Fatigue driving prompting device and system based on face detection

Abstract

The invention discloses a fatigue driving prompting device and system based on face detection, and relates to the technical field of intelligent traffic safety, and the fatigue driving prompting device comprises an image acquisition module, a face detection module, a fatigue determination module, a prompting alarm module and a data storage module. The device and the system thereof can monitor the fatigue state of a driver in real time, prompt and alarm in time once detecting that the driver has a fatigue sign, enhance the alertness of the driver, effectively prevent traffic accidents caused by fatigue driving, remarkably improve the driving safety, have a data storage function, and are convenient to use. The driving state and characteristic data of a driver can be recorded in real time, historical accident data are stored, precious data are provided for follow-up driving behavior analysis and accident prevention research, objective evidence support is provided for possible legal disputes, installation is convenient, too much space in a compartment is not occupied, and the system is simple, convenient and easy to adjust. And meanwhile, excessive requirements on the sitting posture of a driver are not made.

Description

Technical Field

[0001] The present invention relates to the field of intelligent traffic safety technology, and in particular to a fatigue driving prompting device based on face detection and a system thereof. Background Art

[0002] Fatigue driving refers to a state in which the driver's driving skills decline, reaction becomes slow, and judgment weakened due to physiological or psychological reasons after driving continuously for a long period of time, thereby increasing the risk of traffic accidents. Fatigue driving is a serious traffic safety hazard that requires joint efforts of drivers, automobile manufacturers and relevant departments to prevent and reduce its occurrence.

[0003] Traditional fatigue driving detection methods have problems such as low accuracy and slow response, and cannot effectively ensure the safety of drivers and passengers. With the development of image processing technology and artificial intelligence algorithms, fatigue driving monitoring systems based on face detection have gradually become a research hotspot. Therefore, there is an urgent need for a fatigue driving reminder device and system based on face detection, which aims to monitor the driver's facial state in real time, remind the driver to avoid fatigue driving in time, and improve road traffic safety. Summary of the invention

[0004] The purpose of the present invention is to provide a fatigue driving reminder device and system based on face detection to solve the above-mentioned deficiencies in the prior art.

[0005] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a fatigue driving prompt system based on face detection, comprising an image acquisition module, a face detection module, a fatigue determination module, a prompt warning module and a data storage module;

[0006] The image acquisition module images and processes the driver's head, extracts feature data and transmits them to the data storage module and the face detection module respectively;

[0007] The face detection module processes the feature data, summarizes the analysis report and transmits it to the fatigue determination module;

[0008] The data storage module receives the feature data from the image acquisition module in real time, forms the persisted data and transmits it to the fatigue determination module, and the data storage module stores the historical accident data imported from the outside and transmits it to the fatigue determination module;

[0009] The fatigue determination module receives the analysis report from the face detection module, the existing data and the historical accident data from the data storage module, and compares the analysis report with the existing data and the historical accident data, and generates a determination report which is transmitted to the prompt alarm module;

[0010] The prompt and warning module analyzes the judgment report and prompts the driver according to the judgment report, and collects driver feedback and warnings.

[0011] Further, the image acquisition module includes a face imaging unit, a standardization processing unit, a face positioning unit, a face alignment unit and a feature extraction unit;

[0012] The face imaging unit continuously and in real time images the driver's head to obtain imaging information;

[0013] The standardization processing unit performs standardization processing on the imaging information, including image brightness and contrast processing, size adjustment, cropping, and removal of irrelevant background information;

[0014] The face positioning unit locates the position and size of the face in the standardized imaging information;

[0015] The face alignment unit detects key points of the face and transforms the face to a standard template face for alignment;

[0016] The feature extraction unit uses a linear discriminant analysis method to extract feature data from the aligned face images. The feature data includes the driver's blinking frequency, eyeball activity, eye tremor, facial activity and head activity.

[0017] The output end of the face imaging unit is connected to the input end of the standardization processing unit, the output end of the standardization processing unit is connected to the input end of the face positioning unit, the output end of the face positioning unit is connected to the input end of the face alignment unit, and the output end of the face alignment unit is connected to the input end of the feature extraction unit.

