A method for real-time driving risk identification and risk level determination for passenger transport drivers

Abstract

The present invention provides a method for real-time driving risk identification and risk level determination of passenger transport drivers, and the method is to use a risk rule table for risk identification and risk grading, wherein the risk rule table includes three parts: abnormal driving behavior description, risk type identification, and risk level determination. First, the platform receives the video stream and vehicle positioning information uploaded by the vehicle-mounted monitoring equipment; secondly, the corresponding information is structured to form a set of driver dynamic behavior characteristics; again, the driver's dynamic behavior characteristic set is identified according to the risk type, and a risk type information record is formed; then, the risk type information is determined in real time according to the risk level determination rules, and the risk level information is recorded; finally, the driver's driving behavior is intervened in real time according to the risk level information. The method of the present invention can provide enterprises with a means of real-time intervention in the driver's driving behavior.

Description

Technical Field

[0001] The present invention relates to the field of road passenger transport enterprises, and in particular to a method for real-time driving risk identification and risk level determination for passenger transport drivers. Background Art

[0002] In the process of driving, the existing technology generally involves installing a video device in the vehicle to allow the safety manager to manage the driver's driving risk through a software platform by means of real-time polling and random sampling. The method is to manually check whether the driver has dangerous driving behavior by viewing the video online in real time, and then manually determine the risk level.

[0003] The inventors of this application found that the existing methods have at least the following problems during the implementation of the present invention:

[0004] Safety officers can only identify risks for a single vehicle, which makes it impossible to effectively manage the safety risks of multiple vehicles in real time. Video polling and reviewing lead to high workload for safety officers and great limitations in identifying risks.

[0005] It can be seen from this that the existing methods have problems such as incomplete manual risk identification, poor effect of risk level determination based on manual experience, and untimely risk intervention. Summary of the invention

[0006] The present invention provides a method for real-time driving risk identification and risk level determination for passenger transport drivers, which is used to solve or at least partially solve the technical problem of poor identification effect existing in the prior art.

[0007] In order to solve the above technical problems, the present invention provides a method for real-time driving risk identification and risk level determination for passenger transport drivers, comprising:

[0008] Risk identification and risk level determination are performed according to the risk rule table, where the risk rule table includes abnormal driving behavior description, risk type identification, and risk level determination. Abnormal driving behavior: used to determine the type of abnormal driving behavior of the driver. Risk type identification is used by the platform to identify the risk type and record the risk type information. Risk level determination determines the risk level based on the number of times the vehicle risk type occurs on that day.

[0009] First, the platform receives the video stream and vehicle positioning information uploaded by the on-board monitoring equipment; secondly, the platform performs structured processing on the corresponding information to form a set of dynamic behavior characteristics of the driver; thirdly, the platform identifies the risk type of the driver's dynamic behavior characteristics set based on the risk type identification to form a risk type information record; then, the risk type information is determined in real time based on the risk level determination, and the risk level information is recorded; finally, real-time intervention measures are taken on the driver's driving behavior based on the risk level information.

[0010] In one implementation, the platform receives video streams and vehicle location information uploaded by the vehicle monitoring device, including:

[0011] The video uploaded by the on-board equipment of the road transport enterprise's vehicles is parsed through JT / T 1078-2016 "Video Communication Protocol of Satellite Positioning System for Road Transport Vehicles", and the video information obtained includes: alarm sign, start time, end time, file size, longitude, latitude, and alarm duration. The alarm signs include fatigue driving, making or receiving phone calls, and smoking. The start time format is YY-MM-DD-hh-mm-ss, and the end time format is YY-MM-DD-hh-mm-ss.

[0012] JT / T 808-2019 "Terminal Communication Protocol and Data Format of Road Transport Vehicle Satellite Positioning System" parses the location information and alarms uploaded by the on-board equipment of road transport companies' vehicles, and obtains information including: time, status, longitude, latitude, instantaneous speed, alarm sign, and alarm duration. The alarm signs include speeding alarm, fatigue driving alarm, dangerous driving behavior alarm, and overtime driving alarm.

