A face recognition method for online car-hailing, a terminal device and a storage medium

Abstract

The application relates to the technical field of identity recognition, in particular to a network car hailing face recognition method, a terminal device and a storage medium. The method comprises the following steps: acquiring face feature data corresponding to a driver face image and a passenger face image of a current network car hailing respectively, and generating corresponding mouth region features; judging whether the driver and the passenger are in a communication state according to the mouth region features; if the driver and the passenger are in the communication state, acquiring corresponding communication speech; selecting a target voice tone corresponding to a first appearance of a scurrilous word in the speech segment; if the target voice tone is consistent with a driver authentication voice tone, determining a first abnormal state; acquiring a facial expression feature corresponding to the driver face image; and if the facial expression feature is consistent with a fatigue driving feature, determining a second abnormal state. The network car hailing face recognition method, the terminal device and the storage medium provided by the application have the effect of improving the supervision intensity of the network car hailing in a driving process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of identity recognition, in particular to a face recognition method for online car hailing, a terminal device and a storage medium. BACKGROUND

[0002] With the rapid development of the Internet, people's lifestyle is also changing. If people need to travel, they can use mobile software to book a suitable vehicle at any time and anywhere.

[0003] Currently, in the actual passenger carrying process of online car hailing, there are many events such as mismatch between people and vehicles, drivers insulting and beating passengers, dangerous driving, and even indecent behavior in the vehicle, which seriously threatens the safety of online car hailing. SUMMARY

[0004] In order to improve the supervision of online car hailing during driving, the present application provides a face recognition method for online car hailing, a terminal device and a storage medium.

[0005] The present application provides a face recognition method for online car hailing, comprising the following steps:

[0006] Obtain and identify the driver's face image of the current online car hailing, and generate the corresponding identity authentication result;

[0007] Determine whether the identity authentication result is identity authentication success;

[0008] If the identity authentication result is identity authentication success, determine whether the current online car hailing is in the passenger carrying state;

[0009] If the current online car hailing is in the order receiving and passenger carrying state, obtain the face feature data corresponding to the driver's face image and the passenger's face image of the current online car hailing, respectively;

[0010] Identify the face feature data to generate the corresponding mouth region feature;

[0011] Determine whether the driver and the passenger are in communication state according to the mouth region feature;

[0012] If the driver and the passenger are in communication state, obtain the corresponding communication voice;

[0013] Determine whether there is insulting vocabulary in the communication voice;

[0014] If there is insulting vocabulary in the communication voice, obtain the voice segment corresponding to the insulting vocabulary;

[0015] Select the target voice tone corresponding to the first appearance of the insulting vocabulary in the voice segment;

[0016] identifying and determining whether the target voice tone is consistent with the driver authentication voice tone;

[0017] if the target voice tone is consistent with the driver authentication voice tone, determining a first abnormal state;

[0018] if the abusive words do not exist in the communication voice, obtaining corresponding facial features according to the driver face image;

[0019] determining whether the facial features are consistent with the fatigue driving features;

[0020] if the facial features are consistent with the fatigue driving features, determining a second abnormal state;

[0021] according to the first abnormal state or the second abnormal state, issuing a corresponding warning prompt.

[0022] By adopting the above technical solution, based on the successful authentication of the current online car driver, the face images of the current online car driver and the passenger are obtained, so as to generate corresponding mouth region features by identifying the corresponding facial feature data in the face images, to determine whether the driver and the passenger in the current online car are in communication state, further to obtain the communication voice when the driver and the passenger are in communication state, to determine whether the driver and the passenger have quarreled during the driving of the vehicle, and to determine whether the target voice tone corresponding to the first appearance of the abusive words in the voice segment is consistent with the driver authentication voice tone, so as to facilitate the responsibility identification of the quarrel inducer affecting the driving safety, further to determine the first abnormal state, if the communication voice does not contain the relevant abusive words, to obtain the facial features of the driver face image to determine whether the facial features are consistent with the fatigue driving features, to identify the second abnormal state if the facial features are consistent with the fatigue driving features, and to issue a corresponding warning prompt to the driver according to the determination result of the actual abnormal state, thereby improving the supervision intensity of the online car during driving.

[0023] Optionally, the mouth region features include a driver mouth opening and closing period and a passenger mouth opening and closing period, and the determination of whether the driver and the passenger are in communication state according to the mouth region features comprises the following steps:

[0024] determining whether there is continuous and alternating overlap between the driver mouth opening and closing period and the passenger mouth opening and closing period;

[0025] if there is the continuous and alternating overlap between the driver mouth opening and closing period and the passenger mouth opening and closing period, obtaining first voice information in the driver mouth opening and closing period and second voice information in the passenger mouth opening and closing period, respectively;

[0026] determining whether corresponding response information exists in the first voice information and the second language information;

[0027] if the corresponding response information exists in the first voice information and the second language information, determining that the driver and the passenger are in the communication state;

[0028] if the corresponding response information does not exist in the first voice information and the second language information, determining that the driver and the passenger are not in the communication state.

[0029] By using the above technical solution, it is determined whether corresponding continuous and alternating overlap exists between the mouth opening and closing periods of the driver and the passenger, and whether corresponding response information exists in the first voice information corresponding to the driver and the second language information corresponding to the passenger in the above periods, thereby improving the accuracy of determining whether the driver and the passenger are in the communication state.

[0030] Optionally, after the driver and the passenger are in the communication state if the corresponding response information exists in the first voice information and the second language information, the method further comprises the following steps:

[0031] obtaining a communication duration corresponding to the communication state;

[0032] determining whether the communication duration exceeds a dangerous driving communication duration;

[0033] if the communication duration exceeds the dangerous driving communication duration, recognizing the face feature data corresponding to the driver, and generating corresponding eye region features;

[0034] obtaining a corresponding gaze point according to the eye region features;

[0035] determining whether the gaze point exceeds a road region where the current online car-hailing vehicle travels;

[0036] if the gaze point exceeds the road region where the current online car-hailing vehicle travels, obtaining a corresponding gaze duration;

[0037] determining whether the gaze duration reaches an abnormal gaze shift duration;

[0038] if the gaze duration reaches the abnormal gaze shift duration, issuing corresponding attention distraction warning information.

[0039] By adopting the technical scheme, on the basis that the communication duration of the driver and the passenger exceeds the dangerous driving communication duration in the driving safety specification, whether the driver has the inattentive driving condition in the driving process of the current online car-hailing vehicle is determined by further combining whether the gaze point of the driver exceeds the road region currently traveled by the online car-hailing vehicle and whether the gaze duration corresponding to the gaze point exceeding the road region currently traveled by the online car-hailing vehicle exceeds the corresponding abnormal gaze shift duration, so that the supervision intensity of the driving attention of the driver of the online car-hailing vehicle in the driving process is improved.

