Control method and device of infrared light supplement lamp

Abstract

The invention relates to a control method for an infrared light supplement lamp, and the method comprises the following steps: S1, carrying out the image segmentation of an infrared image, and extracting a human face skin region; and S2, calculating a brightness mean value of a human face skin area in the infrared image, and adjusting a duty ratio of a PWM control signal output to a control circuit according to whether the brightness mean value of the human face skin area is within a threshold range so as to control the brightness of an infrared light supplement lamp. According to the control device and method for the infrared light supplement lamp, feedback adjustment of the brightness of the light emitted to the skin area of the human face by the infrared light supplement lamp is achieved, and the control device and method have the advantages of being high in adaptability and automation degree.

Description

Technical Field

[0001] The present invention relates to the field of lighting lamps, and in particular to a control method and device for an infrared fill light. Background Art

[0002] A common method for detecting heart rate is image photoplethysmography (IPPG). IPPG relies on a camera capturing video images of the skin surface and analyzing pixel intensity variations within the video image to extract the pulse signal, enabling heart rate detection. During detection, the camera must effectively capture subtle pixel variations across the skin.

[0003] Due to insufficient lighting conditions, existing technologies use supplemental lighting to illuminate the human body, enabling the camera to capture clear images and improving the signal-to-noise ratio. Furthermore, to minimize disruption to sleep, invisible infrared lighting and cameras are often used to capture images. However, existing technologies for detecting pulses at night still suffer from significant measurement errors. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to overcome the defects or shortcomings of the prior art and design a control method and device for an infrared fill light.

[0005] A method for controlling an infrared fill light comprises the following steps:

[0006] S1: Segment the infrared image and extract the facial skin area;

[0007] S2: Calculate the average brightness of the facial skin area in the infrared image, and adjust the duty cycle of the PWM control signal output to a control circuit according to whether the average brightness of the facial skin area is within a threshold range to control the brightness of the infrared fill light.

[0008] The control method of the infrared fill light of the present invention can adjust the output PWM signal according to the brightness of the facial skin area of the human face in the infrared image, and has the advantage of strong adaptability.

[0009] Furthermore, the step S1 includes the following steps:

[0010] S11: Perform face detection on the infrared image to obtain a face recognition frame;

[0011] S12: Extract the facial skin area of the face in the face recognition frame.

[0012] Through face detection, the position of the face in the infrared image is initially obtained, and then the facial skin area in the infrared image is extracted from the face recognition frame to improve the accuracy of subsequent detection of the brightness of the face area.

[0013] Furthermore, step S2 includes the following steps:

[0014] S21A: Determine whether the average brightness of the facial skin area of the human face is within a threshold range. If so, execute step S1 repeatedly; if not, execute step S21B.

[0015] S21B: Determine whether the average brightness of the facial skin area of the face is lower than the threshold range. If so, execute step S22A; if not, execute step S21C;

[0016] S21C: Determine whether the average brightness of the facial skin area of the face is higher than the threshold range. If so, execute step S22B; if not, loop through step S1;

[0017] S22A: Increase the duty cycle of the output PWM control signal;

[0018] S22B: Reduce the duty cycle of the output PWM control signal.

[0019] By changing the duty cycle of the output PWM control signal, the average brightness of the facial skin area is ensured to be within the threshold range.

[0020] Furthermore, in step S11, face detection is performed using a nested network algorithm to obtain a face recognition frame; and in step S12, a K-means clustering algorithm is used to obtain a facial skin area within the face recognition frame.

[0021] The PIP-Net algorithm is a highly efficient facial landmark detection network that combines the advantages of coordinate regression and heatmap regression. By designing a nearest neighbor regression module and employing a semi-supervised learning strategy, it offers strong robustness and good cross-domain generalization, making it particularly advantageous in complex scenarios with diverse human postures. Even in the dimly lit nighttime environment, the PIP-Net network can still effectively locate faces and generate facial recognition frames.