[0018] Furthermore, the specific steps of the linear discriminant analysis method are as follows:

[0019] Calculate the within-category scatter matrix, the within-category scatter matrix S W Describes the degree of dispersion of samples within each category. For a binary classification problem, there are two categories, 0 and 1. The mathematical expression is:

[0020]

[0021] Among them, X0 and X1 represent the sample sets of two categories respectively, μ0 and μ1 represent the mean vectors of two categories respectively;

[0022] Calculate the inter-class scatter matrix, the inter-class scatter matrix S B Describes the degree of dispersion between the center points of different categories. For the binary classification problem, its mathematical expression is:

[0023] S B =(μ0-μ1)(μ0-μ1) T

[0024] μ0 and μ1 are the mean vectors of the two categories respectively;

[0025] Optimizing the objective function, the goal of the linear discriminant analysis method is to find a projection direction w so that the projected data is separated between categories and clustered within categories. This can be achieved by maximizing the ratio of the scatter matrix between categories to the scatter matrix within categories, that is, optimizing the following objective function:

[0026]

[0027] To find the optimal projection direction, we need to optimize the objective function J(w). The solution of J(w) has nothing to do with the length of w, but only with the direction of w. Let w T S W w=1, transforming the original problem into a constrained optimization problem;

[0028] Using the Lagrange multiplier method, the constrained optimization problem is transformed into an unconstrained optimization problem, and the expression of w is obtained by solving it:

[0029]

[0030] The original data is projected into a low-dimensional space using the optimal projection direction w, and the samples are classified in the low-dimensional space.

[0031] Further, the face detection module includes a blink frequency detection unit, an eyeball activity detection unit, an eye tremor detection unit, a facial activity detection unit, a head activity detection unit and a summarization unit;

[0032] The blink frequency detection unit receives the characteristic data and detects the blink frequency of the driver to obtain blink frequency data;

[0033] The eyeball activity detection unit receives the characteristic data and detects the rotation frequency and angle of the driver's eyeballs to obtain the eyeball activity data;

[0034] The eye tremor detection unit receives the characteristic data and detects the frequency and amplitude of the driver's eye tremor to obtain the eye tremor data;

[0035] The facial activity detection unit receives the feature data and detects the facial activity of the driver to obtain facial activity data;

[0036] The head activity detection unit receives the characteristic data and detects the frequency and amplitude of the driver's head activity to obtain the head activity data;

[0037] The summary unit summarizes the blink frequency data, eyeball activity data, eye tremor data, facial activity data and head activity frequency to generate an analysis report;

[0038] The input ends of the blink frequency detection unit, the eyeball activity detection unit, the eye tremor detection unit, the facial activity detection unit and the head activity detection unit are respectively connected to the output end of the feature extraction unit;

[0039] The output ends of the blink frequency detection unit, the eyeball activity detection unit, the eye tremor detection unit, the facial activity detection unit and the head activity detection unit are respectively connected to the input end of the summarizing unit.

[0040] Further, the data storage module includes a persistence data unit and an accident data unit;

[0041] The continuous data unit receives the feature data from the feature extraction unit in real time and forms continuous data, and its input end is connected to the output end of the feature extraction unit;

[0042] The accident data unit stores historical accident data imported from outside.

[0043] Further, the fatigue determination module receives the analysis report from the summary unit, the ongoing data from the ongoing data unit, and the historical accident data from the accident data unit, and compares the analysis report with the ongoing data and the historical accident data, respectively, to generate a determination report;

[0044] The input end of the fatigue determination module is connected to the output ends of the summary unit, the survival data unit and the accident data unit respectively.

[0045] Furthermore, the prompt alarm module includes a prompt unit, a feedback unit and an alarm unit;

[0046] The prompt unit receives the determination report from the fatigue determination module and gives a prompt according to the determination report content;

[0047] The feedback unit collects driver feedback within a preset time threshold;

[0048] The alarm unit issues an alarm based on the feedback information;

[0049] The input end of the prompt unit is connected to the output end of the fatigue determination module, the input end of the feedback unit is connected to the output end of the prompt unit, and the input end of the alarm unit is connected to the output end of the feedback unit.

