A method and system for risk management of vehicle-based orders

Abstract

This invention relates to the technical field of safety risk supervision for ride-hailing orders, and discloses a method and system for risk management of ride-hailing orders. It uses real-time driver driving characteristic image information combined with intelligent image recognition algorithms and standard violation service attitude image information of ride-hailing drivers to perform visual detection-based precise monitoring of drivers' violation service attitudes during ride-hailing operations. It also uses real-time driver voice text information combined with intelligent search algorithms and standard violation service attitude voice text keywords to perform deep and reliable analysis of drivers' violation service attitudes during ride-hailing operations. This achieves intelligent and accurate supervision of violation service attitudes throughout the entire ride-hailing process based on visual and voice multimodal data fusion, realizing intelligent safety monitoring of ride-hailing order driving risks from multiple perspectives, including driver legitimacy, driving behavior, and service attitude.

Description

Technical Field

[0001] This invention relates to the technical field of vehicle order security risk supervision, specifically a vehicle order risk management method and system. Background Technology

[0002] Vehicle-related order risk management is a core and highly specialized topic in logistics, mobility, car rental, and fleet management. It refers to a series of management activities involving the identification, assessment, monitoring, and prevention of potential risks throughout the entire process of converting vehicle-related services (such as transportation, delivery, car rental, and ride-hailing) into standardized orders. Simply put, it ensures that every ride-hailing or car-hailing order is completed safely, on time, compliantly, and economically. Ride-hailing order risk management is particularly crucial as it involves passenger safety and property security. Current ride-hailing order risk management systems fail to provide comprehensive legal oversight of drivers throughout their journeys, nor can they intelligently monitor drivers' driving behavior and service attitudes. This not only increases the safety risks of ride-hailing travel but also reduces the passenger experience.

[0003] Chinese invention patent CN113112391B, published on July 19, 2024, discloses an order processing method and apparatus. This application, by obtaining the location attribute information of the target pick-up location corresponding to a travel order, can determine the waiting risk level of the service provider accepting the travel order at the target pick-up location. It can perform risk assessment of waiting at the target pick-up location and, if it determines that the service requesting party has not arrived at the target pick-up location within the waiting time after the start of the scheduled travel time, send a reminder message to the service provider indicating that the travel order can be cancelled. Thus, by controlling the time it takes for passengers to arrive at the target pick-up location, it can reduce the risk of illegal parking while ensuring drivers can pick up passengers smoothly, improving driver efficiency. However, the above technical solution cannot achieve full legal supervision of the driver throughout the entire travel process, nor can it achieve intelligent supervision of the driver's driving behavior and service attitude. This not only increases the safety risks of ride-hailing travel but also reduces the passenger's experience. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing ride-hailing order risk management systems, which fail to provide comprehensive legal oversight of drivers throughout their journeys and lack intelligent monitoring of their driving behavior and service attitudes, thus increasing safety risks and compromising passenger experience, the goal is to achieve dynamic and legal oversight of drivers throughout their journeys, intelligent monitoring of their driving behavior, and precise monitoring of their service attitudes. This will effectively reduce ride-hailing order safety risks and improve the safety and passenger experience of ride-hailing services.

[0005] (II) Technical Solution This invention is achieved through the following technical solution: a method for risk management of vehicle-based orders, the method comprising the following steps: The system collects ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information. Based on the ride-hailing order information, it searches and processes the facial images of legally registered drivers to obtain facial image information of registered drivers. It analyzes the legality of driver identities during ride-hailing operations to obtain legality analysis information for drivers. It identifies and processes drivers' illegal driving behaviors during ride-hailing operations to obtain illegal driving behavior identification information for drivers. Finally, it uses visual detection to identify drivers' illegal service behavior during ride-hailing operations. The system analyzes and processes information to obtain visual information on drivers' poor service attitude during ride-hailing operations; it also analyzes and processes voice information on drivers' poor service attitude during ride-hailing operations; and it assesses the driving risks of drivers during ride-hailing order execution to obtain ride-hailing order driving risk assessment information. When it is safe, the system continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed; when it is dangerous, the system constructs a summary of ride-hailing driving risks, performs ride-hailing order driving risk push operations, and continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed.

