TWI795765B - Non-contact physiological signal measurement apparatus, system and method - Google Patents

Abstract

A non-contact physiological signal measurement apparatus, system and method are provided. The non-contact physiological signal measurement apparatus includes a storage device, an image signal sensing element, and a processor. The storage device is used to store a plurality of modules. The image signal sensing element is used to obtain continuous images. The processor is coupled to the image signal sensing element and the storage device. The processor is used to execute the plurality of modules to detect skin images from the continuous images, and extract a skin reflection signal from the skin images in the continuous images. The processor calculates a photoplethysmography signal based on the skin reflection signal, and calculates physiological information based on the photoplethysmography signal.

Description

Non-contact physiological signal measurement equipment, system and method

The present invention relates to the technical field of physiological information identification, in particular to a measuring device, system and method for providing non-contact and comprehensive physiological information.

With the improvement of medical level, people's attention to health also increases. The most basic way to evaluate the health status of the body is through the measurement of physiological characteristic values. That is to say, the current state of health is judged through the data of physiological characteristic values. In particular, patients with chronic diseases need to monitor their physiological characteristic values at all times. However, the current physiological characteristic measurement equipment usually needs to wear the measurement equipment on a certain part of the body of the person to be measured in order to collect data.

On the other hand, with the phenomenon of aging, the demand for long-term care is also increasing. In this regard, long-term care often requires a large amount of manpower to provide immediate care for the elderly. Therefore, the shortage of long-term care manpower is also one of the main problems of long-term care. In view of this, how to establish an automated care system to reduce the demand for long-term care manpower and provide immediate care functions, the following will propose solutions in several embodiments.

The present invention provides a non-contact physiological signal measuring device, system and method, which can effectively acquire physiological parameters in a non-contact manner.

The non-contact physiological signal measuring device of the present invention includes a storage device, an image signal sensing element and a processor. The storage device is used for storing multiple modules. The image signal sensing element is used to obtain continuous images. The processor is coupled to the image signal sensing element and the storage device. The processor is used to execute the multiple modules to detect the skin image from the continuous image, and extract the skin reflection signal from the skin image in the continuous image, and the processor calculates the photoplethysmography signal according to the skin reflection signal. The processor calculates physiological information according to the photoplethysmographic signal, and the physiological information includes at least one of an estimated heartbeat curve, an estimated heartbeat variation curve, an estimated respiration value, an estimated blood oxygen value, and an estimated blood pressure value.

The non-contact physiological signal measuring system of the present invention includes a monitoring host and a plurality of measuring devices. Multiple measuring devices are coupled to the monitoring host. The plurality of measurement devices are used for non-contact physiological signal measurement of a plurality of measurement objects, and are used to determine whether to output corresponding warning signals to the monitoring host, so that the monitoring host monitors the plurality of physiological information and corresponding warning signs. Each of the plurality of measuring devices is used to detect a respective skin image in the continuous images of the corresponding measurement object, extract the skin reflection signal from the skin image, and calculate the photoplethysmography from the skin reflection signal signal. Each of the plurality of measurement devices calculates corresponding physiological information from the photoplethysmographic signal, wherein the physiological information includes an estimated heartbeat curve, an estimated heartbeat variation curve, an estimated respiration value, an estimated blood oxygen value, and an estimated blood pressure value. at least one of them.

The non-contact physiological signal measurement method of the present invention includes the following steps: obtaining continuous images through the image signal sensing element; detecting skin images from the continuous images, and extracting skin reflection signals from the skin images in the continuous images; calculating a photoplethysmographic signal from the reflected signal; and calculating physiological information based on the photoplethysmographic signal, wherein the physiological information includes at least one of an estimated heartbeat curve, an estimated heartbeat variation curve, an estimated respiration value, an estimated blood oxygen value, and an estimated blood pressure value one.

Based on the above, the non-contact physiological signal measurement device, system and method of the present invention can provide a non-contact physiological signal measurement function, and can automatically generate various physiological signal physiological information of one or more measurement targets.

In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail together with the accompanying drawings.

