position in motion videos, this invention provides Embodiment Four based on Embodiment One. Specifically, the image processing system for health monitoring during motion further includes an image preprocessing device for:

[0067] Set horizontal reference lines for specific parts of the test subject based on the subject's height;

[0068] Collect, extract, and integrate original motion videos of specific body parts of the subject being tested being above the horizontal reference line during motion, and generate the first motion video;

[0069] Collect, extract, and integrate original motion videos of specific body parts of the test subject being below the horizontal reference line during motion, and generate a second motion video;

[0070] The first motion video and / or the second motion video are used as the input source for the non-contact health monitoring device.

[0071] In this embodiment, the image preprocessing device first monitors the height of the test subject and sets a horizontal reference line at a specific part of the test subject in a static state. Preferably, in this embodiment, the horizontal reference line is set at the nose of the test subject. During jumping, the image preprocessing device extracts and integrates the motion video of the nose being higher than the horizontal reference line to generate a first motion video, and integrates the motion video of the nose being lower than the horizontal reference line to generate a second motion video.

[0072] In practice, the first motion video can be obtained by extracting and integrating the original motion video that conforms to the first motion state, such as a motion video image of a body part exceeding the horizontal reference line, and used as the input source of the non-contact health monitoring device. Alternatively, the second motion video can be obtained by extracting the original motion video that conforms to the second motion state, such as a motion video image of a body part below the horizontal reference line, and used as the input source of the non-contact health monitoring device.

[0073] Based on the specific application scenario, the motion video image with better video quality from either the first or second motion video, which yields more accurate results, can be selected as the input source.

[0074] This embodiment uses an image preprocessing device to set a horizontal reference line and integrates facial motion videos from different heights. This allows the algorithm to quickly locate the face and calculate physiological health parameters, effectively shortening the duration of health monitoring during movement and improving the accuracy of health monitoring results.

[0075] To verify the effectiveness of this embodiment, the present invention also provides a comparison data table of the image processing system with image preprocessing device of the present invention and the invention patent method for monitoring exercise heart rate as described in the table below. The comparison data is obtained by simultaneously monitoring the same test subject in different running states. Among them, the heart rate data obtained by monitoring the same test subject in the same exercise state using a health monitoring contact wristband is used as a control. Three sets of comparison results were obtained for three test subjects.

[0076] The comparative example used in this embodiment is the exercise heart rate obtained by the monitoring method of the invention patent with publication number CN112381011A (hereinafter referred to as the comparative prior art).

[0077] Table 4 shows the heart rate data of the test subjects during slow rope skipping.

[0078] Table 4

[0079]

[0080] Table 5 shows the heart rate data of the subjects during rapid rope skipping.

[0081] Table 5

[0082]

[0083] Table 5

[0084] As can be seen from Tables 4 and 5, the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the traditional non-contact monitoring method without this system differs from the heart rate monitoring results measured by the contact wristband by approximately ±10. Therefore, it can be seen that the image processing system for health monitoring under exercise provided in this application can greatly improve the accuracy of health monitoring results under exercise.

[0085] Example 5

[0086] Based on Embodiment 4, the present invention also provides Embodiment 5, which is as follows:

[0087] If the first motion video and the second motion video are used as the input sources for the non-contact health monitoring device, the non-contact health monitoring device outputs the health data obtained from the test corresponding to the first motion video and the health data obtained from the test corresponding to the second motion video, and finally the two data are averaged.

[0088] The image preprocessing device integrates and stitches the first or second video image as the input source, so that all the original motion images acquired are used, ensuring that no motion state of the object under test is ignored.

[0089] Example 6

[0090] Based on Embodiment 1, this invention also provides Embodiment 6, specifically, an image processing system for health monitoring during motion further includes:

[0091] An image preprocessing device is used to calculate the frequency of the body movement of the test object based on the motion state of the test object; and to set the image acquisition interval time point according to the frequency, and to acquire motion video at the image acquisition interval time point.

