Intelligent alcohol induction type starting control system and method for new energy automobile

Abstract

The invention discloses an intelligent alcohol induction type starting control system for a new energy automobile, and belongs to the technical field of new energy driving modes. The system is composed of a pressure sensor, an alcohol detection module, a vehicle-mounted computer and a drunk driving automatic brake. The pressure sensor detects the pressure of a driver seat and judges whether a driver sits or not; the alcohol detection module detects whether the driver drinks; and the vehicle-mounted computer receives the information of the two and makes a judgment, and if the driver drinks, the drunk driving automatic brake is controlled to drive the brake pedal to a brake state. The drunk driving automatic brake comprises a plurality of parts, and the braking action is achieved through a specific transmission structure. In addition, an alternative scheme of the driver is provided, and the cost is reduced. The system effectively blocks drunk driving starting, does not interfere with starting of an engine of the extended-range and plug-in new energy automobile, ensures normal use of part of functions of the automobile, and provides powerful guarantee for traffic safety.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy drive methods, and particularly to an intelligent alcohol induction start control system for new energy vehicles. Background Art

[0002] Drunk driving poses a major risk to traffic safety; although traffic police conduct law enforcement inspections on the road from time to time to curb drunk driving behavior, there are still individual drivers who take chances and choose to drive after drinking. Drunk driving not only poses a serious threat to the life safety of the driver himself, but also greatly endangers the safety of other pedestrians and vehicles on the road, resulting in frequent traffic accidents and causing a large number of casualties and property losses.

[0003] In order to completely curb the drunk driving behavior of drivers, some technologies for detecting drinking and braking in the vehicle have emerged. For example, the publication number CN202448965U discloses an anti-drunk driving device for a vehicle, which belongs to the technology of "starting against the will of the driver or passenger" in the classification of B60T7 / 12; however, these existing technologies have many deficiencies. Some devices have low detection accuracy and are prone to misjudgment, resulting in the possibility of braking interruption even during normal driving; some devices have complex structures and high costs, making it difficult to promote and apply them on a large scale; there are also some devices that, when applied to new energy vehicles, will interfere with the normal operation of the vehicle, such as affecting engine start or battery management system, resulting in some functions of the vehicle being unable to be used normally, bringing inconvenience to the driver. Summary of the Invention

[0004] The present invention provides an intelligent alcohol induction start control system for new energy vehicles, and the technical problem to be solved is: to overcome the problems existing in the existing technologies for detecting drinking and braking in the vehicle, such as low detection accuracy, complex structure and high cost, and interference with the normal operation of new energy vehicles, and to provide an intelligent alcohol induction start blocking and warning control system with accurate detection, reasonable structure, controllable cost and no impact on the normal functions of new energy vehicles. In order to achieve the above invention purpose, the technical solution adopted by the present invention is:

[0005] In order to achieve the above invention purpose, the technical solution adopted by the present invention is: The system mainly consists of a pressure sensor, an alcohol detection module, an in-vehicle computer and a drunk driving automatic brake. The pressure sensor is installed in the driver's seat cushion and judges whether the driver is sitting in the driver's position by detecting the change of the seat cushion pressure. The alcohol detection module is arranged in the driver's area and can accurately detect whether the driver is drinking. The in-vehicle computer is electrically connected to the pressure sensor and the alcohol detection module, receives the detection information transmitted by the two, and analyzes and judges whether the driver in the driver's seat is drinking. When it is judged that the driver is drinking, the in-vehicle computer controls the drunk driving automatic brake to work, drives the brake pedal to rotate to the braking state, and prevents the vehicle from starting.

[0006] Furthermore, the automatic brake for drunk driving includes components such as a rotating sleeve, a compression expansion assembly, and a servo electric cylinder. The rotating sleeve is fixed to the rotating shaft of the brake pedal, and the inner circumference of its blind hole is an internal gear surface, and the end surface is a cut-off surface. The coupling shaft has an external expansion piece and a coupling rod, and the external expansion piece and the coupling rod are matched through an inclined surface, and a reset spring is provided in the coupling shaft. The spiral rod and the spiral sleeve form a spiral pair, one end of the spiral rod is fixed to the coupling shaft, and the spiral sleeve is slidably connected to the outer shell. The driver (such as a linear drive unit such as a servo electric cylinder) pushes the spiral rod and the coupling shaft to expand the external expansion piece and mesh with the internal gear surface. When the servo electric cylinder pushes the spiral sleeve to move linearly, it drives the rotating sleeve and the brake pedal to rotate to achieve braking.