[0013] In one embodiment, the vehicle-mounted device data is structured and processed to form a driver dynamic behavior feature set, including:

[0014] The platform integrates the received video stream and vehicle positioning information to form a set of driver dynamic behavior features, whose vectors include: alarm information (T_AM), alarm location information details (T_AD), and abnormal driving behavior (T_ADB). The alarm information (T_AM) records the alarm information uploaded by the vehicle-mounted device, including the fields: alarm flag (T_AM_A), start time (T_AM_t), end time (T_AM_s), and alarm duration (T_AM_o). The alarm location information details (T_AD) records the location details within the time period when the alarm information occurs, including the fields: alarm flag (T_AD_o), time (T_AD_t), longitude (T_AD_l), latitude (T_AD_la), and instantaneous speed (T_AD_iv).

[0015] In one embodiment, vehicle risk type identification is performed on the driver's dynamic behavior feature set according to the risk type identification, and a risk type information record (T_RTI) is formed. The risk type information record includes: risk type (T_RTI_rt), start time (T_RTI_t), end time (T_RTI_s), and risk occurrence duration (T_RTI_o), including:

[0016] Perform real-time identification on T_AM in the driver's dynamic behavior feature set according to the risk type identification rule of OD in the risk rule table. When T_AM_A = OD and T_AM_o > 30 seconds, in T_AD, within the time period of T_AM_t < T_AD_t < T_AM_s, compare each instantaneous speed with the speeding threshold. If all instantaneous speeds are greater than the speeding threshold, then identify the driving speeding risk and record it in the risk type information.

[0017] Perform real-time identification on T_AM in the driver's dynamic behavior feature set according to the risk type identification rule of DT in the risk rule table. When T_AM_A is DT and T_AM_o > 14400 seconds, then identify the driving overtime risk and record it in the risk type information.

[0018] Perform real-time identification on T_AM in the driver's dynamic behavior feature set according to the risk type identification rule of CU in the risk rule table. When T_AM_A is CU and T_AM_o > 30 seconds, then identify the risk of making or receiving calls and record it in the risk type information.

[0019] Perform real-time identification on T_AM in the driver's dynamic behavior feature set according to the risk type identification rule of DS in the risk rule table. When T_AM_A is DS and T_AM_o > 60 seconds, then identify the smoking risk and record it in the risk type information.

[0020] Perform real-time identification on T_AM in the driver's dynamic behavior feature set according to the risk type identification rule of FD in the risk rule table. When T_AM_A is FD and T_AM_o > 40 seconds, then identify the fatigue driving risk and record it in the risk type information.

[0021] In one embodiment, perform real-time risk level determination on T_RTI according to the risk level judgment rule, and record the risk level information, including:

[0022] According to the risk level determination rules of OD in the risk rule table, the OD of the day in T_RTI_rt is counted in real time to obtain the number of OD of the day: count(OD). When 1≤count(OD)<5, the OD is recorded as a slight risk, corresponding to level 4; when 5≤count(OD)<10, the speeding is recorded as a general risk, corresponding to level 3; when 10≤count(OD)<15, the speeding is recorded as a greater risk, corresponding to level 2; when count(OD)≥15, the speeding is recorded as a major risk, corresponding to level 1.

[0023] According to the risk level determination rules of DT in the risk rule table, the DT of the day in T_RTI_rt is counted in real time to obtain the number of DT of the day: count(DT). When count(DT)=1, the record is that the timeout is a greater risk, and the corresponding level is level 2; when count(DT)≥2, the record is that the timeout is a major risk, and the corresponding level is level 1.

[0024] According to the risk level determination rule of CU in the constructed risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of CU of the day: count(CU). If count(CU)=1, the record of making or receiving calls is a general risk, and the corresponding level is level three; if 2≤count(CU)≤3; the record of making or receiving calls is a large risk, and the corresponding level is level two; if count(CU)>3, the record of making or receiving calls is a major risk, and the corresponding level is level one.