[0006] Preferably, the following steps are taken: collecting ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information; and then searching and processing the facial images of legitimate drivers registered for ride-hailing orders based on the ride-hailing order information to obtain the facial image information of drivers registered for ride-hailing orders. The system obtains ride-hailing order information confirmed by drivers online through a ride-hailing management platform and generates ride-hailing order text information. This text information includes driver identity information, vehicle information, passenger identity information, passenger origin and destination information, and ride-hailing route information. The ride-hailing management platform includes any one of Didi Chuxing, T3 Chuxing, and Cao Cao Chuxing. During the ride-hailing order execution process, the system also collects images of the driver's face and body movements via a mobile terminal camera and generates real-time driver driving characteristic images. Furthermore, during the ride-hailing order execution process, the system collects the driver's voice and text communication during driving via a mobile terminal voice module and generates real-time driver voice and text information. The mobile terminal includes any one of smartphones and in-vehicle computers. Based on the ride-hailing order text information, a search and processing of the facial images of legitimate drivers registered for ride-hailing orders is performed to obtain the facial image information of drivers registered for ride-hailing orders.

[0007] Preferably, the process of searching for and processing the facial images of legitimate drivers registered for ride-hailing services based on the ride-hailing order text information to obtain the facial image information of drivers registered for ride-hailing services includes the following steps: Obtain the ride-hailing order text information; The KD-tree nearest neighbor search algorithm is used to search for facial image information of ride-hailing drivers who have legally registered ride-hailing orders on the ride-hailing management platform based on the ride-hailing order text information, and to construct the facial image information of drivers registered for ride-hailing orders.

[0008] Preferably, the process of analyzing and processing the legality of the driver's identity during the ride-hailing process to obtain legality analysis information for the ride-hailing driver; identifying and processing the driver's illegal driving behavior during the ride-hailing process to obtain illegal driving behavior identification information for the ride-hailing driver; analyzing and processing the driver's poor service attitude during the ride-hailing process based on visual detection to obtain poor service attitude analysis information for the ride-hailing driver from a visual perspective; and analyzing and processing the driver's poor service attitude during the ride-hailing process based on voice recognition to obtain poor service attitude analysis information for the ride-hailing driver from a voice perspective includes the following steps: Based on the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order, the legality analysis of the identity of the driver during the ride-hailing process is performed to obtain legality analysis information of the ride-hailing driver. Based on the real-time driver driving feature image information and the standard illegal driving behavior image information matrix of the ride-hailing driver, the illegal driving behavior of the driver during the ride-hailing process is identified and processed to obtain the illegal driving behavior identification information of the ride-hailing driver. Based on the real-time driving feature image information of the ride-hailing driver and the standard violation service attitude image information matrix of the ride-hailing driver, the driver's violation service attitude during the ride-hailing process is analyzed and processed using visual detection to obtain the driver's visual violation service attitude analysis information; based on the real-time voice text information of the ride-hailing driver and the standard violation service attitude voice text keyword matrix of the ride-hailing driver, the driver's violation service attitude during the ride-hailing process is analyzed and processed using speech recognition to obtain the driver's voice violation service attitude analysis information.

[0009] Preferably, the process of analyzing the legality of the driver's identity during the ride-hailing process based on the real-time driver's driving feature image information and the driver's facial image information registered in the ride-hailing order, to obtain the legality analysis information of the ride-hailing driver, includes the following steps: Obtain the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order; The SURF image feature matching algorithm is used to match the real-time driving feature image information of the ride-hailing driver with the facial image information of the driver registered in the ride-hailing order. Based on the image feature matching results, legal analysis information of the ride-hailing driver object is generated. When the real-time driving feature image information of the ride-hailing driver is successfully matched with the facial image information of the driver registered for the ride-hailing order, it indicates that the current ride-hailing driver is a driver who has legally registered for the ride-hailing order. Then, the legality analysis information of the ride-hailing driver object is output as legal. If the real-time driving feature image information of the ride-hailing driver fails to match the facial image information of the driver registered in the ride-hailing order, it indicates that the current ride-hailing driver is not a driver legally registered for the ride-hailing order. In this case, the legality analysis information of the ride-hailing driver object will be output as invalid.

