Monitoring system for vehicle
By combining a light source, a light source control unit, and an AI recognition module in the vehicle monitoring system, the distance and degree of overlap between glare and the pupil are identified, solving the glare problem when the driver is wearing glasses and achieving high-precision vision monitoring.
Patent Information
- Application Number
- CN202511857397.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2025-06-20
- Filing Date
- 2025-12-10
- Publication Date
- 2026-02-06
AI Technical Summary
Existing vehicle monitoring systems are prone to glare when drivers wear glasses, making it difficult to accurately identify the eyes. This glare is even more severe when using under-display lens technology, affecting the accuracy of vision monitoring.
By combining multiple light sources, a light source control unit, an AI recognition module, and a camera module, the system can accurately determine whether to switch lighting sources by identifying the distance and degree of overlap between glare and the pupil, thereby improving the accuracy of vision monitoring.
It significantly improves the accuracy of driver vision monitoring, ensuring that the pupil position can still be accurately identified even in the presence of glare, and reducing the interference of glare on the monitoring system.
Smart Images
Figure CN121486677A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a monitoring system, and more particularly to a vehicle monitoring system. Background Technology
[0002] In driver monitoring systems (DMS), illumination is required to ensure the system can accurately monitor and identify the driver's facial features. However, if the driver wears glasses, glare may occur on the lenses, preventing the monitoring system from correctly identifying the driver's eyes. If the monitoring system uses under-display camera (UDC) technology, this glare phenomenon will also appear as a diffraction pattern, expanding its impact area. Summary of the Invention
[0003] This invention provides a vehicle monitoring system with high accuracy.
[0004] According to one embodiment of the present invention, a vehicle monitoring system is provided, including multiple light sources, a light source control unit, an AI recognition module, and a camera module. The light sources are used to emit illumination light. The light source control unit is configured to switch the light sources. The AI recognition module is connected to the light source control unit. The camera module is connected to the AI recognition module and is used to capture images of a face illuminated by the illumination light to generate sensing data. The AI recognition module generates recognition data based on the sensing data, and the recognition data includes distance data. The distance data includes the minimum distance between the illumination light on the face and the pupil. The light source control unit switches the light sources based on the recognition data.
[0005] According to another embodiment of the present invention, a vehicle monitoring system is provided, including multiple light sources, a light source control unit, and a camera module. The light sources are used to emit illumination light. The light source control unit is configured to switch the light sources. The camera module is connected to the light source control unit and includes a camera and an image signal processor, wherein the camera is connected to the image signal processor, and the image signal processor includes an AI recognition module. The camera is used to capture images of a face illuminated by the illumination light, and the AI recognition module is used to generate recognition data. The recognition data includes distance data, which includes the minimum distance between the illumination light on the face and the pupil. The light source control unit switches the light sources according to the recognition data.
[0006] Based on the above, the vehicle monitoring system provided in this embodiment of the invention uses an AI recognition module to identify the distance between glare and the pupil, thereby accurately determining whether it is necessary to switch the lighting source again, which greatly improves the accuracy of vision monitoring.
[0007] To make the above features and advantages of the present invention more apparent and understandable, specific embodiments are described below, and detailed descriptions are provided in conjunction with the accompanying drawings. Attached Figure Description
[0008] Figure 1 A schematic diagram of a vehicle monitoring system according to a first embodiment of the present invention is shown.
[0009] Figure 2 A schematic diagram of a vehicle monitoring system according to a second embodiment of the present invention is shown.
[0010] Figure 3 A schematic diagram of a vehicle monitoring method according to some embodiments of the present invention is shown.
[0011] Figure 4 A schematic diagram of the monitoring status is shown according to some embodiments of the present invention.
[0012] Figure 5 A schematic diagram of the monitoring status is shown according to some embodiments of the present invention.
[0013] Explanation of reference numerals in the attached figures: 1: Vehicle monitoring system 2: Vehicle monitoring system 100: Camera Module 101: Sensor Data 110: Camera 120: Image Signal Processor 200: AI Recognition Module 201: Identification Data 300: Light source control unit 401: Light source 402: Light source D1, D2, D3: Minimum distance FC: Face GL: Glasses LL: Illuminating light SL, SL1, SL2, SL3, SL4, SL5: Glare 1000: Vehicle Monitoring Methods 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900: Steps Detailed Implementation
[0014] Reference Figure 1 The diagram illustrates a vehicle monitoring system according to a first embodiment of the present invention. The vehicle monitoring system 1 is adapted to monitor the driver's line of sight and includes light sources 401, 402, a light source control unit 300, an AI recognition module 200, and a camera module 100.