[0050] A fatigue driving reminder device based on face detection, which is applicable to the fatigue driving reminder system based on face detection, comprises a base;

[0051] A warning light, which is fixedly mounted on one side of the top of the base;

[0052] The voice device is fixedly installed on the other side of the top of the base.

[0053] Furthermore, two clamping parts are provided on one side of the base, one side of the top is rotatably connected to a turntable, the top of the turntable is rotatably connected to support rod one, and the other end of support rod one is rotatably connected to support rod two, and cushion layers are fixedly installed on opposite sides of the two clamping parts.

[0054] Furthermore, it also includes an imaging device, which is rotatably connected to the other end of the supporting rod.

[0055] Compared with the prior art, the present invention provides a fatigue driving reminder device and system based on face detection:

[0056] 1. It can monitor the driver's fatigue status in real time, including key indicators such as blinking frequency, eyeball activity, eye tremor, facial activity and head activity. Once the driver shows signs of fatigue, it will give timely prompts and warnings to enhance the driver's alertness, so that the driver can be aware of potential dangers at the early stage of fatigue, and take proactive measures to make adjustments, effectively preventing traffic accidents caused by fatigue driving and significantly improving driving safety;

[0057] 2. It has data storage function, which can record the driver's driving status and characteristic data in real time, and store historical accident data, providing valuable information for subsequent driving behavior analysis and accident prevention research. At the same time, the data recording function also provides objective evidence support for possible legal disputes;

[0058] 3. The device is easy to install, does not take up too much space in the car, is simple and easy to adjust, can adapt to a variety of vehicle models, and does not make too many requirements on the driver's sitting posture. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0060] Figure 1 A first schematic diagram of the overall structure provided for Embodiment 1 and Embodiment 2 of the present invention;

[0061] Figure 2 A schematic diagram of the structure of an image acquisition module provided in Embodiment 1 of the present invention;

[0062] Figure 3 A first schematic diagram of a local structure provided in Embodiment 1 of the present invention;

[0063] Figure 4 A second schematic diagram of a local structure provided in Embodiment 1 of the present invention;

[0064] Figure 5 A third schematic diagram of a local structure provided in Embodiment 1 of the present invention;

[0065] Figure 6 A partial structural diagram of the second embodiment of the present invention;

[0066] Figure 7 A second schematic diagram of the overall structure provided for Embodiment 2 of the present invention;

[0067] Figure 8 This is a schematic diagram of the overall structure of the third embodiment of the present invention.

[0068] Description of reference numerals:

[0069] 1. Image acquisition module; 11. Face imaging unit; 12. Standardization processing unit; 13. Face positioning unit; 14. Face alignment unit; 15. Feature extraction unit; 2. Face detection module; 21. Blinking frequency detection unit; 22. Eye movement detection unit; 23. Eye tremor detection unit; 24. Facial movement detection unit; 25. Head movement detection unit; 26. Summarization unit; 3. Fatigue determination module; 4. Prompt alarm module; 41. Prompt unit; 42. Feedback unit; 43. Alarm unit; 5. Data storage module; 51. Continuous data unit; 52. Accident data unit; 6. Base; 61. Snap-in; 62. Turntable; 63. Support rod one; 64. Support rod two; 65. Cushion; 7. Warning light; 8. Voice device; 9. Imaging device. DETAILED DESCRIPTION

[0070] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0071] Embodiment 1:

[0072] See also Figure 1-Figure 5 , a fatigue driving prompt system based on face detection, comprising an image acquisition module 1, a face detection module 2, a fatigue determination module 3, a prompt warning module 4 and a data storage module 5;

[0073] The image acquisition module 1 images and processes the driver's head, extracts feature data and transmits it to the data storage module 5 and the face detection module 2 respectively;

[0074] The face detection module 2 processes the feature data, summarizes the analysis report and transmits it to the fatigue determination module 3;