100: Non-contact physiological signal measurement equipment

110: Processor

120: image signal sensing element

130: storage device

300_1~300_N: continuous image

301: Human skin area

302:Environmental area

310: heartbeat estimation curve

320:Heartbeat variation estimation curve

330: respiration estimate

340: blood oxygen estimate

350: blood pressure estimate

401: Skin reflex signal

402: Equidistant reflection signal

403: Photoplethysmography signal

410: Skin reflection signal acquisition module

411: Skin detection unit

412: Skin reflection signal acquisition unit

420: Skin reflection signal calculation module

421: equal interval signal sampling unit

422: Signal conversion unit

430: Physiological signal analysis module

431: Heartbeat signal calculation unit

432:Respiratory signal calculation unit

433: blood oxygen signal calculation unit

434: Blood pressure signal calculation unit

600: Non-contact physiological signal measurement equipment

610: Monitor host

620_1~620_N: Measuring equipment

S210~S240, S510~S593: steps

FIG. 1 is a schematic diagram of a non-contact physiological signal measuring device according to an embodiment of the present invention.

FIG. 2 is a flowchart of a non-contact physiological signal measurement method according to an embodiment of the present invention.

FIG. 3 is a schematic measurement diagram of an embodiment of the present invention.

FIG. 4 is a schematic diagram of a plurality of modules according to an embodiment of the present invention.

FIG. 5 is a flowchart of a non-contact physiological signal measurement method according to another embodiment of the present invention.

FIG. 6 is a schematic diagram of a non-contact physiological signal measurement system according to an embodiment of the present invention.

In order to make the content of the present invention more comprehensible, the following specific embodiments are taken as examples in which the present invention can actually be implemented. In addition, where possible, elements/components/steps using the same reference numerals in the drawings and embodiments represent the same or similar parts.

FIG. 1 is a schematic diagram of a non-contact physiological signal measuring device according to an embodiment of the present invention. Referring to FIG. 1 , the non-contact physiological signal measurement device 100 may include a processor 110 , an image signal sensing element 120 and a storage device 130 . The processor 110 is coupled to the image signal sensing element 120 and the storage device 130 . In this embodiment, the storage device 130 can be used to store multiple modules to realize image processing, analysis and signal processing in various embodiments of the present invention. The image signal sensing element 120 may include a visible light (Visible light) and/or an infrared light (Infrared; IR) camera. In this embodiment, the non-contact physiological signal measurement device 100 can use the image signal sensing element 120 to measure the physiological information of the measurement target in a non-contact manner.

In this embodiment, the processor 110 may include, for example, a central processing unit (Central Processing Unit; CPU), a microprocessor (Microprocessor Control Unit; MCU) or a field programmable gate array (Field Programmable Gate Array; FPGA), but the present invention is not limited thereto. The processor 110 can drive and control the image signal sensing element 120 , and the processing processor 110 can access the storage device 130 and execute multiple modules.

In this embodiment, the storage device 130 may include, for example, random access memory (Random-Access Memory; RAM), read-only memory (Read-Only Memory; ROM), optical disc (Optical disc), magnetic disk (Magnetic disk) ), Hard drive, Solid-state drive, Flash drive, Security digital (SD) card, Memory stick, Compact flash ( Compact flash; CF card or any type of storage device, but the present invention is not limited thereto. In this embodiment, the storage device 130 can be used to store multiple modules (software programs or firmware) and image data described in various embodiments of the present invention.

In this embodiment, the non-contact physiological signal measurement device 100 can be a portable sensing device, for example, the user can hold the non-contact physiological signal measurement device 100 to measure one or more sensing devices. The object measures physiological information in a non-contact sensing (image sensing) manner. Alternatively, in an embodiment, the non-contact physiological signal measurement device 100 may be a fixed monitoring device. The image signal sensing element 120 can be disposed around the bed, for example, to monitor the sensing objects on the bed in real time.