[0010] Preferably, the process of identifying and processing the driver's illegal driving behavior during the ride-hailing process based on the real-time driver's driving feature image information and the standard illegal driving behavior image information matrix of the ride-hailing driver includes the following steps: Establish a matrix of image information on standard driving violations by ride-hailing drivers. ,in Indicates the first The standard image information of ride-hailing drivers corresponding to different types of illegal driving behavior includes driving with eyes closed, driving while talking on the phone, driving while eating, driving with hands off the steering wheel, and driving while picking up items. The standard image information of ride-hailing drivers represents the standard image information of drivers' driving behavior during ride-hailing operations set for different types of illegal driving behavior. The real-time driving feature image information of the ride-hailing driver is compared with the standard illegal driving behavior image information matrix of the ride-hailing driver. Image information of standard illegal driving behavior of ride-hailing drivers described in the article Image feature matching is performed, and information on illegal driving behavior of ride-hailing drivers is generated based on the image feature matching results. The specific steps for generating the information on illegal driving behavior of ride-hailing drivers are as follows: Step 2221, Initialization Phase: Update the maximum number of iterations In the image information matrix of standard illegal driving behaviors of ride-hailing drivers Initialize and update driving behavior within the search space to search for the location of the female spider wasp; the formula for initializing and updating the location is: ,in Indicates the first A female spider wasp searching for driving behavior in a spatial dimension of [missing information]. The image information matrix of standard illegal driving behaviors of ride-hailing drivers The initial position in the search space; Indicates the value random function, and These represent the image information matrix of standard illegal driving behaviors of female spiders and wasps in the ride-hailing vehicles. The upper and lower boundaries of the search space; Step 2222, Search Phase: The female spider wasp searches for driving behavior at a constant step size in the image information matrix of the standard illegal driving behavior of the ride-hailing driver. Randomly search the search space to find image information of standard illegal driving behavior of the ride-hailing driver that matches the real-time driving feature image information of the ride-hailing driver. Simultaneously, the matrix of images showing the standard illegal driving behaviors of female spiders searching for driving behavior in ride-hailing vehicles is updated. The search position is determined within the search space, where the formula for updating the position of the female spider wasp during the search phase, based on driving behavior, is: ,in and These represent the search phase, number 1 and 2 respectively. After the nth iteration and the th After the nth iteration A matrix of images showing standard driving violations by female spiders and wasps in ride-hailing vehicles. The location within the search space; and These represent the search phase, number 1 and 2 respectively. After the nth iteration, the first one is randomly selected. and the A matrix of images showing standard driving violations by female spiders and wasps in ride-hailing vehicles. The location within the search space; This indicates the constant step size for searching female spider wasps based on their driving behavior. ,in and They represent the values ​​respectively. A random function;

[0011] Step 2223, Follow and Hunt Phase: Searching for female spiders by driving behavior at random steps within the image information matrix of standard illegal driving behavior of ride-hailing drivers. Randomly search the search space to find image information of standard illegal driving behavior of the ride-hailing driver that matches the real-time driving feature image information of the ride-hailing driver. The prey, the female spider wasp, searches for driving behavior by targeting the image information of the ride-hailing driver whose driving characteristics best match the image information of the real-time driver's driving characteristics, and identifies the standard illegal driving behavior image information of the ride-hailing driver. The prey direction is moved and the position is updated; the formula for updating the position of the female spider wasp during the following and hunting phases is as follows: ,in This indicates the first stage of the following and hunting phase. After the nth iteration A matrix of images showing standard driving violations by female spiders and wasps in ride-hailing vehicles. The location within the search space; This indicates the first stage of the following and hunting phase. After the next iteration, the female spider wasp searched for driving behavior in the image information matrix of the standard illegal driving behavior of the ride-hailing driver. Search the search space to find the image information of the standard illegal driving behavior of the ride-hailing driver that best matches the real-time driving feature image information of the ride-hailing driver. The optimal position; This represents the random step size used to search for female spider wasps based on their driving behavior. Indicates the value A random function; Step 2224: When the maximum number of iterations is met, output the standard illegal driving behavior image information of the ride-hailing driver that best matches the real-time driving feature image information of the ride-hailing driver. ; Step 2225: Based on the real-time driving characteristic image information of the ride-hailing driver and the standard illegal driving behavior image information of the ride-hailing driver. Image feature matching results are used to generate identification information on illegal driving behavior of ride-hailing drivers; When the real-time driving characteristic image information of the ride-hailing driver is compared with the standard illegal driving behavior image information of the ride-hailing driver. Image feature matching was successful, indicating that the current ride-hailing driver has a [missing information - likely a typo or incomplete sentence]. If a certain type of illegal driving behavior is detected, the system will output the illegal driving behavior identification information of the ride-hailing driver, indicating that an illegal driving behavior has occurred, and simultaneously output the first... Text information on various types of illegal driving behavior; When the real-time driving characteristic image information of the ride-hailing driver is compared with the standard illegal driving behavior image information of the ride-hailing driver. If no matching of image features is found, it indicates that the current ride-hailing driver has not committed any illegal driving behavior. Therefore, the output of the ride-hailing driver illegal driving behavior identification information is "no illegal driving behavior".