[0040] Optionally, after the gaze duration reaches the abnormal gaze shift duration, the method further comprises the following steps:

[0041] determining whether the gaze point is shifted to the road region within a preset gaze correction duration;

[0042] if the gaze point is not shifted to the road region within the preset gaze correction duration, switching the driving mode of the current online car-hailing vehicle to a safe automatic driving mode, and acquiring real-time positioning of the current online car-hailing vehicle.

[0043] By adopting the technical scheme, if the gaze point of the driver of the current online car-hailing vehicle is not shifted to the road region within the preset gaze correction duration specified in the driving safety specification, the driving mode of the current online car-hailing vehicle is switched to the safe automatic driving mode, and the real-time positioning of the current online car-hailing vehicle at this time is acquired, so that the safety of the vehicle driving when the driver of the current online car-hailing vehicle is inattentive is improved.

[0044] Optionally, after determining that the driver and the passenger are not in the communication state if the corresponding response information does not exist in the first voice information and the second voice information, the method further comprises the following steps:

[0045] acquiring road condition information corresponding to the road traveled by the current online car-hailing vehicle;

[0046] acquiring vehicle driving actions corresponding to the current online car-hailing vehicle;

[0047] determining whether the vehicle driving actions conform to road condition driving safety standards corresponding to the road condition information;

[0048] if the vehicle driving actions do not conform to the road condition driving safety standards corresponding to the road condition information, generating a corresponding safe driving control instruction according to the road condition driving safety standards to control the driving actions of the current online car-hailing vehicle.

[0049] By adopting the technical scheme, the current online car is controlled by combining the road condition information of the current online car and the safe driving control instruction corresponding to the road condition driving safety standard, so that the occurrence of relevant driving safety accidents caused by the driver of the current online car being distracted in the driving process due to long-time communication with the outside world is reduced.

[0050] Optionally, after judging whether the abusive words exist in the communication voice, the method further includes the following steps:

[0051] According to the driver's face image, the corresponding facial expression feature is obtained;

[0052] It is judged whether the facial expression feature conforms to the fatigue driving feature;

[0053] If the facial expression feature conforms to the fatigue driving feature, the current online car is controlled to drive to a safe parking area according to the prepared emergency control instruction, and the driving control instruction corresponding to the current online car is stopped.

[0054] By adopting the technical scheme, on the basis of the driver of the current online car arguing with the passenger, it is further judged whether the driver's facial expression feature conforms to the fatigue driving feature. If the driver's facial expression feature conforms to the fatigue driving feature, the current online car is controlled to drive to a safe parking area according to the prepared emergency control instruction, so that the safety supervision intensity of the current online car in the driving process is improved.

[0055] Optionally, after the target voice tone conforms to the driver's authentication voice tone, it is determined that the first abnormal state, the method further includes the following steps:

[0056] The steering wheel monitoring video corresponding to the current online car is obtained;

[0057] It is judged whether the driver's handheld operation action appears in the steering wheel monitoring video;

[0058] If the driver's handheld operation action does not appear in the steering wheel monitoring video, the current online car is braked according to the emergency braking instruction.

[0059] By adopting the technical scheme, on the basis of the driver of the current online car arguing with the passenger, it is further judged whether the driver's handheld operation action appears in the steering wheel monitoring video. If the driver's handheld operation action does not appear, the current online car is braked according to the emergency braking instruction, so that the safety supervision intensity of the current online car in the driving process is improved.

[0060] Optionally, after the identity authentication result is identity authentication success, it is judged whether the current online car is in the passenger carrying state, the method further includes the following steps:

[0061] If the current online car hailing is in the unaccepted ride state, an in-vehicle monitoring picture corresponding to the current online car hailing is acquired;

[0062] According to the in-vehicle monitoring picture, it is determined whether the passenger is in the current online car hailing;

[0063] If the passenger is in the current online car hailing, a face image corresponding to the passenger is acquired and recognized, and a corresponding recognition authentication result is generated;

[0064] If the recognition authentication result shows that the identity authentication is abnormal, an engine-off control instruction is acquired and the current online car hailing is controlled to be turned off according to the engine-off control instruction.

[0065] By adopting the above technical solution, it is determined according to the in-vehicle monitoring picture whether the passenger is in the current online car hailing in the unaccepted ride state, the passenger is further subjected to face identity authentication, if the recognition authentication result shows that the identity authentication is abnormal, the current online car hailing is controlled to be turned off according to the engine-off control instruction, thereby reducing the occurrence of the situation that the driver of the current online car hailing accepts private jobs.

[0066] In a second aspect, the present application provides a terminal device, which adopts the following technical solution:

[0067] A terminal device includes a memory and a processor, the memory stores computer instructions capable of running on the processor, and when the processor loads and executes the computer instructions, the above-mentioned online car hailing face recognition method is adopted.

[0068] By adopting the above technical solution, the above-mentioned online car hailing face recognition method is used to generate computer instructions, which are stored in the memory to be loaded and executed by the processor, so that the terminal device is made of the memory and the processor, and is convenient to use.

[0069] In a third aspect, the present application provides a computer readable storage medium, which adopts the following technical solution:

[0070] A computer readable storage medium stores computer instructions, and when the computer instructions are loaded and executed by a processor, the above-mentioned online car hailing face recognition method is adopted.

[0071] By adopting the above technical solution, the above-mentioned online car hailing face recognition method is used to generate computer instructions, which are stored in the computer readable storage medium to be loaded and executed by the processor, and the computer readable storage medium is convenient for reading and storing the computer instructions.

[0072] In summary, the present application includes at least one of the following beneficial technical effects: based on the current online car driver identity authentication success, the current online car driver and passenger face images are obtained respectively, so as to generate the corresponding mouth region features by identifying the corresponding facial feature data in the above face images, to judge whether the driver and passenger in the current online car are in communication state, further by obtaining the communication voice when the driver and passenger are in communication state, it is convenient to judge whether the driver and passenger have a quarrel in the process of vehicle driving, and whether the target voice tone corresponding to the first appearance of abusive vocabulary selected from the voice segment conforms to the driver authentication voice tone, so as to facilitate the responsibility identification of the quarrel inducer affecting the driving safety, and further determine the first abnormal state, if there is no related abusive vocabulary in the communication voice, the corresponding facial features in the driver face image are obtained to determine whether they conform to the fatigue driving features, and the situation conforming to the fatigue driving features is identified as the second abnormal state, and the corresponding warning prompt is sent to the driver combined with the determination result of the actual abnormal state, so as to improve the supervision intensity of online car in the process of driving. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 is a flowchart of steps S101 to S116 in the online car face recognition method provided by the present application.

[0074] Figure 2 is a flowchart of steps S201 to S205 in the online car face recognition method provided by the present application.

[0075] Figure 3 is a flowchart of steps S301 to S308 in the online car face recognition method provided by the present application.

[0076] Figure 4 is a flowchart of steps S401 to S402 in the online car face recognition method provided by the present application.

[0077] Figure 5 is a flowchart of steps S501 to S504 in the online car face recognition method provided by the present application.