FIG. 2 is a flowchart of a non-contact physiological signal measurement method according to an embodiment of the present invention. FIG. 3 is a schematic measurement diagram of an embodiment of the present invention. Referring to FIG. 1 to FIG. 3, the non-contact physiological signal measurement device 100 can perform the steps shown in FIG. 2 S210-S240, to realize non-contact physiological signal measurement. In step S210 , the non-contact physiological signal measuring device 100 can obtain continuous images 300_1 - 300_N through the image signal sensing element 120 , wherein N is a positive integer. The consecutive images 300_1˜300_N may include a plurality of visible light images and/or a plurality of infrared light images. In step S220, the non-contact physiological signal measurement device 100 can detect skin images from the continuous images 300_1˜300_N, and extract skin reflection signals from the skin images in the continuous images. The continuous images 300_1 - 300_N respectively include a human skin area 301 and an environment area 302 . In step S230, the non-contact physiological signal measurement device 100 can calculate a photoplethysmography (PPG) signal according to the skin reflection signal. In step S240, the non-contact physiological signal measurement device 100 can calculate physiological information according to the photoplethysmographic signal. In this embodiment, the physiological information may include at least one of an estimated heartbeat curve 310 , an estimated heartbeat variation curve 320 , an estimated respiration value 330 , an estimated blood oxygen value 340 , and an estimated blood pressure value 350 . Alternatively, the physiological information may include combined information of at least two of the estimated heartbeat curve 310 , the estimated heartbeat variation curve 320 , the estimated respiration value 330 , the estimated blood oxygen value 340 , and the estimated blood pressure value 350 . Therefore, the non-contact physiological signal measurement device 100 of this embodiment can realize automatic (or called passive) non-contact physiological signal measurement without the need for the user to manually operate the measurement device. In addition, an exemplary embodiment will be presented below to describe in detail how to generate the human skin region 301 , the photoplethysmography signal, and the physiological information in this embodiment.

FIG. 4 is a schematic diagram of a plurality of modules according to an embodiment of the present invention. FIG. 5 is a flowchart of a non-contact physiological signal measurement method according to another embodiment of the present invention. ginseng Referring to Figure 1, Figure 2, Figure 4 and Figure 5, in some embodiments of the present invention, the storage device 130 may include a skin reflection signal acquisition module 410, a skin reflection signal calculation module 420, and a physiological signal analysis module 430 . In this embodiment, the skin reflection signal acquisition module 410 may include a skin detection unit 411 and a skin reflection signal acquisition unit 412 . The skin reflection signal calculation module 420 may include an equally spaced signal sampling unit 421 and a signal conversion unit 422 . The signal converting unit 422 may be a Photoplethysmography (PPG) unit, a Remote-Photoplethysmography (rPPG) unit or an Image-Photoplethysmography (iPPG) unit. The physiological signal analysis module 430 may include a heartbeat signal calculation unit 431 , a respiratory signal calculation unit 432 , a blood oxygen signal calculation unit 433 , and a blood pressure signal calculation unit 434 . The non-contact physiological signal measurement device 100 of the present invention can realize the image recognition operation and light volume of the human skin area 301 described in the present invention by executing the software programs, firmware or algorithms of the above-mentioned multiple modules and units. A signal generation operation of the change tracing signal and an information generation operation of the physiological information. However, the storage device 130 of the present invention is not limited to storing software programs, firmware or algorithms of the above-mentioned multiple modules and units. In some other embodiments of the present invention, the non-contact physiological signal measurement device 100 can also realize the image recognition operation, signal generation operation and information described in the present invention by executing other known software programs, firmware or algorithms. generate operations.

In this embodiment, the non-contact physiological signal measurement device 100 may execute steps S510-S593 as shown in FIG. 5 , and in step S510, the processor 110 may acquire continuous images 300_1-300_N. In step S520, the processor 110 may execute the skin reflection signal The skin detection unit 411 of the number extraction module 410 detects skin images from the continuous images 300_1~300_N. The skin detection unit 411 can be a deep learning (Deep Learning) module, such as a neural network (Neural Network, NN) module, and can recognize human body regions in images after training, especially skin regions of human bodies can be recognized . In step S530, the processor 110 may execute the skin reflection signal acquisition unit 412 of the skin reflection signal acquisition module 410 to extract the skin reflection signals from the continuous images 300_1˜300_N. The skin reflection signal acquisition module 410 can provide the skin reflection signal 401 to the skin reflection signal calculation module 420 .