[0075] The data storage module 5 receives the feature data from the image acquisition module 1 in real time, forms the persisted data and transmits it to the fatigue determination module 3, and the data storage module 5 stores the historical accident data imported from the outside and transmits it to the fatigue determination module 3;

[0076] The fatigue determination module 3 receives the analysis report from the face detection module 2, the existing data and the historical accident data from the data storage module 5, and compares the analysis report with the existing data and the historical accident data, and generates a determination report which is transmitted to the prompt alarm module 4;

[0077] The prompting and warning module 4 analyzes the determination report and prompts the driver according to the determination report, and collects driver feedback and warnings.

[0078] The image acquisition module 1 includes a face imaging unit 11, a standardization processing unit 12, a face positioning unit 13, a face alignment unit 14 and a feature extraction unit 15;

[0079] The face imaging unit 11 continuously and in real time images the driver's head to obtain imaging information;

[0080] The standardization processing unit 12 performs standardization processing on the imaging information, including image brightness and contrast processing, size adjustment, cropping, and removal of irrelevant background information;

[0081] The face locating unit 13 locates the position and size of the face in the standardized imaging information;

[0082] The face alignment unit 14 detects key points of the face and transforms the face to a standard template face for alignment;

[0083] The feature extraction unit 15 extracts feature data from the aligned face images using a linear discriminant analysis method, wherein the feature data includes the driver's blinking frequency, eyeball activity, eye tremor, facial activity, and head activity;

[0084] The output end of the face imaging unit 11 is connected to the input end of the standardization processing unit 12, the output end of the standardization processing unit 12 is connected to the input end of the face positioning unit 13, the output end of the face positioning unit 13 is connected to the input end of the face alignment unit 14, and the output end of the face alignment unit 14 is connected to the input end of the feature extraction unit 15.

[0085] The specific steps of the linear discriminant analysis method are as follows:

[0086] Calculate the within-category scatter matrix, the within-category scatter matrix S W Describes the degree of dispersion of samples within each category. For a binary classification problem, there are two categories, 0 and 1. The mathematical expression is:

[0087]

[0088] Among them, X0 and X1 represent the sample sets of two categories respectively, μ0 and μ1 represent the mean vectors of two categories respectively;

[0089] Calculate the inter-class scatter matrix, the inter-class scatter matrix S B Describes the degree of dispersion between the center points of different categories. For the binary classification problem, its mathematical expression is:

[0090] S B =(μ0-μ1)(μ0-μ1) T

[0091] μ0 and μ1 are the mean vectors of the two categories respectively;

[0092] Optimizing the objective function, the goal of the linear discriminant analysis method is to find a projection direction w so that the projected data is separated between categories and clustered within categories. This can be achieved by maximizing the ratio of the scatter matrix between categories to the scatter matrix within categories, that is, optimizing the following objective function:

[0093]

[0094] To find the optimal projection direction, we need to optimize the objective function J(w). The solution of J(w) has nothing to do with the length of w, but only with the direction of w. Let w T S W w=1, transforming the original problem into a constrained optimization problem;

[0095] Using the Lagrange multiplier method, the constrained optimization problem is transformed into an unconstrained optimization problem, and the expression of w is obtained by solving it:

[0096]

[0097] The original data is projected into a low-dimensional space using the optimal projection direction w, and the samples are classified in the low-dimensional space.

[0098] Specific implementation method: The face imaging unit 11 continuously obtains the image information of the driver's head by means of camera imaging, and transmits it to the standardization processing unit 12. The standardization processing unit 12 adjusts the brightness and contrast of the image, resizes and crops it, and removes irrelevant background information to facilitate subsequent further processing, and transmits the standardized image information to the face positioning unit 13. The face positioning unit 13 locates the position and size of the face in the standardized image information, and transmits the located and standardized image information to the face alignment unit 14. The face alignment unit 14 detects the key points of the face, transforms the face to a standard template face for alignment, and transmits the transformed image information to the feature extraction unit 15. The feature extraction unit 15 extracts feature data from the aligned face image using a linear discriminant analysis method;

[0099] When a driver is driving fatigued, his eyes will show some obvious different characteristics, which can be used as an important basis for judging whether he is in a fatigue state. These characteristics include increased blinking frequency, dull face, reduced eye movement, red eyes, eye tremors and involuntary drooping of the head, etc. Therefore, the feature data extracted by the feature extraction unit 15 includes the driver's blinking frequency, eye movement, eye tremors, facial movement and head movement.