[0012] Preferably, the analysis of the driver's service attitude violations during ride-hailing operations is performed based on visual detection using the real-time driver driving feature image information and the standard violation service attitude image information matrix of the ride-hailing driver, to obtain the driver's visual service attitude violation analysis information; the analysis of the driver's service attitude violations during ride-hailing operations is performed based on speech recognition using the real-time driver voice text information and the standard violation service attitude voice text keyword matrix of the ride-hailing driver, to obtain the driver's voice service attitude violation analysis information, including the following steps: Establish separate image information matrices for ride-hailing drivers' standard violations and service attitudes. ,in Indicates the first The system includes standard image information of ride-hailing drivers exhibiting different service attitudes, corresponding to various types of such attitudes. These attitude types include contempt, anger, disdain, resentment, verbal abuse, and intimidation. The standard image information represents facial and body language images of drivers during ride-hailing operations, tailored to different service attitude types. A keyword matrix of voice and textual descriptions of standard service attitude violations by ride-hailing drivers is also established. ,in Indicates the first The standard voice text keywords for ride-hailing drivers corresponding to different types of service attitude violations are defined as standard voice text keywords for driver communication during ride-hailing operations, and include phrases such as "get lost," "try it," "you don't deserve it," "are you sick?" "I can't be bothered with you," "useless," and "are you tired of living?" The FLANN image feature matching algorithm is used to combine the real-time driving feature image information of the ride-hailing driver with the standard violation service attitude image information matrix of the ride-hailing driver. Image information of ride-hailing drivers' standard violations of service attitude. Perform image feature matching, and generate analysis information on the driver's visual behavior regarding service attitude violations based on the image feature matching results; When the real-time driving characteristic image information of the ride-hailing driver is being compared with the standard violation service attitude image information of the ride-hailing driver, the information is being compared. Image feature matching was successful, indicating that the current ride-hailing driver has a [missing information - likely a typo or incomplete sentence]. If a certain type of service attitude is found to be non-compliant, the analysis information on the non-compliant service attitude from the driver's visual perspective will be output as "a non-compliant service attitude exists," and the first... Textual information regarding unacceptable service attitude; When the real-time driving characteristic image information of the ride-hailing driver is being compared with the standard violation service attitude image information of the ride-hailing driver, the information is being compared. If no matching of image features is found, it indicates that the current ride-hailing driver does not have any illegal service attitude. Therefore, the output of the ride-hailing driver's visual illegal service attitude analysis information is "no illegal service attitude". The Rabin-Karp text search algorithm was used to combine the real-time voice text information of the ride-hailing driver with the keyword matrix of the driver's standard violation service attitude voice text. Keywords related to ride-hailing drivers' standards, violations, service attitude, voice text, etc. Perform voice-text information matching, and generate analysis information on the illegal service attitude of ride-hailing drivers based on the voice-text information matching results; When the ride-hailing vehicle is in motion, the real-time voice and text information of the driver is compared with the keywords of the driver's standard, non-compliant, and poor service attitude in the voice and text. Successful matching of voice and text information indicates that the current ride-hailing driver has a [number] [unclear - possibly related to a previous sentence or paragraph]. If a certain type of service attitude is found to be non-compliant, the system will output the analysis information of the non-compliant service attitude from the ride-hailing driver's voice terminal, indicating that a non-compliant service attitude exists. Simultaneously, the system will output the first... Textual information regarding unacceptable service attitude; When the ride-hailing vehicle is in motion, the real-time voice and text information of the driver is compared with the keywords of the driver's standard, non-compliant, and poor service attitude in the voice and text. If neither voice nor text information is successfully matched, it indicates that the current ride-hailing driver has not engaged in any service misconduct. Therefore, the output of the ride-hailing driver's voice terminal service misconduct analysis information is "no service misconduct."