[0022] The K-means algorithm iteratively assigns the brightness values of pixels within the facial recognition frame to K clusters. Each pixel is assigned to the cluster corresponding to the cluster center closest to it, thereby clustering pixels with similar brightness values. K-means clustering can distinguish different areas of the face by brightness, avoiding the uneven brightness of the skin when infrared light is projected onto the face due to the uneven shape of the face, thereby identifying skin areas with a better signal-to-noise ratio.

[0023] Furthermore, the threshold range of the brightness mean of the facial skin area of the human face is 70-90.

[0024] The facial skin area was illuminated using infrared fill light, with the brightness controlled between 10% and 90% in 20% increments. The heart rate measured by IPPG served as the experimental group, while the heart rate measured by a PPG heart rate monitoring device served as the control group. It was found that IPPG heart rate detection achieved the highest accuracy when the average brightness of the facial skin area was around 80. Therefore, the brightness threshold range for the facial skin area was set to 70-90 for optimal detection.

[0025] A control device for an infrared fill light, comprising

[0026] Infrared fill light, used to emit infrared light towards the user's face;

[0027] an infrared camera for capturing infrared images of the user's face;

[0028] An image processor, configured to receive the infrared image captured by the infrared camera and execute the above-mentioned method for controlling the infrared fill light;

[0029] The control circuit is used to supply power to the infrared fill light and receive the PWM control signal generated by the image processor to control the brightness of the infrared fill light.

[0030] Furthermore, the control circuit includes a DC power supply DC1 and a transistor Q1. The DC power supply DC1 is connected in series with the collector and emitter of the transistor Q1, and the input and output of the infrared fill light 40 (H1). The image processor is connected to the base of the transistor Q1 and sends a PWM control signal to the transistor Q1. The PWM signal controls the on / off state of the transistor Q1, thereby controlling the amount of power received by the infrared fill light per second and thus controlling the brightness of the infrared fill light.

[0031] Furthermore, the control circuit also includes a linear regulator U1, the input and output ends of the linear regulator U1 are connected in series with the DC power supply DC1, and its ground end is connected to the negative electrode of the DC power supply. The linear regulator U1 stabilizes the current output by the DC power supply DC1 and reduces the input 15V voltage to 5V.

[0032] Furthermore, the control circuit includes a first filter capacitor C1 and a second filter capacitor C2; the first filter capacitor C1 is connected in parallel with the DC power supply DC1, and the second filter capacitor C2 is connected between the ground terminal and the output terminal of the linear regulator U1. The first filter capacitor C1 is used to stabilize the voltage output by the DC power supply DC1, and the second filter capacitor C2 is used to stabilize the voltage output by the linear regulator U1.

[0033] A computer-readable storage medium stores a computer program, wherein the computer program implements the above-mentioned infrared fill light control method when executed by a processor.

[0034] A computer device comprising:

[0035] at least one memory and at least one processor;

[0036] The memory is used to store one or more programs;

[0037] When the one or more programs are executed by the at least one processor, the at least one processor implements the steps of the above-mentioned method for controlling an infrared fill light.

[0038] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a schematic structural diagram of a control device for an infrared fill light according to the present invention;

[0040] Figure 2 This is a flow chart of the control method of the infrared fill light of the present invention;

[0041] Figure 3 FIG. 4 is a circuit diagram of a control circuit of the present invention. DETAILED DESCRIPTION

[0042] The present invention analyzes the reasons why the existing technology produces large errors when monitoring human heart rate at night. It is found that when monitoring human heart rate through IPPG, since this detection technology relies on a camera to capture a video image of the human skin surface; when the pixel brightness in the video image is too low, the signal-to-noise ratio of the video image is too small, making it difficult to accurately identify the heart rate signal; however, when the pixel brightness in the video image is too high, some detailed pixel information in the video image will be lost due to overexposure. Therefore, if the pixel brightness in the video image cannot be guaranteed to be within a certain range, the accuracy of heart rate recognition will deviate significantly.