[0100] The face detection module 2 includes a blink frequency detection unit 21, an eyeball activity detection unit 22, an eye tremor detection unit 23, a facial activity detection unit 24, a head activity detection unit 25 and a summarizing unit 26;

[0101] The blink frequency detection unit 21 receives the characteristic data and detects the blink frequency of the driver to obtain blink frequency data;

[0102] The eyeball activity detection unit 22 receives the characteristic data and detects the rotation frequency and angle of the driver's eyeballs to obtain the eyeball activity data;

[0103] The eye tremor detection unit 23 receives the characteristic data and detects the frequency and amplitude of the driver's eye tremor to obtain the eye tremor data;

[0104] The facial activity detection unit 24 receives the feature data and detects the facial activity of the driver to obtain facial activity data;

[0105] The head activity detection unit 25 receives the characteristic data and detects the frequency and amplitude of the driver's head activity to obtain the head activity data;

[0106] The summarizing unit 26 summarizes the blink frequency data, the eyeball activity data, the eye tremor data, the facial activity data and the head activity frequency to generate an analysis report;

[0107] Input ends of the blink frequency detection unit 21, the eyeball activity detection unit 22, the eye tremor detection unit 23, the facial activity detection unit 24 and the head activity detection unit 25 are respectively connected to the output end of the feature extraction unit 15;

[0108] The output ends of the blink frequency detection unit 21 , the eyeball activity detection unit 22 , the eye tremor detection unit 23 , the facial activity detection unit 24 and the head activity detection unit 25 are respectively connected to the input end of the summarizing unit 26 .

[0109] Specific implementation method: The blinking frequency detection unit 21, the eye movement detection unit 22, the eye twitching detection unit 23, the facial activity detection unit 24 and the head movement detection unit 25 respectively analyze the feature data received from the feature extraction unit 15, obtain the blinking frequency data, the eye movement data, the eye twitching data, the facial activity data and the head movement data and transmit them to the summary unit 26 respectively. The summary unit 26 generates an analysis report after summarizing and transmits it to the fatigue determination module 3.

[0110] The data storage module 5 includes a persistence data unit 51 and an accident data unit 52;

[0111] The persistence data unit 51 receives the feature data from the feature extraction unit 15 in real time and forms persistence data, and its input end is connected to the output end of the feature extraction unit 15;

[0112] The accident data unit 52 stores historical accident data imported from the outside.

[0113] The fatigue determination module 3 receives the analysis report from the summary unit 26, the ongoing data from the ongoing data unit 51, and the historical accident data from the accident data unit 52, and compares the analysis report with the ongoing data and the historical accident data, respectively, to generate a determination report;

[0114] The input end of the fatigue determination module 3 is connected to the output ends of the summary unit 26 , the survival data unit 51 and the accident data unit 52 respectively.

[0115] Specific implementation method: The persistence data unit 51 receives the feature data within a period of time from the feature extraction unit 15, generates persistence data, and is used to perform longitudinal comparison on the driver's state, and judge whether the driver is fatigued by the changes in the driver's head and facial features;

[0116] The accident data unit 52 imports historical accident data by wireless or wired input, and is used to compare the driver's state horizontally, and judge whether the driver is fatigued by comparing the driver's state with the driver's state before the accident;

[0117] The fatigue determination module 3 compares the analysis report from the summary unit 26 with the existing data and the historical accident data respectively, determines the similarity between the analysis report and the normal driver data in the existing data, determines the similarity between the analysis report and the driver data before the accident in the historical accident data, sets a warning threshold, and generates a determination report.