[0007] Furthermore, in order to reduce costs, an alternative solution for the driver is designed. The driver consists of a pressure rod, a connecting rod, and a connector. The connector is rotatably connected to the spiral rod, the connecting rod connects the pressure rod and the connector, and the pressure rod is rotatably connected to the housing. When the output end of the servo electric cylinder presses down the pressure rod, the driver pushes the coupling shaft to make the outer expansion piece mesh with the inner gear surface, and then the servo electric cylinder pushes the spiral sleeve to realize the braking action.

[0008] Furthermore, a guide rod is fixed to the outer shell, and the guide rod passes through the spiral sleeve to ensure that the spiral sleeve is connected to the outer shell in a linear sliding manner, thereby improving movement stability.

[0009] Furthermore, a main reset spring is provided, with two ends of which are respectively connected to the outer shell and the spiral sleeve. When the spiral sleeve is not acted upon by the servo electric cylinder, the spiral sleeve is maintained at a preset position to facilitate system reset.

[0010] Furthermore, the alcohol detection module includes a detection shell and an alcohol detection sensor module. The detection shell is provided with an air inlet and an air outlet. The alcohol detection sensor module is located inside the shell and is electrically connected to the on-board computer to ensure accurate detection of the alcohol content in the vehicle.

[0011] Furthermore, a cavity is arranged around the outer periphery of the coupling rod in the coupling shaft, and a secondary reset spring is built in, whose elastic force causes the coupling rod to move in a direction away from the inner gear surface, so as to facilitate the reset of the coupling rod and the outer expansion member in a non-working state.

[0012] The beneficial effects of the present invention are:

[0013] Effectively curb drunk driving: Through accurate alcohol detection modules and reliable brake control mechanisms, the vehicle is automatically blocked from starting after the driver drinks, forcing the vehicle to enter a braking state, eliminating drunk driving from the source, significantly improving the level of road traffic safety, and reducing traffic accidents caused by drunk driving.

[0014] Strong adaptability: For range-extended and plug-in new energy vehicles, the braking state of this system is mainly achieved by driving the brake pedal for braking, without interfering with the engine startup. Even when the vehicle battery has a low charge or the driver is waiting for a substitute driver after drinking, the engine can still be used to charge the on-vehicle battery, ensuring stable power supply for the air conditioner and entertainment system, thus enhancing the user experience and the convenience of vehicle use.

[0015] Controllable cost: An alternative solution for the driver is designed. Compared with the original drive structure, it can reduce the cost of the driver for the servo electric cylinder. On the premise of ensuring the system functions, it improves the cost performance of the product, which is conducive to the large-scale popularization and application of this system.

[0016] Reasonable structure and high stability: The outer shell is provided with guide rods to ensure the linear sliding of the helical sleeve. The design of the main return spring and the secondary return spring enables the components of the system to automatically reset in the non-working state, ensuring the stability and reliability of the system during multiple uses and extending the service life of the system.

[0017] Accurate detection: The alcohol detection module adopts a detection shell with an air inlet and an air outlet and a high-precision alcohol detection sensor module, which can accurately detect whether the driver is drinking, reduce the occurrence of misjudgment, and ensure the normal operation of the system. Description of the drawings

[0018] Figure 1 It is a schematic diagram of the concept of the present invention;

[0019] Figure 2 is Figure 1 the enlarged view at A in

[0020] Figure 3 is Figure 2 the enlarged schematic view at B in

[0021] Drunk driving automatic brake 100, rotating sleeve 110, pressure-expansion component 120, servo electric cylinder 130, coupling shaft 140, screw rod 150, helical sleeve 160, driver 170, main return spring 180;

[0022] Brake pedal 200;

[0023] On-vehicle computer 300, alcohol detection module 400. Detailed implementation manners

[0024] The following further illustrates the detailed implementation manners of the present invention with reference to the drawings. The same components are denoted by the same reference numerals.

[0025] It should be noted that the terms "front", "rear", "left", "right", "up" and "down" used in the following description refer to the directions in the drawings, and the terms "inner" and "outer" respectively refer to the directions towards or away from the geometric center of a specific component.