In step S540, the processor 110 can execute the equidistant signal sampling unit 421 of the skin reflection signal calculation module 420 to convert the equidistant reflection signals with equal time intervals from the skin reflection signals 401 whose time intervals are not equidistant. 402. In step S550, the processor 110 can execute the signal conversion unit 422 of the skin reflection signal calculation module 420 to reconstruct the equidistant reflection signal 402 into a photoplethysmographic signal 403 with pulse information according to different wavelength signals. The skin reflection signal calculation module 420 can provide the photoplethysmography signal 403 to the physiological signal analysis module 430 .

In step S561, the processor 110 can execute the heartbeat signal calculation unit 431 of the physiological signal analysis module 430 to convert the photoplethysmography signal 403 into the frequency domain and perform band filtering operation to extract the frequency with the highest amplitude as the heartbeat estimate. value. In step S562, the heartbeat signal calculation unit 431 can transform the signal of the adjacent part of the estimated heartbeat value back to the time domain to generate an estimated heartbeat curve. In step S563, the heartbeat signal calculation unit 431 can calculate the heartbeat intensity variation curve according to the heartbeat estimation curve Line (intensity variation), heartbeat amplitude variation curve (amplitude variation), and heartbeat frequency variation curve (frequency variation). The processor 110 can provide the variation curve of the heartbeat intensity, the variation curve of the heartbeat amplitude and the variation curve of the heartbeat frequency to the respiratory signal calculation unit 432 . In step S564, the heartbeat signal calculation unit 431 can output the estimated heartbeat curve and the estimated heartbeat variation curve. The estimated heartbeat variation curve includes at least one of the aforementioned heartbeat intensity variation curve, heartbeat amplitude variation curve, and heartbeat frequency variation curve.

In step S571, the processor 110 can execute the respiratory signal calculation unit 432 of the physiological signal analysis module 430 to calculate the heartbeat intensity variation curve, the heartbeat amplitude variation curve according to the heartbeat intensity variation curve, the heartbeat amplitude variation curve, and the heartbeat frequency variation curve And the respective peak numbers of the heartbeat frequency variation curve per unit time. In step S572, the respiration signal calculation unit 432 can generate three respiration estimation values according to the respective peak numbers of the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve in a unit time. In step S573, the respiration signal calculation unit 432 may output the estimated respiration value 330 (output at least one of the above-mentioned estimated respiration values).

In step S581 , the processor 110 can execute the blood oxygen signal calculation unit 433 of the physiological signal analysis module 430 to extract the first reference signal and the second reference signal with different wavelengths from the photoplethysmography signal 403 . In step S582, the blood oxygen signal calculation unit 433 can multiply the ratio of the two signal values of the first reference signal and the second reference signal by a preset coefficient to obtain the difference with the maximum blood oxygen concentration, and calculate the maximum blood oxygen concentration The difference is subtracted from the concentration to produce a blood oxygen estimate. In step S583, The blood oxygen signal calculation unit 433 can output an estimated blood oxygen value.

In step S591, the processor 110 can execute the blood pressure signal calculation unit 434 to perform quadratic differentiation on the estimated heartbeat curve to generate a third reference signal. In step S592, the blood pressure signal calculation unit 434 can obtain the signal features of every two adjacent valley intervals in the third reference signal, convert the signal features into blood pressure signals through a regression method, and generate blood pressure estimation values based on the blood pressure signals. In step S593, the blood pressure signal calculation unit 434 may output an estimated blood pressure value.

Therefore, the non-contact physiological signal measurement device 100 and the non-contact physiological signal measurement method of this embodiment can provide a non-contact physiological signal measurement function, and can automatically generate a variety of physiological signals of the measurement target. Information.

In addition, regarding the above-mentioned skin detection unit 411, during the manufacturing process of the skin detection unit 411, the device manufacturer can first train the deep learning module corresponding to the skin detection unit 411, so that the deep learning module can It has multiple continuous sensing images that can be computed and analyzed. The device manufacturer can install or write the trained deep learning module into the storage device 130 . Alternatively, the device manufacturer can also pre-train the deep learning module by operating the processor of the skin detection unit 411 .