[0078] Figure 6 is a flowchart of steps S601 to S603 in the online car face recognition method provided by the present application.

[0079] Figure 7 is a flowchart of steps S701 to S703 in the online car face recognition method provided by the present application.

[0080] Figure 8is a flowchart of steps S801 to S804 in a method for face recognition of online car-hailing provided by the present application. DETAILED DESCRIPTION

[0081] The following will be described in detail in combination with the accompanying Figures 1-8 The present application will be further described in detail.

[0082] Embodiments of the present application disclose a method for face recognition of online car-hailing, as shown in the accompanying drawings, comprising the following steps: Figure 1

[0083] S101. Obtain and identify the driver's face image of the current online car-hailing, and generate a corresponding identity authentication result;

[0084] S102. Determine whether the identity authentication result is identity authentication success;

[0085] S103. If the identity authentication result is identity authentication success, determine the driving state type of the current online car-hailing;

[0086] S104. If the driving state type of the current online car-hailing is order-accepting and passenger-carrying state, obtain the face feature data corresponding to the driver's face image and the passenger's face image of the current online car-hailing respectively;

[0087] S105. Identify the face feature data, and generate the corresponding mouth region feature;

[0088] S106. Determine whether the driver and the passenger are in communication state according to the mouth region feature;

[0089] S107. If the driver and the passenger are in communication state, obtain the corresponding communication speech;

[0090] S108. Determine whether there is abusive vocabulary in the communication speech;

[0091] S109. If there is abusive vocabulary in the communication speech, obtain the speech segment corresponding to the abusive vocabulary;

[0092] S110. Select the target voice tone corresponding to the first appearance of the abusive vocabulary in the speech segment;

[0093] S111. Identify and determine whether the target voice tone conforms to the driver's authentication voice tone;

[0094] S112. If the target voice tone conforms to the driver's authentication voice tone, determine that it is the first abnormal state;

[0095] S113. If there is no abusive vocabulary in the communication speech, obtain the expression feature corresponding to the driver's face image;

[0096] S114. Determine whether the expression feature conforms to the fatigue driving feature; ​

[0097] S115. If the facial features match the fatigue driving features, determine as the second abnormal state;

[0098] S116. According to the first abnormal state or the second abnormal state, a corresponding warning prompt is issued.

[0099] In actual application, with the increase of illegal passenger vehicles, the market order is seriously disturbed. Moreover, due to the non-strict audit mechanism of online car-hailing, some black cars that do not meet the operation standards enter the online car-hailing industry. Whether the driver identity authentication information screening is not strict or the online car-hailing order account does not match the actual driver, it shows that there are loopholes in the access mechanism measures of the relevant platform.

[0100] In order to increase the supervision and audit of the current online car-hailing driver identity information in step S101, the face image of the current online car-hailing driver, i.e. the driver face image, is obtained, and the identity authentication result corresponding to the current online car-hailing driver is generated through the authentication of the online car-hailing supervision platform.

[0101] It should be noted that the current online car-hailing is one-to-one corresponding to the driver, i.e. the current online car-hailing driver. The driver in front of the driver is equipped with a camera with face recognition function. If the driver wants to drive and take orders, he can only start after the face recognition is successful, i.e. the identity authentication result in step S103 shows that the identity authentication is successful. The driver face image in step S101 is collected through the above-mentioned face recognition function camera. The driver face authentication image matched by the current online car-hailing is recorded in the online car-hailing supervision platform. By comparing the driver face image with the driver face authentication image recorded in the online car-hailing supervision platform, if the two are matched successfully, the online car-hailing supervision platform will output the identity authentication result of successful identity authentication.

[0102] Moreover, the front camera will perform face recognition every certain period of time during the vehicle driving process. Once it is found that the driver is not the matched driver of the vehicle, the platform will immediately stop the order of the vehicle and transfer it to the manual for further understanding of the specific situation.

[0103] Among them, if the current online car-hailing has a quarrel with the passenger during the driving process, it will seriously affect the safety of the current online car-hailing during the driving process, so the driving state type of the current online car-hailing is judged in step S103. The driving state type refers to the state type of the current online car-hailing during the driving process, i.e. the state type of taking orders or not taking orders. The above-mentioned driving state type is analyzed and judged on the basis that the current online car-hailing is in the driving process.

[0104] Further, if the current ride-hailing vehicle is in the order-accepting passenger-carrying state, i.e., the order-accepting passenger-carrying state in step S104, the face feature data corresponding to the driver face image and the passenger face image in the current ride-hailing vehicle is obtained. The type of the current ride-hailing vehicle can be obtained through the ride-hailing vehicle supervision platform. Each order of the current ride-hailing vehicle is obtained through the order allocation of the ride-hailing vehicle supervision platform.

[0105] It should be noted that each seat area in the current ride-hailing vehicle is configured with a face recognition camera. The driver face image and the passenger face image in step S104 can be obtained through the face recognition camera. The driver face image refers to the driver face image of the driver seat area in the current ride-hailing vehicle captured by the face recognition camera. The passenger face image refers to the passenger face image of the passenger seat area in the current ride-hailing vehicle captured by the face recognition camera. The face feature data corresponding to the driver and the passenger can be obtained by further identifying and analyzing the driver face image and the passenger face image. The face feature data refers to the feature data of the eyes, pupils, mouth, nose, and other facial organs obtained by identifying and analyzing the face image.

[0106] In order to further determine whether the driver and the passenger in the current ride-hailing vehicle are in the communication state during driving, the corresponding mouth area feature in step S105 is obtained according to the face feature data of the driver and the passenger. The mouth area feature refers to the feature of the mouth area in the face feature data, such as the action, frequency, and size of the mouth opening and closing.

[0107] It should be noted that the opening and closing action of the mouth of the driver and the passenger can be used to determine whether the driver and the passenger are communicating with the outside world at the same time. Further, the voice information of the driver and the passenger communicating with the outside world can be obtained through the recording device in the current ride-hailing vehicle. Further, the voice information can be analyzed to determine whether the driver and the passenger are in the communication state.

[0108] In order to further determine whether the driver and the passenger are arguing, the communication voice of the driver and the passenger in step S107 is obtained on the basis that the driver and the passenger in the current ride-hailing vehicle are in the communication state. The communication voice refers to the voice information of the driver and the passenger communicating with each other. The communication voice can be collected and recorded through the recording device in the current ride-hailing vehicle.

[0109] Further, the communication voice is recognized to determine whether there is a abusive vocabulary in the communication voice. The abusive vocabulary refers to negative vocabulary in language communication, such as insulting, mocking, and degrading. If the communication voice contains the corresponding abusive vocabulary, it is determined that the current online car driver and the passenger have an argument. Then, the voice segment corresponding to the abusive vocabulary in the communication voice is obtained. The voice recognition system corresponding to the recording device in the current online car stores a vocabulary library corresponding to the abusive vocabulary. The communication voice is recognized by the voice recognition system to determine whether the corresponding vocabulary in the communication voice matches the vocabulary in the vocabulary library. If the matching is successful, it means that the communication voice contains the corresponding abusive vocabulary. If the matching is not successful, it means that the communication voice does not contain the corresponding abusive vocabulary.