[0026] In order to make the content of the present invention easier to be clearly understood, the following will combine the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention.

[0027] As Figures 1 to 3 shown, the intelligent alcohol induction start control system for new energy vehicles consists of a pressure sensor, an alcohol detection module, and an in-vehicle computer.

[0028] The pressure sensor is arranged in the driver's seat cushion and is used to detect the pressure change of the cushion; when a person sits on the driver's seat cushion, the pressure sensor will detect the pressure change of the driver's seat cushion. The pressure sensor is mainly used to detect whether the driver is sitting on the driver's seat cushion.

[0029] The alcohol detection module 400 is arranged in the driver's area and is used to detect whether the driver drinks alcohol.

[0030] The in-vehicle computer is electrically connected to the pressure sensor and the alcohol detection module. The in-vehicle computer is used to receive the detection information of the pressure sensor and the alcohol detection module; specifically, when the in-vehicle computer learns that there is someone in the driver's seat through the pressure sensor, it actively obtains the detection information of the alcohol detection module on the alcohol content in the vehicle, and then judges whether the driver in the driver's seat drinks alcohol.

[0031] The drunk driving automatic brake is used to drive the brake pedal to rotate to the braking state.

[0032] Among them, when a person sits on the driver's seat cushion of the vehicle, the in-vehicle computer receives the signal of the increased pressure of the pressure sensor. Subsequently, the in-vehicle computer actively obtains the detection information of the alcohol detection module. When the in-vehicle computer receives the information that the driver drinks alcohol, the in-vehicle computer drives the brake pedal to rotate to the braking state through the drunk driving automatic brake.

[0033] To sum up, after the driver drinks alcohol, after the in-vehicle computer obtains the drinking information of the driver through the alcohol detection module, the in-vehicle computer drives the brake pedal to rotate to the braking state through the drunk driving automatic brake; the above braking state is determined by the driver's drinking information and cannot be changed by the driver's own will.

[0034] In addition, for extended-range and plug-in new energy vehicles, since the above braking state mainly relies on driving the brake pedal to brake, it will not interfere with the start of the engine. Therefore, even if the battery power of extended-range and plug-in new energy vehicles is low, the driver can still use the engine to charge the in-vehicle battery to ensure the stable power supply of the air conditioner and entertainment system. In this way, even if the driver needs to wait for someone else to drive after drinking alcohol, they can still enjoy the air conditioner and entertainment functions of the vehicle safely and comfortably.

[0035] Further, the drunk driving automatic brake 100 includes a rotating sleeve 110, a compression and expansion component 120, and a servo electric cylinder 130;

[0036] The housing 120 is fixedly arranged.

[0037] The rotating sleeve 110 is fixedly connected to the rotating shaft of the brake pedal. A blind hole is formed in the rotating sleeve 110. The inner peripheral surface of the blind hole is an internal gear surface 111, and the end surface of the blind hole is a cut-off surface.

[0038] The coupling shaft 140 is radially provided with a plurality of sliding coupling rods 141 and axially provided with a sliding outer expansion member 142. The contact surface between the outer expansion member 142 and the coupling rods 141 is an inclined surface. In the initial state, the outer expansion member 142 and the coupling rods 141 of the coupling shaft 140 are located in the blind hole; wherein, during the process that the coupling shaft 140 drives the outer expansion member 142 to move towards the cut-off surface of the blind hole, the cut-off surface presses the outer expansion member 142 to move relative to the coupling rods 141, and the outer end of the outer expansion member 142 meshes with the internal gear surface 111. A return spring 143 is arranged inside the coupling shaft 140. The return spring 143 acts on the outer expansion member 142 by its own elastic force. When the outer expansion member 142 is not in contact with the cut-off surface, the elastic force of the return spring 143 always maintains the outer expansion member 142 at a preset position inside the coupling shaft 140.

[0039] The external surface of the screw rod 150 is provided with a helical groove, and the helical sleeve 160 is provided with a helical protrusion. The helical groove matches the helical protrusion, that is, the screw rod 150 and the helical sleeve 160 form a screw pair; the screw rod 150 is sleeved inside the helical sleeve 160, and the helical sleeve 160 is slidably arranged with the housing 120. One end of the screw rod 150 is fixedly connected to the coupling shaft 140;

[0040] The driver 170 is used to linearly push the screw rod 150 and the coupling shaft 140 to achieve the state where the outer expansion member 142 is unfolded and meshes with the internal gear surface 111; the driver 170 is a linear drive unit, such as a servo electric cylinder or the like.