[0015] Light sources 401 and 402 are used to emit illumination light LL towards the driver's face FC; the illumination light LL can be, for example, infrared light. The light source control unit 300 is used to switch the light sources 401 and 402 on or off. The number of light sources 401 and 402 is not limited to... Figure 1 The two shown can be any number of times, and can be placed anywhere near the driver's seat of the vehicle.
[0016] The camera module 100 is connected to the AI recognition module 200, and the AI recognition module 200 is connected to the light source control unit 300. In some embodiments, the AI recognition module 200 may be implemented as multiple code snippets. These code snippets are stored in an edge computing device and executed by the edge computing device, but are not limited thereto.
[0017] The camera module 100 may include a camera 110 for capturing images of the driver's face FC illuminated by illumination light LL, thereby enabling the camera module 100 to generate sensing data 101. The AI recognition module 200 then generates recognition data 201 based on the sensing data 101.
[0018] However, this invention is not limited to Figure 1 The architecture shown is referenced. Figure 2 This diagram illustrates a vehicle monitoring system according to a second embodiment of the present invention. The vehicle monitoring system 2 of this second embodiment includes light sources 401 and 402, a light source control unit 300, and a camera module 100. The camera module 100 includes a connected camera 110 and an image signal processor (ISP) 120. The main difference between the vehicle monitoring system 2 and the vehicle monitoring system 1 is that the AI recognition module 200 is configured within the image signal processor 120. The vehicle monitoring system 2 is similar to the vehicle monitoring system 1 in that the light sources 401 and 402 emit illumination light LL towards the driver's face FC; the light source control unit 300 switches the light sources 401 and 402 on or off; the camera 110 captures the driver's face FC illuminated by the illumination light LL; and the AI recognition module 200 generates recognition data 201.
[0019] Simultaneously refer to Figures 1 to 3 ,in Figure 3 A schematic diagram of a vehicle monitoring method according to some embodiments of the present invention is shown. The vehicle monitoring method 1000 is adapted to monitor the driver's line of sight and includes: Step 1100: Turn on at least one of the multiple light sources 401, 402 to illuminate the driver's face FC; Step 1200: Use the camera module 100 and the AI recognition module 200 to search for the driver's pupils and generate a pupil search result; specifically, when the driver is wearing glasses GL, glare SL may be generated on the glasses GL, causing one or two pupils to be completely covered by the glare SL (e.g. Figure 4 (as shown); Therefore, in this step, the camera 110 of the camera module 100 is used to capture the driver's face FC, and the AI recognition module 200 is used to identify whether the camera 110 has captured the driver's two pupils, and the recognition result is stored as a pupil search result. Step 1300: If any pupil in the pupil search result is completely covered by glare SL (e.g.) Figure 4 (as shown), then the light source control unit 300 switches the light sources 401 and 402 until no pupil is completely covered by the glare SL. Step 1400: If, after switching light sources 401 and 402, the camera 110 does indeed capture the driver's two pupils (i.e., no pupil is completely obscured by glare, such as...), Figure 5 As shown), the AI recognition module 200 determines the glare grayscale of each of the glare SL1, SL2, SL3, SL4, and SL5 to form glare grayscale data. Based on the grayscale threshold built into the AI recognition module 200, it determines which of the glare SL1, SL2, SL3, SL4, and SL5 is valid glare. Figure 5 For example, if the gray levels of glare SL1, SL2, SL3, and SL4 are greater than or equal to the gray level threshold built into the AI recognition module 200, and the gray level of glare SL5 is less than the gray level threshold, then glare SL1, SL2, SL3, and SL4 are considered valid glare, and glare SL5 is considered invalid glare; in some embodiments, the gray level threshold may fall within the range of 230 to 255, but is not limited thereto. Step 1500: As Figure 5 As shown, distance data is generated for effective glare SL1, SL2, SL3, and SL4; this distance data includes the minimum distance D1 between glare SL1 and the pupil, the minimum distance D2 between glare SL2 and the pupil, and the minimum distance D3 between glare SL3 and the pupil; on the other hand, glare SL4 partially overlaps with the pupil. Step 1600: Perform overlap ratio identification for glare SL4 that overlaps with the pupil. Specifically, the camera 110 of the camera module 100 captures the driver's face FC, and the AI recognition module 200 identifies the area of the part of glare SL4 that overlaps with the pupil (i.e., the overlapping area) and the pupil area, and stores the ratio of the overlapping area to the pupil area as an overlap ratio. If the overlap ratio is greater than or equal to an overlap ratio threshold, it means that the glare SL4 will interfere with vision monitoring. Accordingly, step 1700 needs to be executed to switch the light sources 401 and 402 through the light source control unit 300 and return to step 1200. On the other hand, if the overlap ratio is less than the overlap ratio threshold, it means that the glare SL4 will not interfere with vision monitoring, so vision monitoring can be performed (step 1800). In some embodiments, the overlap ratio threshold may be, for example, 25%, but is not limited thereto.