[0110] It should be noted that, in order to further regulate the professional quality of the current online car driver and the related responsibility identification, the online car supervision platform records the voice backup data corresponding to the driver of the current online car. Further, the target voice corresponding to the first abusive vocabulary in the voice segment in step S110 is selected, and it is determined whether the target voice matches the voice stored in the voice backup data corresponding to the voice of the current online car driver. If it matches, it is determined that the current online car driver is the inducer of the argument, and a first abnormal state is sent to the current online car supervision platform, so as to hold the current online car driver responsible. The first abnormal state refers to an abnormal state that reminds the online car supervision platform that the current online car is likely to induce a safety accident during driving.

[0111] Further, if there is no abusive vocabulary in the communication voice, the corresponding facial expression data of the current online car driver in step S113 is further obtained by the face recognition function camera. The facial expression data refers to the expression information reflected in the driver's face image, such as full concentration, great surprise, and dull eyes. The facial expression data can be distinguished by the representation of facial organs, for example, by raising eyebrows, enlarging eyes, and opening mouth, it can be determined that the facial expression is in a great surprise state.

[0112] It should be noted that, by comparing the facial expression data corresponding to the current online car driver with the corresponding fatigue driving characteristics, it can be determined whether the current online car driver is in a fatigue driving state. Fatigue driving refers to the phenomenon that the driver's driving skills decline after a long time of continuous driving, and the driver's sleep quality is poor or insufficient. Long time driving can easily cause fatigue. Driving fatigue can affect the driver's attention, feeling, perception, thinking, judgment, will, decision, and movement. The fatigue driving characteristics refer to the related characteristics of fatigue driving shown by the facial expression of the driver.

[0113] For example, through face feature detection of the camera system by the face recognition function, the mouth region feature in the current online car driver's facial expression data is extracted, and whether there is a yawning behavior is determined according to the degree of mouth opening and closing, or the eye region in the facial expression data is located, the eye opening and closing degree in the region is calculated, and whether there is a closing eye behavior is determined according to the threshold control, and the yawning behavior and the closing eye behavior are both fatigue driving features.

[0114] If the above facial expression features meet the fatigue driving features, it is determined that the second abnormal state is sent to the online car supervision platform, and the second abnormal state refers to an abnormal state that reminds the online car supervision platform that the current online car driving process is easy to induce safety accidents.

[0115] The online car supervision platform analyzes the received first abnormal state or second abnormal state and sends a corresponding warning prompt to the driver of the current online car. The online car supervision platform can also send a remote control instruction to the current online car according to the actual situation to control the current online car to stop, stop and other operations.

[0116] The online car face recognition method provided in the embodiment is based on the successful identity authentication of the current online car driver, and the face images of the current online car driver and the passenger are obtained respectively, so as to generate corresponding mouth region features by identifying the corresponding facial feature data in the above face images, to determine whether the driver and the passenger in the current online car are in a communication state, and further to obtain the communication voice of the driver and the passenger in the communication state, to determine whether the driver and the passenger have a quarrel in the vehicle driving process, and to determine whether the target voice tone corresponding to the first appearance of the abusive vocabulary selected from the voice segment meets the driver authentication voice tone, so as to facilitate the responsibility identification of the quarrel inducer affecting the driving safety, and further determine the first abnormal state. If there is no related abusive vocabulary in the communication voice, the corresponding facial expression features in the driver's face image are obtained to determine whether they meet the fatigue driving features, and the situation meeting the fatigue driving features is identified as the second abnormal state. The corresponding warning prompt is sent to the driver according to the determination result of the actual abnormal state, so as to improve the supervision intensity of the online car in the driving process.

[0117] In one of the embodiments of the present embodiment, the mouth region feature includes the driver's mouth opening and closing period and the passenger's mouth opening and closing period, and step S106 of determining whether the driver and the passenger are in a communication state according to the mouth region feature includes the following steps:

[0118] S201. Determine whether there is a continuous alternating overlap between the driver's mouth opening and closing period and the passenger's mouth opening and closing period.

[0119] S202. If there is continuous alternation and overlap between the driver's mouth opening and closing period and the passenger's mouth opening and closing period, the first voice information in the driver's mouth opening and closing period and the second language information in the passenger's mouth opening and closing period are obtained respectively;

[0120] S203. Determine whether there is corresponding response information in the first voice information and the second language information;

[0121] S204. If there is corresponding response information in the first voice information and the second language information, it is determined that the driver and the passenger are in a communication state;

[0122] S205. If there is no corresponding response information in the first voice information and the second language information, it is determined that the driver and the passenger are not in a communication state.

[0123] The driver's mouth opening and closing period in step S201 refers to the time period corresponding to the continuous opening and closing action of the current online taxi driver's mouth, and the passenger's mouth opening and closing period refers to the time period corresponding to the continuous opening and closing action of the current online taxi passenger's mouth.

[0124] In actual application, if there is communication between the current online taxi driver and the passenger, there must be continuous alternation or overlap between the driver's mouth opening and closing period and the passenger's mouth opening and closing period. Generally, the communication between people has an alternating connection or overlap before and after the mouth opening and closing action. The alternating connection or overlap before and after the mouth opening and closing action can be achieved by selecting the mouth camera picture of the driver and the passenger when the driver and the passenger have mouth opening and closing action at the same time through the face recognition function camera. If the communication mode of the two people is a question and answer mode, this mode corresponds to the continuous alternation between the mouth opening and closing periods of the two people. If the communication mode of the two people is a simultaneous response mode, this mode corresponds to the overlap or step overlap between the mouth opening and closing periods of the two people, for example, two people singing a song.

[0125] In order to further determine whether the current online taxi driver and the passenger are in a communication state, the driver's corresponding first language information and the passenger's corresponding second language information are obtained through the recording device in the current online taxi. The first language information refers to the language voice information sent by the current online taxi driver to the outside world in the corresponding mouth opening and closing period, and the second language information refers to the language voice information sent by the current online taxi passenger to the outside world in the corresponding mouth opening and closing period.

[0126] By analyzing whether there is corresponding response information in the above first language information and second language information, it is determined whether the current online taxi driver and the passenger are in a communication state. The response information refers to the language voice response information sent by the current online taxi driver and the passenger to the outside world in the corresponding mouth opening and closing period.

[0127] For example, the acquired first language information is "Where is your hometown", and the acquired second language information is "My hometown is in the northeast". It can be determined that the first language information and the second language information have corresponding relevance response information, and it is further determined that the current driver and the passenger of the online car-hailing are in a communication state.

[0128] For another example, the acquired first language information is "Where is your hometown", and the acquired second language information is "My wife is Xiaoli". It can be determined that the first language information and the second language information do not have corresponding relevance response information, and it is further determined that the current driver and the passenger of the online car-hailing are not in a communication state.