[0118] Embodiment 2:

[0119] See also Figure 1 , Figure 6-Figure 7 This embodiment provides a technical solution based on the first embodiment:

[0120] The prompt alarm module 4 includes a prompt unit 41, a feedback unit 42 and an alarm unit 43;

[0121] The prompting unit 41 receives the determination report from the fatigue determination module 3 and gives a prompt according to the determination report content;

[0122] The feedback unit 42 collects driver feedback within a preset time threshold;

[0123] The alarm unit 43 issues an alarm based on the feedback information;

[0124] The input end of the prompt unit 41 is connected to the output end of the fatigue determination module 3 , the input end of the feedback unit 42 is connected to the output end of the prompt unit 41 , and the input end of the alarm unit 43 is connected to the output end of the feedback unit 42 .

[0125] Specific implementation method: After the prompt unit 41 receives the judgment report from the fatigue judgment module 3, it uses this as a basis to remind the driver by voice reminder. Within a preset time threshold, the feedback unit 42 receives the driver's feedback and transmits the feedback information to the alarm unit 43. The alarm unit 43 determines whether to flash the light and sound a buzzer warning based on the feedback information.

[0126] Embodiment three:

[0127] See also Figure 8 , a fatigue driving prompting device based on face detection, comprising a base 6;

[0128] A warning light 7, which is fixedly mounted on one side of the top of the base 6;

[0129] The voice device 8 is fixedly mounted on the other side of the top of the base 6 .

[0130] Two clamping parts 61 are provided on one side of the base 6, and a turntable 62 is rotatably connected to one side of the top thereof, a support rod 1 63 is rotatably connected to the top of the turntable 62, and a support rod 2 64 is rotatably connected to the other end of the support rod 1 63, and a cushion layer 65 is fixedly installed on the opposite sides of the two clamping parts 61.

[0131] The imaging device 9 is rotatably connected to the other end of the second supporting rod 64 .

[0132] Specific implementation method: two clamping parts 61 are clamped to the guide plate of the air outlet of the automobile air conditioner. The cushion layer 65 can be deformed and has a high friction coefficient, which can ensure that the clamping part 61 does not shake after the clamping is completed. The base 6 has a built-in storage power supply. At the same time, the storage power supply can be connected to the power supply in the car through a transmission line and has a certain self-sustaining force. After the fixation is completed, the turntable 62 is manually rotated, and the support rod 1 63, the support rod 2 64 and the imaging device 9 are rotated at the same time to ensure that the imaging device 9 can face the driver as directly as possible and can completely include the driver's head in the imaging range. After rotating to a certain angle, the turntable 62, the support rod 1 63, the support rod 2 64 and the imaging device 9 can stabilize the angle after rotation without shaking. The base 6 has a built-in PLC controller. The imaging device 9 includes but is not limited to a camera, which can image in real time, and is electrically connected to the internal power supply. At the same time, it is also controlled by the internal PLC programming program;

[0133] The warning light 7 can flash to stimulate the driver's vision and warn him. The warning light 7 is electrically connected to the internal power supply and is also controlled by the internal PLC programming program;

[0134] The voice device 8 can issue inquiries to the driver, receive feedback from the driver and emit buzzing sounds to stimulate the driver's hearing. The voice device 8 is electrically connected to the internal power supply and is also controlled by the internal PLC programming program.

[0135] Working principle: When in use, the base 6 is placed at an appropriate position in the cockpit to ensure that the imaging device 9 can completely include the driver's head in the imaging range. The imaging information is processed in the image acquisition module 1, and the feature data is extracted and transmitted to the data storage module 5 and the face detection module 2 respectively. The face detection module 2 processes the feature data, summarizes the analysis report and transmits it to the fatigue judgment module 3. The data storage module 5 receives the feature data from the image acquisition module 1 in real time, forms the existing data and transmits it to the fatigue judgment module 3, and the data storage module 5 stores the historical accident data imported from the outside and transmits it to the fatigue judgment module 3. The fatigue judgment module 3 receives the analysis report from the face detection module 2, the existing data and the historical accident data from the data storage module 5, and compares the analysis report with the existing data and the historical accident data respectively, generates a judgment report and transmits it to the prompt alarm module 4. The prompt alarm module 4 first inquires the driver through the voice device 8 and receives the driver's feedback within the preset time threshold. If the driver cannot give normal feedback, the warning light 7 is used to flash an alarm and the voice device 8 is used to emit a buzzing sound to stimulate the driver.