[0092] During the exercise, the frequency of the test subject's movement will not change significantly over a period of time. The system is also equipped with an image preprocessing device, which can calculate the frequency of the test subject's body movement based on the test subject's movement state, set the image acquisition interval time point according to the frequency, and acquire motion video at the image acquisition interval time point.

[0093] For example, during running, the human body moves up and down with the movement. The current running frequency can be collected; for instance, if the body moves to the same posture and height within a one-second interval, the image acquisition interval can be set to one second, capturing motion video every one second. Furthermore, as the running speed gradually increases, the image preprocessing device can adjust the image acquisition interval accordingly based on the monitored running speed.

[0094] To verify the effectiveness of this embodiment, the present invention also provides a comparison data table of the image processing system with image preprocessing device of the present invention and the invention patent method for monitoring exercise heart rate as described in the table below. The comparison data is obtained by simultaneously monitoring the same test subject in different running states. Among them, the heart rate data obtained by monitoring the same test subject in the same exercise state using a health monitoring contact wristband is used as a control. Three sets of comparison results were obtained for three test subjects.

[0095] The comparative example used in this embodiment is the exercise heart rate monitored by the invention patent method with publication number CN112381011A (hereinafter referred to as the comparative prior art).

[0096] Table 6 shows the heart rate data of the test subjects during low-paced, steady-speed running.

[0097] Table 6

[0098]

[0099] As can be seen from Table 6, the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the traditional non-contact monitoring method without this system differs from the heart rate monitoring results measured by the contact wristband by approximately ±10. Therefore, it can be seen that the image processing system for health monitoring under exercise provided in this application can greatly improve the accuracy of health monitoring results under exercise.

[0100] Example 7

[0101] Existing non-contact health monitoring devices mostly calculate human health data based on weak signal changes on the skin surface of the test subject. Therefore, whether the lighting conditions of the acquired images meet the requirements of non-contact health monitoring devices is crucial. However, in reality, due to the unevenness of natural light sources, the acquired images often appear too dark, affecting the detection results. Based on any of the above embodiments two to six, this invention also provides embodiment seven. In this embodiment, the image preprocessing device, if the lighting conditions are insufficient during image acquisition, is further used for:

[0102] The acquired motion video images are used as input sources and fed into a non-contact health monitoring device to calculate health data indicators.

[0103] The detection is performed using contact medical devices, and the data obtained from the detection is compared with the data calculated by the algorithm.

[0104] Based on the comparison results, adjust the illumination and focusing parameters of the image acquisition device and perform another comparison;

[0105] Repeat the above steps until the data comparison error is within a certain range, then lock the fill light and focus parameters of the image acquisition device to acquire motion video.

[0106] In practice, motion video images are first acquired using the image acquisition device of this system and used as input to a non-contact health monitoring device to calculate health data indicators. While the system is acquiring data, a contact medical device, such as a contact fitness tracker, is used to detect the human body. The data obtained from the contact fitness tracker is compared with the data obtained from the system. Based on the comparison results, the image preprocessing device adjusts the illumination and focusing parameters of the image acquisition device for another comparison. This process is repeated until the data comparison error is within a certain range. At this point, the illumination and focusing parameters of the image acquisition device are determined for acquiring motion video.

[0107] To verify the effectiveness of this embodiment, the present invention also provides a comparison data table below, showing the image processing system of the present invention with an image preprocessing device and the system without an image preprocessing device for adjusting the supplementary lighting parameters to monitor exercise heart rate. This table represents the comparison data obtained when the same test subject is monitored in different running states at the same time. Among them, the heart rate data obtained by a health monitoring contact bracelet monitoring the same test subject in the same exercise state is used as a control. Three test subjects were tested, and three sets of comparison results were obtained.

[0108] The comparative example used in this embodiment is the monitoring of exercise heart rate without the addition of an image preprocessing device to adjust the supplementary lighting parameters (hereinafter referred to as "comparison technology").

[0109] Table 7 shows the heart rate data of the test subjects during outdoor running.

[0110] Table 7

[0111]

[0112] As shown in Table 7, the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the system without image preprocessing device to adjust the supplementary light parameters differs from the heart rate monitoring results measured by the contact wristband by approximately ±15. This demonstrates that adding an image preprocessing system to the system to adjust the supplementary light parameters can greatly improve the accuracy of health monitoring results during exercise.