For example, the skin detection unit 411 may receive the continuous images 300_1 - 300_N shown in FIG. 3 in advance for training. In this embodiment, each of the images 300_1 - 300_N may respectively include a human skin area 301 and an environment area 302 . It should be noted that since the non-contact physiological signal measurement device 100 may be operated in different usage environments (such as operating under conditions of different ambient light intensities or operating under conditions of different environmental images) or for sensing different quantities type of target (e.g. Such as different race categories or race skin color), so the skin detection unit 411 can firstly let the deep learning module of the skin detection unit 411 be input by inputting the sensing results corresponding to various usage environments and various measurement target types It can learn to recognize images corresponding to various usage environments and various measurement target types, and can effectively distinguish the human skin area 301 and the environment area 302 in the image, and further analyze the human skin area 301 in the image and calculations to generate a skin reflection signal. It is worth noting that the image signal sensing element 120 can be used to perform a continuous sensing operation for a period of time, or a plurality of continuous sensing operations continuously, so that the processor 110 can update in real time according to the image analysis results successively. Physiological information.

For another example, since the non-contact physiological signal measurement device 100 may be used in different situations or objects each time, the aforementioned image analysis results may correspond to different usage environments or different sensing object types, the skin detection unit 411's deep learning modules require different image processing algorithms or different learning experiences. Therefore, the non-contact physiological signal measurement device 100 of the present invention can also perform an adaptive learning function. The skin detection unit 411 can perform mutual verification based on the detection results of the continuous images 300_1~300_N and the detection results of another continuous image of the same measurement target in the next detection period, and can further feed back the verification results to the deep learning The module performs calibration, so that the skin detection unit 411 can provide accurate identification results for the human skin area 301 in the image, and also has an adaptive self-correction function, which can be used in various usage environments and various Measurement target type.

FIG. 6 is a schematic diagram of a non-contact physiological signal measurement system according to an embodiment of the present invention. Referring to FIG. 6, the non-contact physiological signal measurement system 600 of this embodiment For example, it can be constructed in a long-term care environment or a hospital medical environment to perform real-time, non-contact and automatic physiological signal measurement and monitoring operations for a large number of measurement targets. In this embodiment, the non-contact physiological signal measurement system 600 may include a plurality of measurement devices 620_1~620_N. Each of the measurement devices 620_1˜620_N can independently execute and realize the non-contact physiological signal measurement operation of the above-mentioned embodiments. In this embodiment, each of the measurement devices 620_1˜620_N may have related hardware and software configurations as in the above-mentioned non-contact physiological signal measurement device 100 of FIG. 1 . The monitoring host 610 is coupled to the measurement devices 620_1~620_N (wired or wireless communication methods can be used). The monitoring host 610 may include a processor, a storage device and a communication interface. The monitoring host 610 can be, for example, a cloud server device, and is configured to receive physiological information provided by each of the measuring devices 620_1 - 620_N.

In this embodiment, the measurement devices 620_1~620_N can be used to perform non-contact physiological signal measurement for multiple measurement objects, and to determine whether to output corresponding warning signals to the monitoring host 610, so that the monitoring host 610 can Monitor the warning signals corresponding to these physiological information. In this regard, when the measurement devices 620_1~620_N determine that at least one of the estimated heartbeat curve, estimated heartbeat variation curve, estimated respiration value, estimated blood oxygen value, and estimated blood pressure value of any of these physiological information is an abnormal value, the corresponding The measurement equipment can output a warning signal to the monitoring host 610 so that the monitoring host 610 can provide reminder information of the corresponding measurement target through a warning method such as video or sound, so that the monitor can obtain it through video or sound Real-time abnormal information of physiological signal measurement.

In addition, in other embodiments of the present invention, computing resources can also be concentrated In the monitoring host 610 , overall image processing, analysis and data calculation can be performed through the monitoring host 610 . In other words, the measuring devices 620_1~620_N may also only have image signal sensing elements, and the measuring devices 620_1~620_N respectively upload continuous images to the monitoring host 610. The processor of the monitoring host 610 can perform independent image processing, analysis, and data calculation on the continuous images provided by the measurement devices 620_1~620_N, so as to generate a plurality of corresponding physiological information.