[0129] The online car-hailing face recognition method provided in the embodiment determines whether there is corresponding continuous and alternating overlap between the mouth opening and closing periods of the driver and the passenger, and determines whether there is corresponding response information in the first language information corresponding to the driver and the second language information corresponding to the passenger in the above period, thereby improving the accuracy of determining whether the driver and the passenger are in a communication state.

[0130] In one of the embodiments of the present embodiment, as shown in FIG. 2, after it is determined that the driver and the passenger are in a communication state in step S204, the method further includes the following steps: Figure 3

[0131] S301. Acquire the communication duration corresponding to the communication state;

[0132] S302. Determine whether the communication duration exceeds the dangerous driving communication duration;

[0133] S303. If the communication duration exceeds the dangerous driving communication duration, identify the face feature data corresponding to the driver, and generate the corresponding eye region feature;

[0134] S304. Acquire the corresponding gaze point according to the eye region feature;

[0135] S305. Determine whether the gaze point exceeds the road area where the current online car-hailing is driving;

[0136] S306. If the gaze point exceeds the road area where the current online car-hailing is driving, acquire the corresponding gaze duration;

[0137] S307. Determine whether the gaze duration reaches the abnormal gaze shift duration;

[0138] S308. If the gaze duration reaches the abnormal gaze shift duration, issue a corresponding attention distraction warning information.

[0139] ​In actual application, the long communication time between the current online car driver and the passenger may cause the driver to be distracted during driving, thereby affecting the driving safety of the current online car.

[0140] In order to reduce the occurrence of the above situation, it is further determined whether the communication time exceeds the minimum communication safety time specified by the road driving safety specification, that is, the dangerous driving communication time in steps S302 to S303, and it is further determined whether the communication time between the current online car driver and the passenger affects the safe driving of the driver. The communication time can be timed by a timer for voice information collected by a recording device in the current online car.

[0141] For example, the dangerous driving communication time is 2 minutes, which means that the communication time between the current online car driver and the passenger exceeds 2 minutes, which violates the road driving safety specification and may affect the safe driving of the current online car driver. It is known that the communication time is 5 minutes, so it can be determined that the communication time has exceeded the above-mentioned corresponding dangerous driving communication time.

[0142] In order to further analyze the attention of the current online car driver, the corresponding eye region features are generated by recognizing the corresponding facial feature data of the driver, and the eye region features refer to eye rotation actions, pupil line of sight angles and other eye features. According to the eye region features, the fixation point of the eyes of the current online car driver can be obtained, and the fixation point refers to the fixation area of the pupil of the current online car driver.

[0143] If the fixation point of the current online car driver exceeds the road area driven by the current online car, it means that the driver is focusing on other things at this time, and the road area driven by the current online car refers to the drivable range area of the current online car on the current driving road. The road driven by the current online car can be obtained by satellite positioning, and the specific road area that the vehicle can pass through can be obtained according to the high-precision map.

[0144] It should be noted that during the driving of the current online car, if the driver has behaviors such as playing mobile phones, looking for things, turning back to chat and other behaviors that are not focused, it may cause the driver to have improper operation during driving, and cause related safety vehicle safety accidents. By detecting whether the fixation point of the current online car driver exceeds the road area driven by the current online car, it is determined whether the driver is distracted, mainly using the principle of human eye imaging, real-time detecting the pupil point of the driver to calculate the line of sight angle, and combining the normal driving line of sight angle value of the driver to determine whether the above line of sight angle deviates from the road area driven by the current online car.

[0145] In order to further detect the distraction behavior of the current online car-hailing driver, it is determined whether the gaze duration of the driver's gaze point deviating from the road area reaches the abnormal gaze shift duration. The abnormal gaze shift duration refers to the minimum gaze point deviation duration specified by the vehicle driving safety standard, that is, the minimum gaze duration of the driver's gaze point in the non-road area. If the driver's gaze duration is equal to or exceeds the abnormal gaze shift duration, the probability of vehicle safety accidents will increase.

[0146] For example, the abnormal gaze shift duration is 3 seconds, and the current online car-hailing driver's gaze point deviates from the road area for 10 seconds. It can be determined that the gaze duration reaches the corresponding abnormal gaze shift duration, and the online car-hailing supervision platform sends a corresponding attention distraction warning prompt to the current online car-hailing driver.

[0147] For example, the abnormal gaze shift duration is 3 seconds, and the current online car-hailing driver's gaze point deviates from the road area for 1.5 seconds. It can be determined that the gaze duration does not reach the corresponding abnormal gaze shift duration, and the online car-hailing supervision platform continues to monitor and analyze the driver's gaze point.

[0148] The online car-hailing face recognition method provided by the embodiment further combines whether the driver's gaze point deviates from the road area where the online car-hailing vehicle is currently driving, and whether the gaze duration corresponding to the gaze point deviating from the road area exceeds the corresponding abnormal gaze shift duration, to determine whether the current online car-hailing driver has inattention during driving, thereby improving the supervision of the driver's driving attention during driving.

[0149] In one embodiment of the present embodiment, as shown in Figure 4 After step S308, that is, if the gaze duration reaches the abnormal gaze shift duration, the corresponding attention distraction warning prompt is sent, the method further includes the following steps:

[0150] S401. Determine whether the gaze point shifts to the road area within the preset gaze correction duration;

[0151] S402. If the gaze point does not shift to the road area within the preset gaze correction duration, switch the driving mode of the current online car-hailing vehicle to the safe automatic driving mode, and obtain the real-time positioning of the current online car-hailing vehicle.

[0152] In order to continue to supervise the subsequent action of the current online car-hailing driver after the attention is not concentrated, the preset gaze correction time length in step S401 is set, which refers to the maximum correction time length that the current online car-hailing driver is allowed to transfer the gaze point to the road area where the current online car-hailing travels from the time when the gaze point of the current online car-hailing driver deviates from the road area.

[0153] For example, the preset gaze correction time length is 2 seconds, the starting time when the gaze point of the current online car-hailing driver deviates from the road area is 18:00:00 seconds obtained by the timer of the face recognition function camera, when the time becomes 18:00:05 seconds, the face recognition function camera recognizes that the gaze point of the current online car-hailing driver has not been transferred to the road area, it can be determined that the gaze point of the current online car-hailing driver has not been transferred to the road area within the preset gaze correction time length, and the driving mode of the current online car-hailing is switched to the safe automatic driving mode.

[0154] The driving mode refers to the driving control mode of the current online car-hailing, and the safe automatic driving mode refers to the safe automatic driving mode prepared when the gaze point of the driver is not in the road area and is not transferred to the road area within the preset gaze correction time length. According to the safe automatic driving mode, the current online car-hailing stops recognizing the driving control instruction issued by the driver, directly drives the current online car-hailing according to the driving control instruction in the safe automatic driving mode, and obtains the real-time positioning of the current online car-hailing, so as to track the real-time position of the current online car-hailing by the online car-hailing supervision platform. The safe automatic driving mode is a driving mode that meets the road driving safety by combining the road area and the road conditions where the current online car-hailing travels.