[0025] According to the risk level determination rule of DS in the constructed risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of DS of the day: count(DS). If count(DS)=1, smoking is recorded as general risk, and the corresponding level is level three; if 2≤count(DS)≤3; smoking is recorded as a greater risk, and the corresponding level is level two; if count(DS)>3, smoking is recorded as a major risk, and the corresponding level is level one.

[0026] According to the risk level judgment rule of FD in the constructed risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of FDs on the day: count(FD). If count(FD)=1, fatigue driving is recorded as a greater risk, and the corresponding level is level 2; if count(FD)>1; fatigue driving is recorded as a major risk, and the corresponding level is level 1.

[0027] In one implementation, the safety officer intervenes in the driver's driving behavior in real time based on the risk level information, including:

[0028] When the risk type level reaches general risk, remote voice reminders will be given through the platform; when the risk type level reaches a higher risk level, it will be reported to the company leaders and key monitoring will be carried out to prevent risk escalation; when the risk type level reaches a major risk, it will be reported to the company leaders and the driver will be asked to park in a safe parking area for remote psychological counseling.

[0029] The above one or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:

[0030] The present invention provides a method for real-time driving risk identification and risk level determination for passenger transport drivers, and risk identification is performed according to a risk rule table, wherein the risk rule table includes three parts: abnormal driving behavior description, risk type identification, and risk level determination. Abnormal driving behavior: used to determine the type of abnormal driving behavior of the driver. Risk type identification is used to identify the risk type by the platform. Risk level determination determines the risk level based on the number of times the risk type occurs. First, the video stream and vehicle positioning information uploaded by the vehicle-mounted monitoring device are received; secondly, the corresponding information is structured to form a set of driver dynamic behavior characteristics; again, the vehicle risk type is identified for the driver's dynamic behavior feature set according to the risk type identification to form a risk type information record; then, the risk type information record is subjected to real-time risk level determination according to the risk level determination, and the risk level information is recorded; finally, real-time intervention measures are taken on the driver's driving behavior according to the risk level information.

[0031] This method establishes scientific risk identification rules and risk level determination rules for risk behaviors that exist during vehicle driving. The abnormal driving behavior of the driver is uploaded to the platform through the on-board equipment, and the corresponding behavior characteristics are obtained by analyzing and processing the video stream and location information. Then, it is matched with the risk rule table to identify various dangerous driving behaviors, which solves the problem of relying on the subjective experience of safety officers for risk identification and improves the effect of risk identification. In addition, the record of the risk generation process is retained, which can solve the problem of daily safety officers manually identifying risks by viewing videos one by one through the platform, greatly improving the identification efficiency. In the past, safety officers could only inspect the safety risks of a single vehicle. Through risk classification information, safety officers can understand the risks of all vehicles in real time and dynamically. According to the different risk levels, timely reminders can be given in a targeted manner, and vehicle risk behaviors can be warned in a comprehensive and real-time manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 The present invention is a flowchart of a method for real-time driving risk identification and risk level determination based on video for road passenger transport enterprises in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The present invention proposes a video-based driving risk identification and risk grading method for road passenger transport enterprises. It uses a pre-constructed risk rule table to match the features in the driver's dynamic behavior feature set, and automatically identifies and determines the risk level of dangerous driving behaviors in video streams and position information. This can solve the problems of safety officers' subjective identification of vehicle safety risks while the vehicle is driving, the problem of the heavy workload of safety officers in risk identification, the problem that safety officers cannot comprehensively identify all vehicle risks in real time and dynamically, and the problem of cumbersome daily hidden danger management work for safety officers.