In summary, the non-contact physiological signal measurement device, system and method of the present invention can obtain continuous continuous images through the image signal sensing element, and perform image processing and analysis on the continuous images to generate photoplethysmographic signals . The non-contact physiological signal measurement device of the present invention can analyze the photoplethysmography signal to obtain multiple physiological information of the measurement object.

Although the present invention has been disclosed above with the embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the technical field may make some changes and modifications without departing from the spirit and scope of the present invention. The scope of protection of the present invention should be defined by the scope of the appended patent application.

100: Non-contact physiological signal measurement equipment 110: Processor 120: image signal sensing element 130: storage device

Claims (24)

A non-contact physiological signal measurement device, comprising: a storage device for storing multiple modules; an image signal sensing element for obtaining a continuous image; and a processor coupled to the image signal sensing Components and the storage device, wherein the processor is used to execute the modules to detect a skin image from the continuous image, and extract a skin reflection signal from the skin image in the continuous image, and the processor A photoplethysmographic signal is calculated according to the skin reflection signal, wherein the processor calculates a physiological information according to the photoplethysmographic signal, and the physiological information includes a heartbeat estimation curve, a heartbeat variation estimation curve, and a breathing estimation value, an estimated value of blood oxygen, and an estimated value of blood pressure, wherein the processor executes a skin reflection signal acquisition module, and the skin reflection signal acquisition module includes: a skin detection unit for The skin image is detected from the continuous image through a deep learning module, wherein the skin detection unit is also based on the detection result of the continuous image and the detection of another continuous image of the same measurement target during the next detection period The results are mutually verified, and the verification results are fed back to the deep learning module to correct the deep learning module, wherein the processor executes a physiological signal analysis module, wherein the physiological signal analysis module includes: a heartbeat signal calculation unit for converting the photoplethysmographic signal into Band filtering is performed after the frequency domain to take out the frequency with the highest amplitude as the estimated heartbeat value, and the signal of the adjacent part of the estimated heartbeat value is converted back to the time domain to generate an estimated heartbeat curve, wherein the heartbeat signal calculation unit is based on the The heartbeat estimation curve calculates a heartbeat intensity variation curve, a heartbeat amplitude variation curve and a heartbeat frequency variation curve. The measurement device according to claim 1, wherein the skin reflection signal acquisition module further includes: a skin reflection signal acquisition unit, configured to extract the skin reflection signal from the continuous images. The measurement device as described in claim 2, wherein the processor executes a skin reflection signal calculation module, and the skin reflection signal calculation module includes: a signal sampling unit with equal intervals, which is used to change from time intervals to non-equal intervals An equidistant reflection signal with equal time intervals is converted from the skin reflection signal; and a signal conversion unit is used to recombine the equidistant reflection signal according to different wavelength signals to obtain the photoplethysmographic signal with pulse information. The measuring device as described in claim 3, wherein the signal conversion unit is a photoplethysmography unit (Photoplethysmography, PPG), a remote photoplethysmography unit (Remote-Photoplethysmography, rPPG) or an image photoplethysmography unit Tracing unit (Image-Photoplethysmography, iPPG). The measurement device as described in claim 1, wherein the physiological signal analysis module further includes: A breathing signal calculation unit, used to calculate the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve in unit time according to the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve The respective peak numbers for the three respiration estimates. The measuring device as described in claim 1, wherein the processor executes the physiological signal analysis module, wherein the physiological signal analysis module further includes: a blood oxygen signal calculation unit for extracting from the photoplethysmography signal A first reference signal and a second reference signal with different wavelengths are taken, and the blood oxygen signal calculation unit multiplies the ratio of the two signal values of the first reference signal and the second reference signal by a preset coefficient A difference value from a maximum blood oxygen concentration is obtained, wherein the blood oxygen signal calculation unit subtracts the difference value from the maximum blood oxygen concentration to generate the estimated blood oxygen value. The measuring device as described in claim 1, wherein the processor executes the physiological signal analysis module, wherein the physiological signal analysis module further includes: a blood pressure signal calculation unit, which is used for quadratic differentiation of the estimated heartbeat curve , to generate a third reference signal, and the processor