[0155] For example, the preset gaze correction time length is 2 seconds, the starting time when the gaze point of the current online car-hailing driver deviates from the road area is 18:00:00 seconds obtained by the timer of the face recognition function camera, when the time becomes 18:00:01 seconds, the face recognition function camera recognizes that the gaze point of the current online car-hailing driver has been transferred to the road area, it can be determined that the gaze point of the current online car-hailing driver has been transferred to the road area within the preset gaze correction time length, and the system continues to monitor the gaze point of the current online car-hailing driver.

[0156] The online car-hailing face recognition method provided by the embodiment can switch the driving mode of the current online car-hailing to the safe automatic driving mode if the gaze point of the current online car-hailing driver is not transferred to the road area within the preset gaze correction time length specified by the driving safety specification, and obtain the real-time positioning of the current online car-hailing at this time, thereby improving the safety of vehicle driving when the current online car-hailing driver is not concentrated.

[0157] In one of the embodiments of the present embodiment, as shown in Figure 5As shown, after determining that the driver and the passenger are not in the communication state in step S205, the following steps are further included:

[0158] S501. Obtain road condition information corresponding to a road on which the current online car hailing vehicle is driving;

[0159] S502. Obtain a vehicle driving action corresponding to the current online car hailing vehicle;

[0160] S503. Determine whether the vehicle driving action conforms to a road condition driving safety standard corresponding to the road condition information;

[0161] S504. If the vehicle driving action does not conform to the road condition driving safety standard corresponding to the road condition information, generate a corresponding safe driving control instruction according to the road condition driving safety standard to control the driving action of the current online car hailing vehicle.

[0162] In actual application, if the driver and the passenger of the current online car hailing vehicle are not in the communication state, the driver may drive while talking on the phone or driving while communicating with people on both sides of the road. The above situations divert the driver's attention while driving and increase the probability of vehicle safety accidents.

[0163] Therefore, in order to reduce the occurrence of the above situations, road condition information corresponding to a road on which the current online car hailing vehicle is driving is obtained. The road condition information refers to road condition information of a road section on which the current online car hailing vehicle is driving. The road condition information can be obtained by a related driving record shooting device such as a driving recorder arranged on the current online car hailing vehicle. The driving record shooting device can shoot environmental information around the current online car hailing vehicle, and the environmental information around the current online car hailing vehicle includes the road condition information.

[0164] Further, a vehicle driving action of the driver of the current online car hailing vehicle is obtained. The vehicle driving action refers to a driving control action of the driver of the current online car hailing vehicle during driving. The vehicle driving action can be obtained by a vehicle control instruction recording system. The vehicle control instruction recording system can record a control instruction issued by the driver of the current online car hailing vehicle during driving in real time. According to the control instruction, the vehicle driving action of the driver of the current online car hailing vehicle can be inferred.

[0165] The road condition driving safety standard in step S503 refers to a safe driving control standard corresponding to the road condition information of the road on which the current online car hailing vehicle is driving.

[0166] For example, according to the road condition information, it can be obtained that there is only one driving road in front of the current online car, and the driving road extends to the right side of the current online car, at this time the corresponding road condition driving safety standard is light accelerator and the steering wheel turns to the right side, at this time the vehicle control instruction displayed by the vehicle control instruction recording system is the accelerator control instruction and the vehicle straight driving control instruction, it can be determined that the vehicle driving action of the current online car driver does not conform to the road condition information corresponding to the road condition driving safety standard, and further according to the road condition driving safety standard, the corresponding safe driving control instruction is generated to control the driving action of the current online car.

[0167] For example, the vehicle control instruction displayed by the vehicle control instruction recording system is the accelerator control instruction and the vehicle right turn control instruction, it can be determined that the vehicle driving action of the current online car driver conforms to the road condition information corresponding to the road condition driving safety standard, then the vehicle control instruction recording system continues to record the control instruction of the current online car driver.

[0168] The safe driving control instruction refers to the vehicle driving control instruction generated according to the safe driving standard corresponding to the road condition information in the road condition driving safety standard when the vehicle driving action of the current online car driver does not conform to the road condition information corresponding to the road condition driving safety standard.

[0169] For example, according to the above, the corresponding safe driving control instruction generated according to the road condition driving safety standard is the vehicle driving control instruction of reducing the speed and turning the vehicle to the right side, further according to the vehicle driving control instruction of reducing the speed to control the current online car to slow down, and according to the vehicle driving control instruction of turning the vehicle to the right side to control the current online car to turn right.

[0170] The online car face recognition method provided by the embodiment combines the road condition information of the road driven by the current online car and the safe driving control instruction corresponding to the road condition driving safety standard to control the driving action of the current online car, thereby reducing the occurrence of related driving safety accidents caused by the current online car driver being distracted during driving due to long-time communication with the outside world.

[0171] In one of the embodiments of the present embodiment, as shown in Figure 6 After step S108, that is, judging whether there is abusive language in the communication voice, the following steps are further included:

[0172] S601. Obtain the corresponding facial expression features according to the driver's face image;

[0173] S602. Judge whether the facial expression features conform to the fatigue driving features;

[0174] S603. If the facial expression features conform to the fatigue driving features, obtain and control the current online car to drive to a safe parking area according to the prepared emergency control instruction, and stop identifying the driving control instruction corresponding to the current online car.

[0175] In actual application, the current online car hailing driver and passenger will quarrel in the driving process, which will distract the driver's attention and affect the driving safety. If the driver is in a state of fatigue driving and other control force decline at this time, the probability of current online car hailing safety accident will be further increased.

[0176] Therefore, in order to reduce the occurrence of the above situation, the corresponding facial features are obtained according to the driver's face image, and it is judged whether the facial features meet the corresponding fatigue driving features. The fatigue driving features refer to the facial features of the current online car hailing driver meeting the judgment standard of fatigue driving.

[0177] It should be noted that the facial features are combined with the fatigue driving features, and the facial features of the driver are intelligently analyzed by the face recognition function camera arranged in the current online car hailing, so as to further judge whether the driver is in a fatigue driving state.

[0178] For example, the fatigue driving features are to judge whether there is a yawn behavior by the degree of mouth opening and closing, and to judge whether there is a closing eye behavior by the degree of eye opening and closing.

[0179] Further, if the facial features of the current online car hailing driver are yawning, it can be determined that the facial features of the driver at this time meet the fatigue driving features described above, and a first warning signal is generated; if the facial features of the current online car hailing driver are closing eyes, a second warning signal is generated; if the facial features of the current online car hailing driver are yawning and closing eyes, a third warning signal is generated. The third warning signal is more urgent than the second warning signal, and the second warning signal is more urgent than the first warning signal, so it is convenient to make accurate analysis combined with the specific situation of the current online car hailing driver.