[0035] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] The embodiment of the present invention provides a method for real-time driving risk identification and risk level determination for a passenger driver, which includes:

[0037] Receive video streams and vehicle positioning information uploaded by vehicle-mounted monitoring equipment;

[0038] Structural processing of video stream and vehicle positioning information to form a set of driver dynamic behavior features;

[0039] The formed set of driver dynamic behavior characteristics is matched with a pre-constructed risk rule table, wherein the risk rule table includes abnormal driving behavior description information, risk type identification rules, and risk level determination rules. The matching process includes: identifying the risk type of the driver's dynamic behavior characteristics set according to the risk type identification rules to form a risk type information record; determining the risk level of the risk type information in real time according to the risk level determination rules and recording the risk level information; and finally taking real-time intervention measures on the driving behavior according to the risk level information.

[0040] Among them, the risk rule table includes abnormal driving behavior description information, risk type identification rules, and risk level determination rules. Abnormal driving behavior: used to determine the type of abnormal driving behavior of the driver. Risk type identification is used for the platform to identify the risk type. Risk level determination determines the risk level based on the number of times the risk type occurs.

[0041] The above steps are implemented by the platform through computer processes.

[0042] Specifically, in the prior art, during the driving process of a vehicle, the danger identification of the driver's driving behavior is generally carried out by installing a video device in the vehicle, and the safety management personnel manually verify whether the driver has dangerous driving behavior by playing the video through the software platform, and then make manual paper records. The prior art mainly has the following shortcomings and technical problems:

[0043] 1. Safety officers can only identify risks for a single vehicle, resulting in the inability to achieve real-time and effective risk management for multiple vehicles.

[0044] 2. The use of video polling and review results in high workload for safety officers and great limitations in identifying risks.

[0045] 3. It is impossible to conduct comprehensive inspection and identification of risk management.

[0046] 4. Safety risks cannot be intervened in real time, which has an adverse impact on the company's production safety risk management.

[0047] Based on this, the present invention proposes a method for real-time driving risk identification and risk level determination for passenger drivers. The method uses the risk type determination rules in the risk rule table to compare the processed data of the video stream and location information uploaded by the vehicle-mounted equipment to automatically identify dangerous behaviors. On the one hand, it improves the accuracy of identification, and on the other hand, it also improves the efficiency of identification.

[0048] In one implementation, the risk types in the risk rule table include speeding, driving overtime, making phone calls while driving, smoking while driving, and fatigue driving.

[0049] Specifically, corresponding to the risk type, the risk rule table gives the risk type judgment rules corresponding to the risk type. For example, for the risk of speeding, the risk type judgment rule can be determined according to the duration of speeding. For example, in Table 4, when the alarm mark is OD, the alarm duration is >30 seconds, and within the alarm start time and the alarm end time, each instantaneous speed in the location information is greater than the speeding threshold, then the driving speeding is recorded once. Similarly, the risk type judgment rule for making phone calls can be determined according to the number of times the phone calls are made, such as the risk level judgment rules for making phone calls in Table 5. The risk level judgment rule judges the risk level according to the number of times the risk type occurs. For example, for the risk of speeding, if the number of speeding in a single day is between 1 and 5 times, it is a minor risk, 5 to 10 times is a general risk, 10 to 15 times is a greater risk, and so on.

[0050] In the specific implementation process, the risk rule table contains three parts: abnormal driving behavior description, risk type identification, and risk level judgment. The abnormal driving behavior description is shown in Table 1, the risk type identification is shown in Table 2, and the risk level judgment is shown in Table 3.

[0051] Table 1 Description of abnormal driving behavior

[0052]

[0053] Table 2 Risk type identification

[0054]

[0055] Table 3 Risk level determination

[0056]

[0057] In one embodiment, the vehicle-mounted device data is structured and processed to form a driver dynamic behavior feature set, including:

[0058] The platform integrates the received video stream and vehicle positioning information to form a set of driver dynamic behavior features, which includes: alarm information (T_AM) and alarm location information details (T_AD). The alarm information (T_AM) records the alarm information uploaded by the vehicle-mounted device, including the fields: alarm flag (T_AM_A), start time (T_AM_t), end time (T_AM_s), and alarm duration (T_AM_o), as shown in Table 4. The alarm location information details (T_AD) records the location details within the time period when the alarm information occurs, including the fields: alarm flag (T_AD_o), time (T_AD_t), longitude (T_AD_l), latitude (T_AD_la), and instantaneous speed (T_AD_iv), as shown in Table 5.