obtains the signal characteristics of every adjacent two valley intervals in the third reference signal, and converts the signal characteristics into a blood pressure signal through linear regression analysis, wherein the processing The device generates the estimated blood pressure value according to the blood pressure signal. The measuring device according to claim 1, wherein the physiological information includes at least one of the estimated heartbeat curve, the estimated heartbeat variation curve, the estimated respiration value, the estimated blood oxygen value, and the estimated blood pressure value Portfolio information. A non-contact physiological signal measurement system, comprising: a monitoring host; and a plurality of measuring devices coupled to the monitoring host, wherein the measuring devices are used to perform non-contact physiological signal measurements on a plurality of measurement objects and used to determine whether to output a corresponding warning signal to the monitoring host, so that the monitoring host monitors the physiological information and the corresponding warning signal, wherein each of the measuring devices is used to detect the corresponding Measuring a respective skin image in a continuous image of the object, extracting a skin reflection signal from the skin image, and calculating a photoplethysmography signal from the skin reflection signal, wherein the measuring devices Each of the photoplethysmographic signals is calculated to correspond to a piece of physiological information, wherein the physiological information includes at least an estimated heartbeat curve, an estimated heartbeat variation curve, an estimated respiration value, an estimated blood oxygen value, and an estimated blood pressure value. One of them, wherein each of the measuring devices includes a skin reflection signal acquisition module, and the skin reflection signal acquisition module includes: a skin detection unit, which is used to learn from the continuous The skin image is detected by the image, wherein the skin detection unit also performs mutual verification according to the detection result of the continuous image and the detection result of another continuous image of the same measurement target during the next detection period, and the verification result Feedback to the deep learning module to correct the deep learning module, wherein each of the measuring devices includes a physiological signal analysis module, wherein The physiological signal analysis module includes: a heartbeat signal calculation unit, which is used to convert the photoplethysmography signal into the frequency domain and perform band filtering operation to extract the frequency with the highest amplitude as the estimated heartbeat value, and calculate the estimated heartbeat The signals of the adjacent parts are converted back to the time domain to generate a heartbeat estimation curve, wherein the heartbeat signal calculation unit calculates a heartbeat intensity variation curve, a heartbeat amplitude variation curve and a heartbeat frequency variation curve according to the heartbeat estimation curve. The non-contact physiological signal measurement system as described in Claim 9, wherein the skin reflection signal acquisition module further includes: a skin reflection signal acquisition unit, configured to extract the skin reflection signal from the continuous images. The non-contact physiological signal measurement system as described in claim 10, wherein each of the measurement devices includes a skin reflection signal calculation module, and the skin reflection signal calculation module includes: a signal sampling unit at equal intervals, It is used to convert the skin reflection signals with equal time intervals from the skin reflection signals with non-equal intervals; and a signal conversion unit is used to recombine the equally spaced reflection signals according to different wavelength signals. The photoplethysmographic signal of pulse information. The non-contact physiological signal measurement system according to claim 11, wherein the signal conversion unit is a photoplethysmography unit, a remote photoplethysmography unit or an image photoplethysmography unit. The non-contact physiological signal measurement system as described in Claim 9, wherein the physiological signal analysis module further includes: a respiratory signal calculation unit, which is used to calculate the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency The variation curves are used to calculate the respective peak numbers of the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve in unit time, so as to generate three kinds of respiration estimation values. The non-contact physiological signal measurement system as described in Claim 9, wherein the physiological signal analysis module further includes: a blood oxygen signal calculation unit, which is used to extract a first signal with different wavelengths from the photoplethysmography signal A reference signal and a second reference signal, and the blood oxygen signal calculation unit multiplies the ratio of the two signal values of the first reference signal and the second reference signal by a preset coefficient to obtain a maximum blood oxygen concentration A difference value, wherein the blood oxygen signal calculation unit subtracts the difference value from the maximum blood oxygen concentration to generate the blood oxygen estimated value. The non-contact physiological signal measurement system as described in Claim 9, wherein the physiological signal analysis module further includes: a blood pressure signal calculation unit, which is used to perform quadratic differentiation on the estimated heartbeat curve to generate a third reference signal, and the processor obtains the signal characteristics of each adjacent two valley intervals