[0113] Example 8

[0114] The above-mentioned embodiment seven involves the system's built-in image preprocessing device adjusting the supplementary lighting parameters of the image acquisition device. This embodiment eight is an improvement on any one of embodiments one through six, specifically as follows:

[0115] The system also includes supplementary lighting devices on both sides of the image acquisition device. Each supplementary lighting device contains multiple independent supplementary lights with equal spacing, and the brightness of each supplementary light is adjustable. The supplementary lighting device determines that the lighting on a certain side is weaker by symmetrically comparing a specific part of the video image acquired by the image acquisition device (such as a human face). Then, it activates the supplementary lighting device on that side and dynamically and independently adjusts the brightness of each supplementary light on that side based on the symmetrical signal-to-noise ratio of the specific part of the video image acquired by the image acquisition device.

[0116] In practice, the image processing device analyzes the motion video captured by the image acquisition device to generate the motion state of the test object. Taking running as an example, the posture and height of the human body change during running, resulting in different motion states. The brightness of the captured images varies depending on the motion state, such as the so-called "half-face" (one half of the face is illuminated by natural light, while the other half is in shadow). The supplementary lighting device with a dynamically adjustable angle structure can dynamically adjust the angle to provide supplementary lighting according to the motion state. Since the supplementary lighting device contains multiple independent supplementary lights with equal spacing, the supplementary lights on the shadowed half of the face can be independently adjusted to different intensities based on the angle, simulating the natural light source of the other half and thus solving the problem of image unevenness caused by this situation.

[0117] Preferably, the supplemental lighting device is also equipped with a polarizing mirror.

[0118] The polarizer and the supplementary lighting device are electrically connected to a dynamically adjustable structure. During use, the polarizer is automatically adjusted according to the brightness of the captured face image. It can also be manually adjusted by the test subject to make it more suitable for the system to capture video images.

[0119] To verify the effectiveness of this embodiment, the present invention also provides a comparison table of heart rate monitoring data between the image processing system with image preprocessing device and the system without supplementary lighting device, as shown in the table below. This table represents the comparison data obtained when the same test subject is monitored in different running states at the same time. Among them, the heart rate data obtained by the health monitoring contact bracelet monitoring the same test subject in the same exercise state is used as a control. Three test subjects were tested and three sets of comparison results were obtained.

[0120] The comparative example used in this embodiment is the exercise heart rate monitored by this system without the addition of a supplementary lighting device (hereinafter referred to as "comparison technology").

[0121] Table 8 shows the heart rate data of the test subjects during outdoor running.

[0122] Table 8

[0123]

[0124] As can be seen from Table 8, the image processing system with image preprocessing device in this embodiment differs from the heart rate monitoring results measured by the contact wristband by approximately ±5, while the system without supplementary lighting device differs from the heart rate monitoring results measured by the contact wristband by approximately ±15. This shows that adding an image preprocessing system that adjusts the supplementary lighting parameters to the system can greatly improve the accuracy of health monitoring results during exercise.

[0125] Example 9

[0126] During rope skipping or burpees, the relative positions of body parts change significantly. If only one image acquisition device is used for recording, the captured motion state may be incomplete due to the limited shooting range of the acquisition device. The test subject may also move out of the image acquisition device's range due to changes in motion. Therefore, this invention proposes Embodiment Nine, an improvement on Embodiment One, specifically:

[0127] The image processing system for health monitoring during motion contains multiple image acquisition devices. The corresponding image acquisition device is activated to acquire video images according to the different motion states of the test object. The acquired video images are then integrated to generate a single motion video image, which is then input into the non-contact health monitoring device.

[0128] Multiple image acquisition devices can be set up at different heights or even different angles to facilitate the acquisition of all motion states of the object under test. The image acquisition device with the highest video quality can be activated to acquire video images based on the motion state. Then, the video images acquired by different acquisition devices are integrated to generate a motion video image as the input source to the non-contact health monitoring device.