[0059] Table 4 Driver dynamic behavior feature set - warning information

[0060] Field Type describe Alarm sign See Table 1 for details of abnormal driving behavior description Start time The format is YY-MM-DD-hh-mm-ss End time The format is YY-MM-DD-hh-mm-ss Duration The format is seconds

[0061] Table 5 Driver dynamic behavior feature set - alarm details information

[0062]

[0063]

[0064] In one embodiment, the vehicle risk type is identified based on the driver's dynamic behavior feature set according to the risk type identification to form a risk type information record, which includes: risk type, start time, end time, and risk occurrence duration, including:

[0065] According to the risk type identification rules for speeding in risk type identification in Table 4, speeding identification is performed in real time. If the alarm sign in Table 1 is a speeding alarm and the alarm duration value is greater than 30 seconds, the alarm location detail record is judged according to Table 3. If the instantaneous speed within the corresponding start time and end time range is greater than the threshold for risk category identification of speeding, the speeding risk is identified.

[0066] Specifically, the vehicle-mounted monitoring device records the location information, and the vehicle-mounted device uploads the location information. If the alarm mark in the driver's dynamic behavior feature set alarm information is a speeding alarm, and the alarm duration is >30 seconds, the instantaneous speed data within the start time and end time range of the alarm is queried in the details of the driver's dynamic behavior feature set alarm location information, and compared with the speeding threshold of driving speeding in the risk category identification. If the instantaneous speed is greater than the speeding threshold, the speeding risk is identified and recorded in the risk category information.

[0067] In one embodiment, risk category recognition is used to identify risk behaviors of a driver's dynamic behavior feature set, including:

[0068] Overtime identification is performed according to the risk type identification rules for driving overtime in the risk category identification in Table 4. If the alarm sign in Table 1 is an overtime alarm and the alarm duration meets the risk type judgment rules for driving overtime, the risk of driving overtime is identified.

[0069] Specifically, the vehicle-mounted monitoring device records the location information, and the vehicle-mounted device uploads the location information. If the alarm mark in the alarm information in the driver's dynamic behavior feature set is an overtime alarm, and the alarm duration is greater than 30 seconds, the driving overtime risk is identified and recorded in the risk category information.

[0070] In one embodiment, risk category recognition is used to identify risk behaviors of a driver's dynamic behavior feature set, including:

[0071] According to the risk type identification rules for making and receiving phone calls in the risk category identification in Table 4, the making and receiving phone calls are identified. If the alarm mark in Table 1 is making and receiving phone calls and the alarm duration meets the risk type judgment rules for making and receiving phone calls, the risk of making and receiving phone calls is identified.

[0072] Specifically, for example, if the alarm mark in the alarm information in the driver's dynamic behavior feature set is a phone call alarm, and the alarm duration exceeds 30 seconds, a phone call risk record is performed and recorded in the risk category information.

[0073] In one embodiment, risk category recognition is used to identify risk behaviors of a driver's dynamic behavior feature set, including:

[0074] Smoking is identified according to the risk type identification rules for smoking in the risk category identification in Table 4. If the alarm sign in Table 1 is smoking and the alarm duration meets the risk type judgment rules for smoking, the smoking risk is identified and recorded in the risk category information.

[0075] Specifically, when the alarm mark in the alarm information in the driver's dynamic behavior feature set is smoking, whether the alarm duration exceeds 60 seconds, if exceeded, the risk of smoking is identified and recorded in the risk category information.

[0076] In one embodiment, risk category recognition is used to identify risk behaviors of a driver's dynamic behavior feature set, including:

[0077] Fatigue driving identification is performed according to the risk type identification rules for fatigue driving in the risk category identification in Table 4. If the alarm sign in Table 1 is fatigue driving and the alarm duration meets the risk type judgment rules for fatigue driving, the fatigue driving risk is identified and recorded in the risk category information.