[0020] It should be noted that in some embodiments, Figure 5 Glare SL4 is absent. That is, there is no glare overlapping the pupil, and the illumination light from light sources 401 and 402 will not interfere with line-of-sight monitoring. Therefore, line-of-sight monitoring can be performed directly after step 1500 (step 1900).
[0021] However, the vehicle monitoring method according to embodiments of the present invention is not limited to the steps described above. In some embodiments, the AI recognition module 200 can be used to first identify the pupil diameter and treat the pupil diameter as a minimum distance threshold. Even Figure 5 If glare SL4 is absent, and the minimum distance D3 of glare SL3 is less than the minimum distance threshold (step 1500), glare SL3 is also considered a valid interference, requiring step 1700 to switch light sources 401 and 402 via the light source control unit 300, and then returning to step 1200. It should be noted that the minimum distance threshold is not limited to being equal to the pupil diameter, but can be any multiple of the pupil diameter.
[0022] In summary, the vehicle monitoring system provided by this embodiment of the invention utilizes an AI recognition module to identify the distance between glare and the pupil, as well as the degree of overlap between glare and the pupil, thereby accurately determining whether it is necessary to switch the lighting source again, which greatly improves the accuracy of vision monitoring.
Claims
1.A vehicle monitoring system, comprising: a plurality of light sources configured to emit illumination light; a light source control unit configured to switch the plurality of light sources; an AI recognition module connected to the light source control unit; and a camera module connected to the AI recognition module and configured to capture a face illuminated by the illumination light to generate a sensing data, wherein the AI recognition module is configured to generate an identification data based on the sensing data, the identification data comprising a distance data, the distance data comprising a minimum distance between the illumination light and a pupil on the face, and the light source control unit is configured to switch the plurality of light sources based on the identification data. 2.A vehicle monitoring system, comprising: a plurality of light sources configured to emit illumination light; a light source control unit configured to switch the plurality of light sources; and a camera module connected to the light source control unit, the camera module comprising a camera and an image signal processor, wherein the camera is connected to the image signal processor and the image signal processor comprises an AI recognition module, wherein the camera is configured to capture a face illuminated by the illumination light and the AI recognition module is configured to generate an identification data, the identification data comprising a distance data, the distance data comprising a minimum distance between the illumination light and a pupil on the face, and the light source control unit is configured to switch the plurality of light sources based on the identification data. 3.The vehicle monitoring system of claim 1 or 2, wherein the identification data comprises a pupil diameter. 4.The vehicle monitoring system of claim 1 or 2, wherein the identification data further comprises an overlap ratio, the overlap ratio being a ratio of an overlap area of the illumination light on the pupil and a pupil area. 5.The vehicle monitoring system of claim 1 or 2, wherein the identification data further comprises a glare gray scale data, and the glare gray scale data comprises a gray scale threshold. 6.The vehicle monitoring system of claim 1 or 2, wherein the identification data further comprises a pupil search result. 7.The vehicle monitoring system of claim 1, wherein the AI recognition module is configured in an edge computing device.