[0041] The piston cylinder output end of the servo electric cylinder 130 is used to push the helical sleeve 160 to move linearly; since the screw rod 150 and the helical sleeve 160 form a screw pair, when the helical sleeve 160 is pushed to slide, the screw rod 150 will drive the coupling shaft 140 to rotate synchronously.

[0042] In summary, when the driver waiting for a designated driver after drinking is sitting in the driver's seat, the in-vehicle computer actively obtains the detection information of the alcohol detection module. First, the in-vehicle computer controls the driver 170 to push the screw rod 150 and the coupling shaft 140 to move synchronously towards the cut-off surface. The cut-off surface presses the outward expansion member 142 to move relative to the coupling rod 141, and the outer end of the outward expansion member 142 meshes with the internal gear surface 111. Subsequently, the in-vehicle computer controls the servo electric cylinder 130 to push the screw sleeve 160 to move linearly, and the screw rod 150 and the coupling shaft 140 rotate synchronously, thereby driving the rotating sleeve 110 and the brake pedal to rotate synchronously, so that the brake pedal rotates to the braking state.

[0043] Further, the driver 170 includes a pressure rod 171, a connecting rod 172, and a connecting head 173;

[0044] The connecting head 173 is rotatably connected to the screw rod 150.

[0045] Both ends of the connecting rod 172 are respectively rotatably connected to the connecting head 173 and the pressure rod 171, and the pressure rod 171 is rotatably connected to the housing 120.

[0046] Among them, during the process of the output end of the servo electric cylinder 130 extending, it first presses down the pressure rod 171. The length dimension of the driver 170 increases in the moving direction of the coupling shaft 140. When the pressure rod 171 reaches the lowest state, the cut-off surface presses the outward expansion member 142 to move relative to the coupling rod 141, and the outer end of the outward expansion member 142 meshes with the internal gear surface 111.

[0047] In summary, since the outer end of the outward expansion member 142 meshes with the internal gear surface 111 when the output end of the servo electric cylinder 130 presses the pressure rod 171 to the lowest state. Therefore, when it is necessary to control the vehicle to brake, the in-vehicle computer only needs to control the output end of the servo electric cylinder 130 to extend; specifically, during the process of the output end of the servo electric cylinder 130 extending, first, the servo electric cylinder 130 presses down the driver 170, so that the outer end of the outward expansion member 142 meshes with the internal gear surface 111; subsequently, the servo electric cylinder 130 pushes the screw sleeve 160 to move linearly, and the screw rod 150 and the coupling shaft 140 rotate synchronously, thereby driving the rotating sleeve 110 and the brake pedal to rotate synchronously, so that the brake pedal rotates to the braking state.

[0048] Further, the housing 120 is also fixedly provided with a guide rod 121, and the guide rod 170 slidably penetrates through the screw sleeve 160. Thus, the screw sleeve 160 is linearly slidably connected to the housing 120.

[0049] Further, it also includes a main return spring 180; both ends of the main return spring 180 are respectively fixedly connected to the housing 120 and the screw sleeve 160.

[0050] When the spiral sleeve 160 is not affected by the servo electric cylinder 130, the elastic force of the main return spring 180 always maintains the spiral sleeve 160 at a preset position.

[0051] Further, the alcohol detection module 400 includes a detection housing 410 and an alcohol detection sensor module 430;

[0052] The detection housing 410 is provided with an air inlet and an air outlet, and the alcohol detection sensor module 430 is located inside the detection housing 410; the alcohol detection sensor module 430 is electrically connected to the vehicle-mounted computer 300.

[0053] Further, a cavity 144 is provided around the outer periphery of the coupling rod 141 of the coupling shaft 140, and a secondary return spring 145 is located in the cavity 144. Both ends of the secondary return spring 145 are fixed to the inner wall of the cavity 144 and the coupling rod 141 respectively; the elastic force of the secondary return spring 145 always moves the coupling rod 141 in a direction away from the inner gear surface 111.