[0136] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A fatigue driving reminder system based on face detection, characterized in that: It comprises an image acquisition module (1), a face detection module (2), a fatigue determination module (3), a prompt alarm module (4) and a data storage module (5); The image acquisition module (1) images and processes the driver's head, extracts feature data and transmits the feature data to the data storage module (5) and the face detection module (2) respectively; The face detection module (2) processes the feature data, summarizes the analysis report and transmits it to the fatigue determination module (3); The data storage module (5) receives the characteristic data from the image acquisition module (1) in real time, forms the continuous data and transmits it to the fatigue determination module (3), and the data storage module (5) stores the historical accident data imported from the outside and transmits it to the fatigue determination module (3); The fatigue determination module (3) receives the analysis report from the face detection module (2), the existing data and the historical accident data from the data storage module (5), and compares the analysis report with the existing data and the historical accident data, respectively, and generates a determination report which is transmitted to the prompt alarm module (4); The prompting and warning module (4) analyzes the judgment report and prompts the driver according to the judgment report, and collects driver feedback and warnings.

2. The fatigue driving reminder system based on face detection according to claim 1 is characterized in that: The image acquisition module (1) comprises a face imaging unit (11), a standardization processing unit (12), a face positioning unit (13), a face alignment unit (14) and a feature extraction unit (15); The face imaging unit (11) continuously and in real time images the driver's head to obtain imaging information; The standardization processing unit (12) performs standardization processing on the imaging information, including image brightness and contrast processing, size adjustment, cropping, and removal of irrelevant background information; The face positioning unit (13) locates the position and size of the face in the standardized imaging information; The face alignment unit (14) detects key points of the face and transforms the face to a standard template face for alignment; The feature extraction unit (15) extracts feature data from the aligned face images using a linear discriminant analysis method, wherein the feature data includes the driver's blinking frequency, eyeball activity, eye tremor, facial activity and head activity; The output end of the face imaging unit (11) is connected to the input end of the standardization processing unit (12), the output end of the standardization processing unit (12) is connected to the input end of the face positioning unit (13), the output end of the face positioning unit (13) is connected to the input end of the face alignment unit (14), and the output end of the face alignment unit (14) is connected to the input end of the feature extraction unit (15).

3. The fatigue driving reminder system based on face detection according to claim 2 is characterized in that: The specific steps of the linear discriminant analysis method are as follows: Calculate the within-category scatter matrix, the within-category scatter matrix S W Describes the degree of dispersion of samples within each category. For a binary classification problem, there are two categories, 0 and 1. The mathematical expression is: Among them, X0 and X1 represent the sample sets of two categories respectively, μ0 and μ1 represent the mean vectors of two categories respectively; Calculate the inter-class scatter matrix, the inter-class scatter matrix S B Describes the degree of dispersion between the center points of different categories. For the binary classification problem, its mathematical expression is: S B =(μ0-μ1)(μ0-μ1) T μ0 and μ1 are the mean vectors of the two categories respectively; Optimizing the objective function, the goal of the linear discriminant analysis method is to find a projection direction w so that the projected data is separated between categories and clustered within categories. This can be achieved by maximizing the ratio of the scatter matrix between categories to the scatter matrix within categories, that is, optimizing the following objective function: To find the optimal projection direction, we need to optimize the objective function J(w). The solution of J(w) has nothing to do with the length of w, but only with the direction of w. Let w T S W w=1, transforming the original problem into a constrained optimization problem; Using the Lagrange multiplier method, the constrained optimization problem is transformed into an unconstrained optimization problem, and the expression of w is obtained by solving it: The original data is projected into a low-dimensional space using the optimal projection direction w, and the samples are classified in the low-dimensional space.