[0078] Specifically, when the alarm sign is fatigue driving in the driver's dynamic behavior feature set, whether the alarm duration exceeds 40 seconds, if exceeded, the risk of fatigue driving is identified and recorded in the risk category information.

[0079] In the prior art, there is a method for identifying and classifying driver driving behaviors based on driving feature groups, which mainly analyzes the driving speed of a large amount of historical driving trajectory data of the driver to obtain a corresponding driving speed model and then infer the driving behavior. The main method used by the present invention for identifying dangerous driving behaviors of drivers is to analyze the driver's behavioral actions and identify the driver's dangerous behavior through the driver's behavior and duration. Compared with the method for identifying and classifying driver driving behaviors based on driving feature groups, the present invention is more intuitive in identifying driving behaviors, and the identified dangerous driving behaviors are photographed, recorded by video, located, and the time of occurrence of dangerous driving behaviors are obtained, so that the judgment is more accurate, thereby reducing the timeliness of safety management personnel's judgment on dangerous driving behaviors and improving the accuracy of dangerous driving behaviors.

[0080] There are also methods for identifying dangerous driving behaviors and predicting dangerous events in the prior art, specifically: collecting the dynamic driving data of the vehicle at the current moment; obtaining the dangerous driving characteristic value at the current moment based on the dynamic driving data; obtaining the historical driving data of the vehicle in a specified historical time period; obtaining the driving behavior characteristic value based on the historical driving data; comparing the dangerous driving characteristic value with the driving behavior characteristic value to obtain a comparison value, and determining whether the vehicle is currently in a dangerous driving behavior based on the comparison value and a preset threshold value, and predicting the probability of a dangerous event. The disadvantage of this method is that if the calculation of the historical driving behavior characteristic value of the vehicle is inaccurate, obtaining the current driving behavior characteristic value for comparison will result in very inaccurate identification of dangerous behaviors, and whether it is a dangerous driving behavior is inferred by the speed change of the vehicle position data. This method is not intuitive enough to judge dangerous driving behaviors, and is prone to misjudgment. Subsequent historical backtracking cannot reveal the true driving situation. Compared with this method, the present invention is more intuitive in identifying driving behavior, and can take photos, record videos, locate the location, and obtain the time of occurrence of dangerous driving behavior for the identified dangerous driving behavior, making the judgment more accurate, thereby reducing the time efficiency of safety management personnel in judging dangerous driving behavior and improving the accuracy of dangerous driving behavior.

[0081] In one embodiment, the method further comprises:

[0082] The rules for determining the risk level of speeding are as follows: if 1≤the number of speeding violations per day<5, speeding is judged to be a slight risk, corresponding to level four; if 5≤the number of speeding violations per day<10, speeding is judged to be a general risk, corresponding to level three; if 10≤the number of speeding violations per day<15, speeding is judged to be a greater risk, corresponding to level two; if the number of speeding violations per day is ≥15, speeding is judged to be a major risk, corresponding to level one.

[0083] The risk level determination rules for overtime driving are as follows: if the number of overtime driving times in a single day = 1, the overtime is determined to be a greater risk, and the corresponding level is level 2; if the number of speeding times in a single day is ≥ 2, the overtime is determined to be a major risk, and the corresponding level is level 1.

[0084] The rules for determining the risk level of making phone calls while driving are as follows: if the number of phone calls made while driving in a single day = 1, then making phone calls while driving is a general risk, and the corresponding level is level three; if 2≤the number of phone calls made while driving in a single day≤3; then making phone calls while driving is determined to be a greater risk, and the corresponding level is level two; if the number of phone calls made while driving in a single day>3, then speeding is determined to be a major risk, and the corresponding level is level one.

[0085] The risk level determination rules for smoking while driving: if the number of times you smoke while driving in a single day = 1, smoking while driving is a general risk, and the corresponding level is level three; if 2≤the number of times you smoke while driving in a single day≤3; then smoking while driving is determined to be a greater risk, and the corresponding level is level two; if the number of times you smoke while driving in a single day>3, then speeding is determined to be a major risk, and the corresponding level is level one.