[0043] Further analysis found that in the application scenario of monitoring human heart rate at night; since users often have unconscious movements when sleeping, this will cause the straight-line distance between the human body and the infrared fill light and infrared camera to change. At this time, the light emitted by the infrared fill light will rapidly attenuate or increase with the change of distance, causing the light intensity of the infrared fill light on the human body surface to fluctuate violently. The image collected by the infrared camera is prone to problems of pixel brightness being too high or too low, resulting in large errors in the results of monitoring human heart rate at night.

[0044] Based on the above technical development process, the control method and device of the infrared fill light of the present invention are finally obtained. The scheme of the present invention is described in detail below with reference to the embodiments.

[0045] See also Figure 1 , which is a schematic structural diagram of a control device for an infrared fill light of the present invention, the control device for the infrared fill light includes an infrared camera 10, an image processor 20, a control circuit 30, and a fill light 40. The infrared fill light 40 emits infrared light toward the user's face, the infrared camera 10 captures an infrared image of the user's face, the image processor 20 receives the infrared image captured by the infrared camera, analyzes the infrared image, and generates a PWM control signal. The control circuit 30 supplies power to the infrared fill light 40 and receives the PWM control signal generated by the image processor to control the power supply intensity to the infrared fill light 40, thereby controlling the brightness of the infrared fill light 10.

[0046] For details, please refer to Figure 2 , which is a flow chart of the control method of the infrared fill light.

[0047] When it is necessary to detect the user's heart rate at night, the infrared fill light 40 emits infrared light to the user's face to illuminate the face, and the infrared camera 10 collects the infrared light reflected back from the face in real time and generates an infrared image.

[0048] The image processor 20 is used to execute the control method of the infrared fill light, including steps S1 and S2.

[0049] S1: Perform image segmentation on the infrared image and extract the facial skin area.

[0050] Specifically, step S1 includes the following steps:

[0051] S11: Perform face detection on the infrared image to obtain a face recognition frame.

[0052] Face detection is performed using a nested network (PIP-Net) algorithm to obtain face recognition frames. PIP-Net is an efficient facial keypoint detection network that combines the advantages of coordinate regression and heatmap regression. By designing a nearest neighbor regression module and adopting a semi-supervised learning strategy, it has strong robustness and good cross-domain generalization performance, showing greater application advantages in complex scenes with different human postures. Even in detection scenarios with poor lighting conditions at night, the PIP-Net network can still effectively locate the face position and obtain the face recognition frame.

[0053] S12: Extract the facial skin area of the face in the face recognition frame.

[0054] The K-means clustering algorithm divides the facial recognition frame into regions of varying brightness. The algorithm iteratively assigns the brightness values of pixels within the facial recognition frame to K clusters, assigning each pixel to the cluster corresponding to the cluster center closest to it, thereby clustering pixels with similar brightness values. K-means clustering can thus distinguish different facial regions by brightness, avoiding the uneven brightness of the skin caused by infrared light projected onto the surface due to the uneven shape of the face, thereby identifying skin regions with a better signal-to-noise ratio.

[0055] S2: Calculate the average brightness of the facial skin area in the infrared image, and adjust the duty cycle of the PWM control signal output to a control circuit according to whether the average brightness of the facial skin area is within a threshold range to control the brightness of the infrared fill light.

[0056] Specifically, step S2 includes the following steps:

[0057] S21A: Determine whether the brightness mean of the facial skin area of the face is within a threshold range. If so, loop through step S1; if not, execute step S21B.

[0058] S21B: Determine whether the average brightness of the facial skin area of the face is lower than the threshold range. If so, execute step S22A; if not, execute step S21C.

[0059] S21C: Determine whether the average brightness of the facial skin area of the face is higher than the threshold range. If so, execute step S22B; if not, loop through step S1.