4. The fatigue driving reminder system based on face detection according to claim 2 is characterized in that: The face detection module (2) comprises a blink frequency detection unit (21), an eyeball activity detection unit (22), an eye tremor detection unit (23), a facial activity detection unit (24), a head activity detection unit (25) and a summarizing unit (26); The blink frequency detection unit (21) receives the characteristic data and detects the blink frequency of the driver to obtain blink frequency data; The eyeball activity detection unit (22) receives the characteristic data and detects the rotation frequency and angle of the driver's eyeballs to obtain eyeball activity data; The eye tremor detection unit (23) receives the characteristic data and detects the frequency and amplitude of the driver's eye tremor to obtain eye tremor data; A facial activity detection unit (24) receives the feature data and detects the driver's facial activity to obtain facial activity data; The head movement detection unit (25) receives the characteristic data and detects the frequency and amplitude of the driver's head movement to obtain the head movement data; The summarizing unit (26) summarizes the blink frequency data, the eyeball movement data, the eye tremor data, the facial movement data and the head movement frequency to generate an analysis report; Input ends of the blink frequency detection unit (21), the eyeball activity detection unit (22), the eye tremor detection unit (23), the facial activity detection unit (24) and the head activity detection unit (25) are respectively connected to the output end of the feature extraction unit (15); The output ends of the blink frequency detection unit (21), the eyeball activity detection unit (22), the eye tremor detection unit (23), the facial activity detection unit (24) and the head activity detection unit (25) are respectively connected to the input end of the summarizing unit (26).

5. The fatigue driving reminder system based on face detection according to claim 4 is characterized in that: The data storage module (5) comprises a continuing data unit (51) and an accident data unit (52); The ongoing data unit (51) receives the feature data from the feature extraction unit (15) in real time and forms ongoing data, and its input end is connected to the output end of the feature extraction unit (15); The accident data unit (52) stores historical accident data imported from outside.

6. The fatigue driving reminder system based on face detection according to claim 5 is characterized in that: The fatigue determination module (3) receives the analysis report from the summary unit (26), the ongoing data from the ongoing data unit (51), and the historical accident data from the accident data unit (52), and compares the analysis report with the ongoing data and the historical accident data, respectively, to generate a determination report; The input end of the fatigue determination module (3) is respectively connected to the output ends of the summary unit (26), the survival data unit (51) and the accident data unit (52).

7. The fatigue driving reminder system based on face detection according to claim 6 is characterized in that: The prompt alarm module (4) comprises a prompt unit (41), a feedback unit (42) and an alarm unit (43); The prompting unit (41) receives the determination report from the fatigue determination module (3) and provides a prompt according to the content of the determination report; The feedback unit (42) collects driver feedback within a preset time threshold; The alarm unit (43) issues an alarm according to the feedback information; The input end of the prompt unit (41) is connected to the output end of the fatigue determination module (3), the input end of the feedback unit (42) is connected to the output end of the prompt unit (41), and the input end of the alarm unit (43) is connected to the output end of the feedback unit (42).

8. A fatigue driving reminder device based on face detection, characterized in that: It is applicable to a fatigue driving reminder system based on face detection as described in any one of claims 1 to 7, comprising a base (6); A warning light (7) is fixedly mounted on one side of the top of the base (6); The voice device (8) is fixedly mounted on the other side of the top of the base (6).

9. The fatigue driving reminder device based on face detection according to claim 8, characterized in that: Two clamping parts (61) are provided on one side of the base (6), one side of the top of which is rotatably connected to a turntable (62), the top of the turntable (62) is rotatably connected to a support rod 1 (63), and the other end of the support rod 1 (63) is rotatably connected to a support rod 2 (64), and cushion layers (65) are fixedly installed on opposite sides of the two clamping parts (61).

10. The fatigue driving reminder device based on face detection according to claim 9, characterized in that: It also includes an imaging device (9) which is rotatably connected to the other end of the second supporting rod (64).

Images