[0086] The risk level determination rules for fatigue driving: If the number of fatigue driving incidents in a single day = 1, fatigue driving is determined to be a greater risk, and the corresponding level is Level 2; if the number of fatigue driving incidents in a single day > 1, speeding is determined to be a major risk, and the corresponding level is Level 1.

[0087] In one embodiment, the method further comprises:

[0088] Based on the risk level information, the risk type and the latest level of the risk type generated by the vehicle in the current time period are obtained in real time. When the risk type level reaches the general risk, a remote voice reminder is given through the platform; when the risk type level reaches the higher risk level, it is reported to the company leadership and key monitoring is carried out to prevent risk escalation; when the risk type level reaches the major risk, it is reported to the company leadership and the driver is asked to park in a safe parking area for remote psychological counseling.

[0089] During the specific implementation process, through the risk level information data generated, the safety officer can check the time, location, risk type and risk level information of the risk according to the risk level information. In the past, the safety officer could only inspect the safety risks of a single vehicle. Through the risk level information, the safety officer can understand the risks of all vehicles in real time and dynamically. According to the different risk levels, timely reminders can be given in a targeted manner, and vehicle risk behaviors can be warned in a comprehensive and real-time manner.

[0090] In general, the advantages and beneficial technical effects of the present invention are as follows:

[0091] 1. Application of risk rule table: establish a risk database and corresponding risk level for risk behaviors during vehicle driving, which includes a set of dynamic behavior characteristics of drivers, risk type identification, and risk level determination. This solves the problem of relying on safety personnel to manually judge risks, resulting in inaccurate risk identification and insufficient facts.

[0092] 2. Automatically identify dangerous behaviors of drivers during driving: By receiving video streams and location information uploaded by on-board devices inside the vehicle, the video streams and location information are structured to form a set of dynamic behavior characteristics of the driver, identify various dangerous driving behaviors, and change the previous safety officer risk identification method of obtaining risky behaviors through manually recorded videos.

[0093] 3. Automatically generate risk level information: Automatically identify the risk level of risk type identification information through risk level determination rules, automatically generate risk level information of different levels and risks, and solve the problem of security officers polling videos to find risky behaviors.