[0180] For example, if the facial features of the current online car hailing driver do not meet the corresponding fatigue driving features, the face recognition function camera arranged in the current online car hailing continues to identify and record the facial features of the driver.

[0181] It should be noted that if the facial features of the current online car hailing driver meet the corresponding fatigue driving features, the online car hailing supervision platform issues corresponding preliminary emergency control instructions to the current online car hailing. The preliminary emergency control instruction refers to the standby emergency vehicle control instruction preset when the driver's vehicle control ability decreases due to fatigue driving and other reasons. The current online car hailing can drive to the safe parking area on the roadside according to the preliminary emergency control instruction, and the current online car hailing stops identifying the driving control instruction corresponding to the current online car hailing, thereby reducing the occurrence of vehicle safety accidents caused by subsequent improper operation of the current online car hailing driver.

[0182] The face recognition method of the online car-hailing provided by the embodiment further determines whether the driver's manner feature conforms to the fatigue driving feature based on the current online car-hailing driver and passenger quarreling, and if the driver's manner feature conforms to the fatigue driving feature, controls the current online car-hailing to drive to a safe parking area through the prepared emergency control instruction, thereby improving the safety supervision strength of the current online car-hailing in the driving process.

[0183] In one of the implementation manners of the embodiment, as shown in Figure 7 the step S112, if the target tone conforms to the driver authentication tone, it is determined as the first abnormal state, and the following steps are further included:

[0184] S701. Obtain the steering wheel monitoring video corresponding to the current online car-hailing;

[0185] S702. Determine whether the driver's handheld operation action appears in the steering wheel monitoring video;

[0186] S703. If the driver's handheld operation action does not appear in the steering wheel monitoring video, obtain and brake the current online car-hailing according to the emergency braking instruction.

[0187] In actual application, if the current online car-hailing driver and passenger quarrel, and the driver is the inducer of this quarrel, it means that the driver's emotion is in an abnormal unstable state at this time, which seriously affects the driving safety of the vehicle.

[0188] Therefore, in order to improve the driving safety of the current online car-hailing, the steering wheel monitoring video in the current online car-hailing is further obtained through the monitoring device in the current online car-hailing, and whether the driver's handheld operation action on the steering wheel appears is further determined according to the steering wheel monitoring video. If the handheld operation action on the steering wheel does not appear, the current online car-hailing is obtained and braked according to the emergency braking instruction.

[0189] It should be noted that the emergency braking instruction refers to the braking instruction sent by the online car-hailing supervision platform to the current online car-hailing when the driver and the passenger are in a quarrel state and do not perform the handheld operation action on the steering wheel. The current online car-hailing realizes emergency braking according to the emergency braking instruction.

[0190] Furthermore, if the driver's handheld operation action on the steering wheel appears in the steering wheel monitoring video, the monitoring device in the current online car-hailing continues to shoot and record the steering wheel area of the current online car-hailing.

[0191] The face recognition method of the online car-hailing provided by the embodiment further determines whether the driver's handheld operation action appears in the steering wheel monitoring video based on the current online car-hailing driver and passenger quarreling, and if the driver's handheld operation action does not appear, brakes the current online car-hailing according to the emergency braking instruction, thereby improving the safety supervision strength of the current online car-hailing in the driving process.

[0192] In one of the implementation manners of the embodiment, as shown in Figure 8 the step S103, if the identity authentication result is successful, after judging whether the current online car is in the passenger-carrying state, the following steps are further included:

[0193] S801. If the current online car is in the un-ordered driving state, the corresponding in-vehicle monitoring picture is obtained;

[0194] S802. According to the in-vehicle monitoring picture, it is judged whether there is a passenger in the current online car;

[0195] S803. If there is a passenger in the current online car, the face image corresponding to the passenger is obtained and recognized, and the corresponding recognition authentication result is generated;

[0196] S804. If the recognition authentication result shows that the identity authentication is abnormal, the current online car is controlled to be turned off according to the turn-off control instruction.

[0197] In actual application, the un-ordered driving state means that the current online car has no ordered task but is in the driving state.

[0198] In order to reduce the current online car to take private jobs, the in-vehicle monitoring picture is further obtained through the monitoring device in the current online car, and it is judged whether there is a passenger in the car according to the in-vehicle monitoring picture. If there is a passenger, the face image corresponding to the passenger is recognized through the face recognition function camera set in the car, and the corresponding recognition authentication result is generated.

[0199] It should be noted that the online car supervision platform stores the identity authentication information of all online car drivers and other online car staff. When the recognition authentication result shows that the identity authentication is abnormal, it means that the passenger is not an online car driver or other online car staff. The online car supervision platform sends a turn-off control instruction to the current online car to control the current online car to be turned off according to the recognition authentication result of the identity authentication.

[0200] If the recognition authentication result shows that the identity authentication is normal, it means that the passenger is an online car driver or other online car staff, and the monitoring device in the current online car continues to take the in-vehicle monitoring picture of the current online car.

[0201] The online car face recognition method provided by the embodiment judges whether the current online car in the un-ordered driving state carries a passenger according to the in-vehicle monitoring picture, further performs face identity authentication on the passenger, and controls the current online car to be turned off according to the turn-off control instruction if the recognition authentication result shows that the identity authentication is abnormal, thereby reducing the occurrence of the current online car driver taking private jobs.

[0202] The embodiment of the present application further discloses a terminal device, comprising a memory, a processor, and computer instructions stored in the memory and capable of running on the processor, wherein the processor executes the computer instructions to adopt any of the online car-hailing face recognition methods in the above embodiments.

[0203] The terminal device can be a computer device such as a desktop computer, a notebook computer or a cloud server, and the terminal device comprises but is not limited to a processor and a memory, for example, the terminal device can further comprise an input / output device, a network access device and a bus.

[0204] The processor can be a central processing unit (CPU), and of course, according to the actual use, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), ready programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. can also be used, and the general-purpose processor can be a microprocessor or any conventional processor, etc. The present application does not make any limitation.

[0205] The memory can be an internal storage unit of the terminal device, for example, a hard disk or a memory of the terminal device, or an external storage device of the terminal device, for example, a plug-in hard disk, a smart memory card (SMC), a secure digital card (SD) or a flash memory card (FC) equipped on the terminal device, and the memory can also be a combination of the internal storage unit and the external storage device of the terminal device. The memory is used to store computer instructions and other instructions and data required by the terminal device, and can also be used to temporarily store data that has been output or will be output, and the present application does not make any limitation.

[0206] The terminal device stores any of the online car-hailing face recognition methods in the above embodiments in the memory of the terminal device, and loads and executes the method on the processor of the terminal device, which is convenient to use.

[0207] The embodiment of the present application further discloses a computer readable storage medium, and the computer readable storage medium stores computer instructions, wherein the computer instructions are executed by the processor to adopt any of the online car-hailing face recognition methods in the above embodiments.