[0054] Further, based on the above system, a method for preventing fraud detection based on face recognition is also loaded in the vehicle-mounted computer, as follows:

[0055] New energy vehicle intelligent alcohol induction start control method,

[0056] S100. Obtain the main driver's door open signal;

[0057] S200. Obtain the facial image of the person entering the driver's seat through the main driver's door, denoted as the driver's facial image; the driver's facial image is obtained by a camera installed in the driving area;

[0058] S300. Obtain the facial image of the person blowing into the air inlet of the alcohol detection module, denoted as the detected person's facial image;

[0059] S400. Identify whether the driver's facial image and the detected person's facial image are the same person through a face recognition algorithm. If they are not the same person, that is, fraud detection, the vehicle-mounted computer drives the brake pedal to rotate to the braking state through the drunk driving automatic brake; if they are the same person and the alcohol detection module passes the detection, the vehicle-mounted computer does not rotate the brake pedal to the braking state.

[0060] Further, the following improvements are made to the traditional face recognition algorithm, specifically including the following steps:

[0061] Optimize image preprocessing: On the basis of existing denoising and grayscale conversion, introduce the adaptive histogram equalization technique. This technique can adaptively adjust the contrast according to the grayscale distribution of local regions of the image, making facial details clearer. For example, it can highlight the textures of key parts such as eyes and eyebrows, enhancing the accuracy of subsequent feature extraction. For facial shadows caused by lighting problems, adopt a shadow removal algorithm based on deep learning. This algorithm is trained with a large number of facial images with and without shadows to learn the feature differences between shadows and normal facial regions, thus effectively removing the shadows in the image and avoiding interference from shadows to subsequent recognition.

[0062] Improve anti-interference ability: For the situation where the driver wears glasses, hats and other occluders, train a special occlusion-robust model. By collecting a large number of facial images with different occluders, train the model so that it can learn the facial feature rules in the case of occlusion, and thus can accurately identify even when there is occlusion. When it is detected that there is occlusion on the face, automatically adjust the recognition strategy, increase the feature weight of the unoccluded area, and preferentially use the features of the unoccluded part for recognition. For the changes in facial features caused by expression changes, construct an expression-invariant model. Using deep learning technology, learn the common facial features under different expressions, making the recognition algorithm have a certain robustness to expression changes, and ensuring that the identity can still be accurately recognized in different expression states such as when the driver smiles or frowns.

[0063] Feature extraction: In addition to traditional facial feature point localization, combine the convolutional neural network (CNN) technology in deep learning. CNN can automatically learn more abstract and representative facial features. For example, through multiple convolutional layers and pooling layers, extract deep texture features, shape features, etc. of the face. Compared with traditional methods, it can describe facial features more accurately and improve the recognition accuracy. Utilize local binary pattern (LBP) features, which have a certain robustness to lighting changes. LBP generates binary patterns by comparing the grayscale values of the central pixel and neighboring pixels to describe the local texture information of the face. Combining LBP features with the features extracted by CNN can further enrich the facial feature expression and improve the recognition effect.

[0064] Enhanced recognition and comparison: Use a combination of cosine similarity and Euclidean distance; cosine similarity is used to measure the directional similarity between feature vectors, and Euclidean distance is used to measure the spatial distance between feature vectors. The combination of the two can more comprehensively evaluate the similarity degree of two facial images and reduce misjudgment.

[0065] The above are only the preferred embodiments of this invention patent and are not intended to limit this invention patent. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this invention patent shall be included within the protection scope of this invention patent.

Claims

1. Intelligent alcohol-sensing start-up control system for new energy vehicles, characterized in that: include: A pressure sensor is provided in the driver's seat cushion to detect changes in the pressure of the seat cushion; The alcohol detection module is set in the main driving area to detect whether the driver has been drinking; The on-board computer is electrically connected to the pressure sensor and the alcohol detection module, and is used to receive detection information from the pressure sensor and the alcohol detection module, and determine whether the driver in the driving seat has drunk alcohol; The automatic brake for drunk driving is used to drive the brake pedal to rotate to the braking state; When the on-board computer receives information that the driver has been drinking, it controls the drunk driving automatic brake to drive the brake pedal to the braking state.

2. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 1 is characterized in that: The drunk driving automatic brake includes a rotating sleeve, a compression and expansion assembly, and a servo electric cylinder. The rotating sleeve is fixedly connected to the rotating shaft of the brake pedal. The rotating sleeve is provided with a blind hole. The inner circumferential surface of the blind hole is an internal gear surface, and the end surface of the blind hole is a cut-off surface; it also includes a coupling shaft. The coupling shaft is radially provided with multiple sliding coupling rods, and an axially provided with a sliding external expansion component. The contact surface between the external expansion component and the coupling rod is an inclined surface, and a reset spring is provided in the coupling shaft; the spiral rod and the spiral sleeve form a spiral pair, one end of the spiral rod is fixedly connected to the coupling shaft, and the spiral sleeve is slidably arranged with the outer shell; the driver is used to linearly push the spiral rod and the coupling shaft to expand the external expansion component and mesh with the internal gear surface; the piston cylinder output end of the servo electric cylinder is used to push the spiral sleeve to move linearly.

3. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 2 is characterized in that: The driver includes a pressure rod, a connecting rod, and a connecting head. The connecting head is rotatably connected to the spiral rod. The two ends of the connecting rod are respectively rotatably connected to the connecting head and the pressure rod. The pressure rod is rotatably connected to the outer shell. When the output end of the servo electric cylinder extends out and presses down the pressure rod, the outer end of the outer expansion piece is meshed with the inner gear surface.

4. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 2 is characterized in that: The shell is also fixedly provided with a guide rod, which slides through the spiral sleeve so that the spiral sleeve is linearly slidably connected with the shell.

5. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 2 is characterized in that: It also includes a main reset spring, the two ends of which are respectively fixedly connected to the outer shell and the spiral sleeve, so as to maintain the spiral sleeve in a preset position.

6. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 1 is characterized in that: The alcohol detection module comprises a detection shell and an alcohol detection sensor module. The detection shell is provided with an air inlet and an air outlet. The alcohol detection sensor module is located in the detection shell and is electrically connected to the on-board computer.

7. The intelligent alcohol-sensing start-up control system for new energy vehicles according to claim 2 is characterized in that: The coupling shaft is provided with a cavity around the outer circumference of the coupling rod, and the secondary return spring is located in the cavity. The two ends of the secondary return spring are respectively fixed to the inner wall of the cavity and the coupling rod, and the elastic force of the secondary return spring always moves the coupling rod in the direction of separating from the inner gear surface.

8. An intelligent alcohol-sensing start-up control method for new energy vehicles, including a facial recognition algorithm, characterized in that: The following steps are involved: S100, obtaining a driver's door opening signal; S200, acquiring a facial image of a person entering the driving seat through the main driving door, recorded as the driver's facial image; the driver's facial image is acquired by a camera installed in the driving area; S300, obtaining a facial image of a person blowing into an air inlet of an alcohol detection module, and recording the image as a facial image of the detection person; S400. Identify through a facial recognition algorithm whether the driver's facial image and the detection person's facial image are the same person. If they are not the same person, it is a false detection, and the on-board computer drives the brake pedal to rotate to the braking state through the drunk driving automatic brake; if they are the same person and the alcohol detection module passes the detection, the on-board computer does not rotate the brake pedal to the braking state.

9. The intelligent alcohol-sensing start-up control method for new energy vehicles according to claim 8 is characterized in that: Optimize image preprocessing: Based on the existing denoising and grayscale, introduce adaptive histogram equalization technology; then adaptively adjust the contrast to make facial details clearer; For facial shadows caused by lighting problems, a shadow removal algorithm based on deep learning is used. Improve anti-interference ability: Train a special occlusion robust model. By collecting a large number of facial images with different occlusions, train the model so that it can learn the facial feature patterns under occlusion, so that it can accurately identify even when there is occlusion. Feature extraction: In addition to traditional facial feature point positioning, the convolutional neural network technology in deep learning is combined to learn more abstract and representative facial features to improve recognition accuracy. LBP is used to generate binary patterns by comparing the grayscale values ​​of the central pixel and the neighboring pixels to describe the local texture information of the face; LBP features are combined with features extracted by CNN to improve recognition results. Strong recognition comparison: A combination of cosine similarity and Euclidean distance is used; cosine similarity is used to measure the directional similarity between feature vectors, and Euclidean distance is used to measure the spatial distance between feature vectors. The combination of the two can more comprehensively evaluate the similarity between two facial images and reduce misjudgment.

Images