in the third reference signal, and converts the signal characteristics into a blood pressure signal through linear regression analysis, wherein the processor generates the blood pressure signal according to the blood pressure signal Estimated blood pressure. The non-contact physiological signal measurement system according to claim 9, wherein the physiological information includes at least one of the estimated heartbeat curve, the estimated heartbeat variation curve, the estimated respiration value, the estimated blood oxygen value, and the estimated blood pressure value A combined information of the two. A non-contact physiological signal measurement method, comprising: obtaining a continuous image through an image signal sensing element; detecting a skin image from the continuous image through a processor, and extracting the skin image from the continuous image taking out a skin reflection signal; calculating a photoplethysmography signal through the processor according to the skin reflection signal; and calculating a physiological information according to the photoplethysmography signal through the processor, wherein the physiological information includes a heartbeat estimate At least one of a curve, an estimated heartbeat variation curve, an estimated respiration value, an estimated blood oxygen value, and an estimated blood pressure value, wherein the step of generating the skin reflection signal includes: executing a deep learning module through the processor, The skin image is detected from the continuous image; and the detection result of the continuous image is mutually verified with the detection result of another continuous image of the same measurement target during the next detection period, and the verification result is fed back to The deep learning module is used to calibrate the deep learning module, wherein the step of calculating the physiological information according to the photoplethysmography signal includes: converting the photoplethysmography signal to a frequency domain through the processor and performing Band filtering operation; the frequency with the highest amplitude is taken out as the heartbeat estimate by the processor, and the signal of the adjacent part of the heartbeat estimate is converted back to the time domain to generate a heartbeat estimate curve; and the heartbeat estimate is estimated by the processor The curve calculates a heartbeat intensity variation curve, a heartbeat amplitude variation curve and a heartbeat frequency variation curve. The non-contact physiological signal measurement method as described in claim 17, wherein the step of generating the skin reflection signal further includes: extracting the skin reflection signal from the continuous image through the processor. The non-contact physiological signal measurement method according to claim 18, wherein the step of calculating the photoplethysmography signal according to the skin reflection signal comprises: converting from the skin reflection signal whose time interval is not equidistant An equidistant reflection signal with equidistant time intervals; and the photoplethysmographic signal with pulse information is recombined from the equidistant reflection signal according to different wavelength signals. The non-contact physiological signal measurement method according to claim 17, wherein the step of calculating the physiological information according to the photoplethysmographic signal further includes: through the processor according to the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve to calculate the difference between the heartbeat intensity variation curve, the heartbeat amplitude variation curve and the heartbeat frequency variation curve in unit time number of peaks to produce three respiration estimates. The non-contact physiological signal measurement method according to claim 17, wherein the step of calculating the physiological information according to the photoplethysmographic signal further includes: extracting signals with different wavelengths from the photoplethysmographic signal through the processor A first reference signal and a second reference signal; through the processor, the ratio of the two signal values of the first reference signal and the second reference signal is multiplied by a preset coefficient to obtain a maximum blood oxygen concentration a difference value; and subtracting the difference value from the maximum blood oxygen concentration by the processor to generate the estimated blood oxygen value. The non-contact physiological signal measurement method according to claim 17, wherein the step of calculating the physiological information according to the photoplethysmographic signal further includes: secondly differentiating the estimated heartbeat curve through the processor to generate A third reference signal; through the processor, the signal characteristics of each adjacent two trough intervals in the third reference signal are obtained, and the signal characteristics are converted into a blood pressure signal through linear regression analysis; and through the processor according to The blood pressure signal generates the blood pressure estimate. The non-contact physiological signal measurement method as described in Claim 17, wherein the physiological information includes the estimated heartbeat curve, the estimated heartbeat variation curve, Combination information of at least two of the estimated respiration value, the estimated blood oxygen value, and the estimated blood pressure value. The non-contact physiological signal measurement method as described in claim 17, including: determining whether to output a corresponding warning signal to a monitoring host according to the physiological information; and monitoring the physiological information and the corresponding warning through the monitoring host signal.

Images