[0208] The computer instructions can be stored in a computer readable medium, the computer instructions include computer instruction codes, the computer instruction codes can be in a source code form, an object code form, an executable file or some intermediate form, etc., and the computer readable medium includes any entity or device capable of carrying the computer instruction codes, recording media, U disks, mobile hard disks, magnetic disks, optical disks, computer memories, read-only memories (ROM), random access memories (RAM), electric carrier signals, telecommunication signals and software distribution media, etc. It should be noted that the computer readable medium includes but is not limited to the above components.

[0209] The computer readable storage medium stores any one of the face recognition methods for the online car hailing in the above embodiments in the computer readable storage medium, and is loaded and executed on the processor to facilitate storage and application of the method.

[0210] The above are preferred embodiments of the application, which do not limit the protection scope of the application, therefore: any equivalent changes made on the structure, shape and principle of the application should be covered within the protection scope of the application.

Claims

1. A facial recognition method for ride-hailing services, characterized in that, Includes the following steps: Acquire and recognize the facial image of the current ride-hailing driver, and generate the corresponding identity authentication result; Determine whether the identity authentication result is successful; If the identity authentication result is successful, then determine the current driving status type of the ride-hailing vehicle; If the current ride-hailing vehicle is in the order-taking and passenger-carrying state, then obtain the facial feature data corresponding to the driver's face image and the passenger's face image of the current ride-hailing vehicle respectively. Identify the facial feature data and generate corresponding mouth region features; Determine whether the driver and passenger are communicating based on the characteristics of the mouth area; If the driver and the passenger are communicating, then the corresponding voice communication is acquired. Determine whether the spoken communication contains abusive language; If the abusive words are present in the spoken communication, then the speech segment corresponding to the abusive words is obtained; Select the target timbre corresponding to the first occurrence of the abusive word in the audio segment; Identify and determine whether the target timbre matches the driver certification timbre; If the target timbre matches the driver's certified timbre, it is determined to be the first abnormal state; If the abusive words are not present in the spoken communication, then the corresponding facial features are obtained from the driver's facial image; Determine whether the facial expression characteristics match those of fatigued driving; If the facial expression characteristics match the fatigue driving characteristics, then it is determined to be the second abnormal state; Based on the first abnormal state or the second abnormal state, issue a corresponding warning message; The mouth region features include the driver's mouth opening and closing cycle and the passenger's mouth opening and closing cycle, and the step of determining whether the driver and passenger are communicating based on the mouth region features includes the following steps: Determine whether there is a continuous alternation or overlap between the driver's mouth opening and closing cycle and the passenger's mouth opening and closing cycle; If there is a continuous alternation and overlap between the driver's mouth opening and closing cycle and the passenger's mouth opening and closing cycle, then the first speech information within the driver's mouth opening and closing cycle and the second speech information within the passenger's mouth opening and closing cycle are obtained respectively. Determine whether there is corresponding response information in the first voice information and the second language information; If there is a corresponding response information in the first voice information and the second language information, then it is determined that the driver and the passenger are in the communication state; If there is no corresponding response information in the first voice information and the second language information, it is determined that the driver and the passenger are not in the communication state. If a corresponding response exists in the first voice information and the second language information, the process further includes the following steps after determining that the driver and the passenger are in a communication state: Obtain the communication duration corresponding to the communication state; Determine whether the duration of the communication exceeds the duration of communication during dangerous driving; If the duration of the communication exceeds the duration of the dangerous driving communication, then the facial feature data corresponding to the driver is identified, and the corresponding eye region features are generated. Based on the characteristics of the eye region, obtain the corresponding fixation point; Determine whether the gaze point exceeds the road area currently being traveled by the ride-hailing vehicle; If the gaze point is outside the road area where the current ride-hailing vehicle is traveling, the corresponding gaze duration is obtained; Determine whether the fixation duration has reached the abnormal fixation shift duration; If the gaze duration reaches the abnormal gaze shift duration, a corresponding attention distraction warning message will be issued; After issuing a corresponding attention distraction warning if the gaze duration reaches the abnormal gaze shift duration, the process further includes the following steps: Determine whether the gaze point shifts to the road area within a preset gaze correction duration; If the gaze point does not shift to the road area within the preset gaze correction time, the driving mode of the current ride-hailing vehicle is switched to the safe autonomous driving mode, and the real-time location of the current ride-hailing vehicle is obtained.

2. The method for facial recognition in ride-hailing services according to claim 1, characterized in that, After determining that the driver and the passenger are not in the communication state if there is no corresponding response information in the first voice information and the second language information, the following steps are also included: Obtain the road condition information corresponding to the road currently being traveled by the ride-hailing vehicle; Obtain the vehicle driving action corresponding to the current ride-hailing vehicle; Determine whether the vehicle's driving actions comply with the road condition driving safety standards corresponding to the road condition information; If the vehicle's driving actions do not conform to the road condition driving safety standards corresponding to the road condition information, then a corresponding safe driving control command is generated based on the road condition driving safety standards to control the current ride-hailing vehicle's driving actions.

3. The method for facial recognition in ride-hailing services according to claim 1, characterized in that, After determining whether abusive language is present in the spoken communication, the following steps are also included: The corresponding facial expression features are obtained from the driver's facial image; Determine whether the facial expression characteristics match the characteristics of fatigued driving; If the facial expression characteristics match the fatigue driving characteristics, then the current ride-hailing vehicle is controlled to drive to a safe parking area according to the prepared emergency control command, and the identification of the driving control command corresponding to the current ride-hailing vehicle is stopped.

4. The method for facial recognition in ride-hailing services according to claim 1, characterized in that, After determining that the first abnormal state is determined if the target timbre matches the driver certification timbre, the following steps are also included: Obtain the steering wheel monitoring video corresponding to the current ride-hailing vehicle; Determine whether the driver's hand-held control action appears in the steering wheel monitoring video; If the driver's hand-held control action is not seen in the steering wheel monitoring video, the current ride-hailing vehicle is braked according to the emergency braking command.

5. The method for facial recognition in ride-hailing services according to claim 1, characterized in that, If the identity authentication result is successful, then after determining whether the current ride-hailing vehicle is carrying passengers, the following steps are also included: If the current ride-hailing vehicle is in a state of not accepting orders, then the corresponding in-vehicle monitoring footage is obtained; Determine whether the passenger is currently in the ride-hailing vehicle based on the in-vehicle monitoring footage; If the passenger is present in the current ride-hailing vehicle, then the facial image corresponding to the passenger is acquired and identified, and a corresponding identification and authentication result is generated; If the identification and authentication result shows an identity authentication error, then the current ride-hailing vehicle is shut down according to the engine shutdown control command.

6. A terminal device, comprising a memory and a processor, characterized in that, The memory stores computer instructions that can run on the processor. When the processor loads and executes the computer instructions, it employs a ride-hailing facial recognition method as described in any one of claims 1 to 5.

7. A computer-readable storage medium storing computer instructions, characterized in that, When the computer instructions are loaded and executed by the processor, a ride-hailing facial recognition method as described in any one of claims 1 to 5 is employed.

Images