[0013] Preferably, the process of assessing and processing the driver's driving risk during the execution of a ride-hailing order to obtain ride-hailing order driving risk assessment information includes the following steps: When it is safe, the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed; when it is dangerous, a summary of ride-hailing driving risk information is constructed, the ride-hailing order driving risk push operation is executed, and the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed. Based on the legality analysis information of the ride-hailing driver, the identification information of the ride-hailing driver's illegal driving behavior, the analysis information of the ride-hailing driver's visual illegal service attitude, and the analysis information of the ride-hailing driver's voice illegal service attitude, the driving risk assessment of the driver during the ride-hailing order execution process is carried out to obtain the ride-hailing order driving risk assessment information. When the legality analysis information of the ride-hailing driver is legal, the identification information of the ride-hailing driver's illegal driving behavior is that there is no illegal driving behavior, the analysis information of the ride-hailing driver's visual service attitude is that there is no illegal service attitude, and the analysis information of the ride-hailing driver's voice service attitude is that there is no illegal service attitude, it means that there is no driver driving risk in the current ride-hailing order during the execution of the ride-hailing order. Then, the ride-hailing order driving risk assessment information is output as safe, and the ride-hailing order risk supervision operation continues until the ride-hailing order is completed. When the legality analysis information of the ride-hailing driver is illegal, or the illegal driving behavior identification information of the ride-hailing driver indicates illegal driving behavior, or the illegal service attitude analysis information of the ride-hailing driver's visual terminal indicates illegal service attitude, or the illegal service attitude analysis information of the ride-hailing driver's voice terminal indicates illegal service attitude, it indicates that there is a driver driving risk in the current ride-hailing order during the execution of the ride-hailing order, and the ride-hailing order driving risk assessment information is output as dangerous; When the ride-hailing order driving risk assessment information is deemed dangerous, the ride-hailing order text information, the real-time driver driving feature image information, the real-time driver voice text information, the ride-hailing order registration driver facial image information, the ride-hailing driver object legality analysis information, the ride-hailing driver illegal driving behavior identification information, the ride-hailing driver visual illegal service attitude analysis information, and the ride-hailing driver voice illegal service attitude analysis information are combined and identified to construct a ride-hailing driving risk summary information; The summarized information on ride-hailing driving risks is pushed to the ride-hailing management platform via the mobile communication network, and a flashing prompt is displayed on the screen to execute the ride-hailing order driving risk push operation. After the ride-hailing order driving risk push operation is completed, the ride-hailing order risk supervision operation continues until the ride-hailing order is completed.