[0060] S22A: Increase the duty cycle of the output PWM control signal by 20%.

[0061] S22B: Reduce the duty cycle of the output PWM control signal by 20%.

[0062] Preferably, in order to ensure that a higher infrared image signal-to-noise ratio can be obtained when the brightness mean of the facial skin area of the human face is within the threshold range, the threshold range is calibrated in this embodiment. Specifically, the facial skin area of the human face is illuminated by an infrared fill light 40, and the infrared fill light 40 is controlled between 10% and 90%, with a step size of 20%. The heart rate obtained by IPPG detection is used as the experimental group, and the heart rate obtained by the PPG heart rate detection device is used as the reference control group. The mean absolute error (MAE) and root mean square error (RMSE) are used as the main performance evaluation indicators; MAE is used to measure the average absolute deviation between the predicted value and the true value, which can reflect the overall prediction accuracy; RMSE is more sensitive to larger errors and can highlight the performance of the algorithm in extreme cases. These indicators can fully reflect the accuracy and stability of the proposed method in heart rate measurement. The obtained heart rate results are shown in Table 1.

[0063] Table 1 Relationship between infrared fill light brightness ratio and heart rate results

[0064] Brightness ratio 10％ 30％ 50％ 70％ 90％ MAE / bpm 12.75 5.32 10.58 18.56 29.58 RMSE / bpm 12.89 5.49 11.24 19.02 29.62

[0065] As can be seen from the table above, the heart rate error obtained by IPPG is highest when the infrared fill light brightness is 30%, with a MAE of only 5.32 and an RMSE of only 5.49. This means that when the infrared fill light brightness is 30%, the accuracy of human heart rate detection by IPPG is the highest. To further obtain the brightness range of the facial skin area with the highest accuracy when IPPG detects human heart rate, the skin brightness values of the facial skin area are statistically analyzed for videos with the infrared fill light brightness of 30%. The final average brightness of the facial skin area is around 80, so the brightness threshold range of the facial skin area is set to 70-90 to obtain the best detection effect.

[0066] See also Figure 3, which is a circuit diagram of the control circuit 30, which includes a DC power supply DC1, a linear regulator U1, a first filter capacitor C1, a second filter capacitor C2, and a transistor Q1. The DC power supply DC1 is connected in series with the collector and emitter of the transistor Q1, the input and output of the infrared fill light 40 (H1), and the input and output of the linear regulator U1. The ground terminal of the linear voltage regulator U1 is connected to the negative electrode (ground terminal) of the DC power supply, and the linear voltage regulator U1 stabilizes the current output by the DC power supply DC1 and reduces the input 15V voltage to 5V; the first filter capacitor C1 is connected in parallel with the DC power supply DC1 to stabilize the voltage output by the DC power supply DC1, and the second filter capacitor C2 is connected between the ground terminal and the output terminal of the linear voltage regulator U1 to stabilize the voltage output by the linear voltage regulator U1; the image processor 20 (H2) is connected to the base of the transistor Q1 to send a PWM control signal to the transistor Q1 to control the shutdown of the collector and emitter of the transistor, thereby controlling the power received by the infrared fill light 40 (H1) per second, thereby controlling the brightness of the infrared fill light 40 (H1).

[0067] The present invention discloses an infrared fill light control device and method. An image processor segments an infrared image captured by an infrared camera to obtain a facial skin region. The image processor then determines whether the average brightness of the facial skin region is within a threshold range. The device then generates a PWM control signal, which is transmitted to a control circuit. The control circuit uses the PWM control signal to control the power output to the infrared fill light, thereby controlling the fill light's brightness. The present invention implements feedback regulation of the brightness of light emitted by the infrared fill light to the facial skin region, offering the advantages of strong adaptability and a high degree of automation.