[0129] Furthermore, those skilled in the art should understand that although many problems exist in the prior art, each embodiment or technical solution of the present invention can be improved in only one or a few aspects, without necessarily solving all the technical problems listed in the prior art or the background art simultaneously. Those skilled in the art should understand that any content not mentioned in a claim should not be construed as a limitation on that claim.

[0130] Although this document frequently uses terms such as original motion video, first motion state, first motion video, second motion state, and second motion video, the possibility of using other terms is not excluded. These terms are used merely for the convenience of describing and explaining the essence of the invention; interpreting them as any additional limitation would contradict the spirit of the invention.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An image processing system for health monitoring in motion, characterized by, include: An image acquisition device is used to acquire raw motion videos under specific motion conditions, including motion speed, motion rate, or body posture. An image processing device is used to extract and generate one or more corresponding motion video images from the original motion video based on the motion state of the object under test; and to input the extracted one or more corresponding motion video images, after noise reduction processing, into a non-contact health monitoring device. An image preprocessing device is used to preset a range of motion speed values ​​or a range of motion rate values; and to collect only the original motion video within the time period that conforms to the range of motion speed values ​​or the range of motion rate values ​​as motion video images.

2. The image processing system for health monitoring in motion state according to claim 1, characterized in that The image preprocessing unit is also used to calculate the image frame size and preset a portrait frame ratio; only the original motion video within the preset portrait frame ratio is collected as motion video images.

3. The image processing system for health monitoring in motion state according to claim 1, wherein, The image preprocessing device is also used for: Set horizontal reference lines for specific parts of the test subject based on the subject's height; Collect, extract, and integrate original motion videos of specific body parts of the subject being tested being above the horizontal reference line during motion, and generate the first motion video; Collect, extract, and integrate original motion videos of specific body parts of the test subject being below the horizontal reference line during motion, and generate a second motion video; The first motion video and / or the second motion video are used as the input source for the non-contact health monitoring device.

4. The image processing system for health monitoring in motion state according to claim 3, wherein, If the first motion video and the second motion video are used as the input sources for the non-contact health monitoring device, the non-contact health monitoring device outputs the health data obtained from the test corresponding to the first motion video and the health data obtained from the test corresponding to the second motion video, and finally the two data are averaged.

5. The image processing system for health monitoring in motion state according to claim 1, wherein, The image preprocessing device is also used to calculate the frequency of the body movement of the test object based on the motion state of the test object; and to set the image acquisition interval time point according to the frequency, and to acquire motion video at the image acquisition interval time point.

6. The image processing system for health monitoring in motion according to any of claims 1-5, characterized in that, If the lighting conditions are insufficient when acquiring images, the image preprocessing device is also used for: The acquired motion video images are used as input sources to a non-contact health monitoring algorithm model to calculate health data indicators. The detection is performed using contact medical devices, and the data obtained from the detection is compared with the data calculated by the algorithm. Based on the comparison results, adjust the illumination and focusing parameters of the image acquisition device and perform another comparison; Repeat the above steps until the data comparison error is within a certain range, then lock the fill light and focus parameters of the image acquisition device to acquire motion video.

7. The image processing system for health monitoring in motion according to any of claims 1-5, characterized in that, A supplementary lighting device is also provided on both sides of the image acquisition device. The supplementary lighting device contains multiple independent supplementary lights with equal spacing. The brightness of the supplementary lights is adjustable. The supplementary lighting device determines that the light on a certain side is weaker by symmetrically comparing a specific part of the video image acquired by the image acquisition device. Then, it activates the supplementary lighting device on that side and dynamically and independently adjusts the brightness of each supplementary light on that side according to the symmetrical signal-to-noise ratio of the specific part of the video image acquired by the image acquisition device.

8. The image processing system for health monitoring in motion state according to claim 1, wherein, The system contains multiple image acquisition devices. The corresponding image acquisition device is activated to acquire video images according to the different motion states of the object under test. The acquired video images are integrated to generate a motion video image as the input source to the non-contact health monitoring device.

Images