[0094] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for real-time driving risk identification and risk level determination for passenger transport drivers, characterized in that: Including: Receiving the video stream and vehicle positioning information uploaded by the in-vehicle monitoring device; Performing structured processing on the video stream and vehicle positioning information to form a set of driver dynamic behavior characteristics; Matching the formed set of driver dynamic behavior characteristics with a pre-constructed risk rule table. The risk rule table includes abnormal driving behavior description information, risk type identification rules, and risk level determination rules. The matching process includes: identifying the risk type of the set of driver dynamic behavior characteristics according to the risk type identification rules to form a risk type information record; determining the risk level of the risk type information in real time according to the risk level determination rules and recording the risk level information; finally, taking real-time intervention measures for the driving behavior according to the risk level information; Among them, performing structured processing on the video stream and vehicle positioning information to form a set of driver dynamic behavior characteristics includes: Integrating the received video stream and vehicle positioning information to form a set of driver dynamic behavior characteristics. The set of driver dynamic behavior characteristics includes T_AM and T_AD. T_AM is the alarm information, and T_AD is the detailed alarm location information. T_AM is used to record the alarm information uploaded by the in-vehicle device, including fields: T_AM_A, T_AM_t, T_AM_s, and T_AM_o. T_AM_A is the alarm flag, T_AM_t is the start time, T_AM_s is the end time, and T_AM_o is the alarm duration; the detailed alarm location information is used to record the location details during the time period when the alarm information occurs, including fields: T_AD_o, T_AD_t, T_AD_l, T_AD_la, and T_AD_iv. T_AD_o is the location alarm flag, T_AD_t is the time, T_AD_l is the longitude, T_AD_la is the latitude, and T_AD_iv is the instantaneous speed; The rule types identified by the risk type identification rules in the risk rule table include speeding, driving overtime, making calls, smoking, and fatigue driving, and the corresponding fields are OD, DT, CU, DS, and FD; Identifying the risk type of the set of driver dynamic behavior characteristics according to the risk type identification rules includes: Performing real-time identification on T_AM in the set of driver dynamic behavior characteristics according to the risk type identification rule of OD in the risk rule table. When T_AM_A = OD and T_AM_o > 30 seconds, in T_AD, within the time period of T_AM_t < T_AD_t < T_AM_s, compare each instantaneous speed with the speeding threshold. If all instantaneous speeds are greater than the speeding threshold, then identify the speeding risk and record it into the risk type information; Performing real-time identification on T_AM in the set of driver dynamic behavior characteristics according to the risk type identification rule of DT in the risk rule table. When T_AM_A is DT and T_AM_o > 14400 seconds, then identify the driving overtime risk and record it into the risk type information; According to the risk type identification rule CU in the risk rule table, T_AM in the driver's dynamic behavior feature set is identified in real time. When T_AM_A is CU and T_AM_o>30 seconds, the risk of making or receiving a phone call is identified and the risk type information is recorded; According to the risk type identification rule of DS in the risk rule table, T_AM in the driver's dynamic behavior feature set is identified in real time. When T_AM_A is DS and T_AM_o>60 seconds, the smoking risk is identified and the risk type information is recorded; According to the risk type identification rule FD in the risk rule table, T_AM in the driver's dynamic behavior feature set is identified in real time. When T_AM_A is FD and T_AM_o>40 seconds, fatigue driving risk is identified and risk type information is recorded; The risk type information record formed is T_RTI, which includes: T_RTI_rt, T_RTI_t, T_RTI_s and T_RTI_o, T_RTI_rt is the risk type, T_RTI_t is the start time, T_RTI_s is the end time, and T_RTI_o is the duration of risk occurrence; Risk level determination rules are used to classify risk types, including: According to the risk level determination rule of OD in the risk rule table, the OD of the day in T_RTI_rt is counted in real time to obtain the number of OD of the day: count(OD). When 1≤count(OD)<5, the OD is recorded as a slight risk, and the corresponding level is level 4; when 5≤count(OD)<10, the speeding is recorded as a general risk, and the corresponding level is level 3; when 10≤count(OD)<15, the speeding is recorded as a greater risk, and the corresponding level is level 2; when count(OD≥15, the speeding is recorded as a major risk, and the corresponding level is level 1; According to the risk level determination rule of DT in the risk rule table, the DT of the day in T_RTI_rt is counted in real time to obtain the number of DT of the day: count(DT); when count(DT)=1, the record is that the timeout is a greater risk, and the corresponding level is level 2; when count(DT)≥2, the record is that the timeout is a major risk, and the corresponding level is level 1; According to the risk level determination rule of CU in the risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of CUs of the day: count(CU). If count(CU) = 1, the record of making or receiving calls is a general risk, and the corresponding level is level 3; if 2≤count(CU)≤3; the record of making or receiving calls is a relatively large risk, and the corresponding level is level 2; if count(CU)>3, the record of making or receiving calls is a major risk, and the corresponding level is level 1; According to the risk level determination rule of DS in the risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of DS of the day: count(DS). If count(DS)=1, the smoking is recorded as a general risk, and the corresponding level is level 3; if 2≤count(DS)≤3; the smoking is recorded as a greater risk, and the corresponding level is level 2; if count(DS)>3, the smoking is recorded as a major risk, and the corresponding level is level 1; According to the risk level determination rule of FD in the risk rule table, the CU of the day in T_RTI_rt is counted in real time to obtain the number of FDs on the day: count(FD). If count(FD)=1, fatigue driving is recorded as a greater risk, and the corresponding level is level 2; if count(FD)>1; fatigue driving is recorded as a major risk, and the corresponding level is level 1.

Images