[0068] Based on the same inventive concept, the present application also provides an electronic device, which can be a terminal device such as a server, a desktop computing device, or a mobile computing device (e.g., a laptop computing device, a handheld computing device, a tablet computer, a netbook, etc.). The device includes one or more processors and a memory, wherein the processor is used to execute a program to implement the control method of the infrared fill light in the embodiment of the present invention; and the memory is used to store a computer program executable by the processor.

[0069] Based on the same inventive concept, the present application also provides a computer-readable storage medium, corresponding to an embodiment of the aforementioned method for controlling an infrared fill light, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for controlling an infrared fill light recorded in any of the aforementioned embodiments.

[0070] The present application may take the form of a computer program product implemented on one or more storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing program code. Computer-usable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. The information can be computer-readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0071] The above-described embodiments merely represent several implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous modifications and improvements without departing from the spirit of the present invention, and the present invention is intended to encompass such modifications and variations.

Claims

1. A method for controlling an infrared fill light, characterized in that: The following steps are involved: S1: Segment the infrared image and extract the facial skin area; S2: Calculate the average brightness of the facial skin area in the infrared image, and adjust the duty cycle of the PWM control signal output to a control circuit according to whether the average brightness of the facial skin area is within a threshold range to control the brightness of the infrared fill light.

2. The control method of the infrared fill light according to claim 1, characterized in that: The step S1 comprises the following steps: S11: Perform face detection on the infrared image to obtain a face recognition frame; S12: Extract the facial skin area of the face in the face recognition frame.

3. The control method of the infrared fill light according to claim 1, characterized in that: The step S2 comprises the following steps: S21A: Determine whether the average brightness of the facial skin area of the human face is within a threshold range. If so, execute step S1 repeatedly; if not, execute step S21B. S21B: Determine whether the average brightness of the facial skin area of the face is lower than the threshold range. If so, execute step S22A; if not, execute step S21C; S21C: Determine whether the average brightness of the facial skin area of the face is higher than the threshold range. If so, execute step S22B; if not, loop through step S1; S22A: Increase the duty cycle of the output PWM control signal; S22B: Reduce the duty cycle of the output PWM control signal.

4. The control method of the infrared fill light according to claim 3, characterized in that: The threshold range of the brightness mean of the facial skin area of the human face is 70-90.

5. A control device for an infrared fill light, characterized in that: include Infrared fill light, used to emit infrared light towards the user's face; an infrared camera for capturing infrared images of the user's face; An image processor, configured to receive the infrared image captured by the infrared camera and execute the control method of the infrared fill light according to claims 1-4; The control circuit is used to supply power to the infrared fill light and receive the PWM control signal generated by the image processor to control the brightness of the infrared fill light.

6. The control device for the infrared fill light according to claim 5, characterized in that: The control circuit includes a DC power supply DC1 and a transistor Q1. The DC power supply DC1 is connected in series with the collector and emitter of the transistor Q1 and the input and output ends of the infrared fill light. The image processor is connected to the base of the transistor Q1 and sends a PWM control signal to the transistor Q1.

7. The control device for the infrared fill light according to claim 6, characterized in that: The control circuit also includes a linear regulator U1, the input and output ends of the linear regulator U1 are connected in series with the DC power supply DC1, and its ground end is connected to the negative electrode of the DC power supply. The linear regulator U1 stabilizes the current output by the DC power supply DC1 and reduces the input voltage.

8. The control device for the infrared fill light according to claim 7, characterized in that: The control circuit further includes a first filter capacitor C1 and a second filter capacitor C2; the first filter capacitor C1 is connected in parallel with the DC power supply DC1, and the second filter capacitor C2 is connected between the ground terminal and the output terminal of the linear regulator U1.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the infrared fill light control method according to claims 1 to 4 are implemented.

10. A computer device, characterized in that: include: at least one memory and at least one processor; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the at least one processor implements the steps of the method for controlling an infrared fill light according to any one of claims 1 to 4.

Images