[0014] A ride-hailing order risk management system is provided to implement the aforementioned ride-hailing order risk management method. The system includes a ride-hailing order service information acquisition module, a ride-hailing order driving safety monitoring module, and a ride-hailing order driving safety early warning module. The ride-hailing order service information acquisition module includes a ride-hailing order information collection unit, a ride-hailing order registration driver facial image search unit, a ride-hailing real-time driver driving feature image collection unit, and a ride-hailing real-time driver voice and text information collection unit. The ride-hailing order information collection unit collects ride-hailing order text information through the ride-hailing management platform; the ride-hailing order registration driver facial image search unit, based on the ride-hailing order text information and in conjunction with the ride-hailing management platform, performs facial image search processing on legitimate drivers registered for ride-hailing orders to obtain ride-hailing order registration driver facial image information; the ride-hailing real-time driving feature image collection unit collects real-time driving feature image information of ride-hailing drivers through the camera lens of a mobile terminal; the ride-hailing real-time driving voice and text information collection unit collects real-time driving voice and text information of ride-hailing drivers through the voice module of a mobile terminal. The ride-hailing order driving safety monitoring module includes a ride-hailing driver object legality analysis unit, a ride-hailing driver standard illegal driving behavior image storage unit, a ride-hailing driver illegal driving behavior recognition unit, a ride-hailing driver standard illegal service attitude image storage unit, a ride-hailing driver visual illegal service attitude analysis unit, a ride-hailing driver standard illegal service attitude voice text keyword storage unit, and a ride-hailing driver voice illegal service attitude analysis unit; The ride-hailing driver legitimacy analysis unit analyzes the legitimacy of the driver's identity during the ride-hailing process based on the real-time driver's driving feature image information and the driver's facial image information registered in the ride-hailing order, obtaining ride-hailing driver legitimacy analysis information; the ride-hailing driver standard illegal driving behavior image storage unit stores ride-hailing driver standard illegal driving behavior image information; the ride-hailing driver illegal driving behavior identification unit identifies the driver's illegal driving behavior during the ride-hailing process based on the real-time driver's driving feature image information and the ride-hailing driver standard illegal driving behavior image information, obtaining ride-hailing driver illegal driving behavior identification information; the ride-hailing driver standard illegal service attitude image storage unit stores ride-hailing driver standard illegal service attitude. The system includes: image information; a driver's visual-end violation service attitude analysis unit, which analyzes and processes the driver's violation service attitude during ride-hailing operations based on visual detection using real-time driver driving feature image information and standard driver violation service attitude image information, to obtain driver's visual-end violation service attitude analysis information; a driver's standard violation service attitude voice text keyword storage unit, which stores driver's standard violation service attitude voice text keywords; and a driver's voice-end violation service attitude analysis unit, which analyzes and processes the driver's violation service attitude during ride-hailing operations based on speech recognition using real-time driver voice text information and standard driver violation service attitude voice text keywords, to obtain driver's voice-end violation service attitude analysis information. The ride-hailing order driving safety early warning module includes a ride-hailing order driving risk assessment unit, a ride-hailing order driving risk information construction unit, and a ride-hailing order driving risk information push unit; The ride-hailing order driving risk assessment unit performs driving risk assessment processing on the driver during the ride-hailing order execution process based on the legality analysis information of the ride-hailing driver, the identification information of the ride-hailing driver's illegal driving behavior, the analysis information of the ride-hailing driver's visual and voice-based illegal service attitude, and combined with data analysis, to obtain ride-hailing order driving risk assessment information. The ride-hailing order driving risk information construction unit constructs ride-hailing driving risk summary information based on the ride-hailing order text information, the real-time driving feature image information of the ride-hailing driver, the real-time voice text information of the ride-hailing driver, the facial image information of the driver registered for the ride-hailing order, the legality analysis information of the ride-hailing driver, the identification information of the ride-hailing driver's illegal driving behavior, the analysis information of the ride-hailing driver's visual and voice-based illegal service attitude, and combined with data processing. The ride-hailing order driving risk information push unit performs ride-hailing order driving risk push operations based on the ride-hailing driving risk summary information and in conjunction with the ride-hailing management platform and display screen.

[0015] (III) Beneficial Effects This invention provides a method and system for risk management of vehicle-related orders. It has the following beneficial effects: I. By utilizing the ride-hailing management platform, mobile terminal camera, and mobile terminal voice module, real-time and efficient acquisition of ride-hailing order text information, real-time driver driving characteristic images, and real-time driver voice text information is obtained, providing real data support for subsequent ride-hailing order driving risk supervision. Based on the ride-hailing order text information and in conjunction with the ride-hailing management platform, efficient and accurate search and processing of facial images of legally registered drivers for ride-hailing orders is achieved, improving the reliability of ride-hailing order driving risk supervision.

[0016] Second, by using real-time driver characteristic image information and facial image information of drivers registered for ride-hailing orders, combined with intelligent image search algorithms, the system can autonomously and efficiently analyze the legality of driver identities during ride-hailing operations. This enables dynamic and legal supervision of drivers throughout the entire ride-hailing process, improving the compliance supervision of ride-hailing order driving risks. Furthermore, by using real-time driver characteristic image information and combining artificial intelligence algorithms with image information of standard illegal driving behaviors of ride-hailing drivers based on big data storage, the system can intelligently identify illegal driving behaviors of drivers during ride-hailing operations, achieving intelligent supervision of illegal driving behaviors throughout the entire ride-hailing process. Finally, the system can utilize real-time driver characteristic image information combined with intelligent image recognition... The algorithm, combined with scientifically stored image information of standard violations and service attitudes of ride-hailing drivers, performs visual detection-based precise monitoring of drivers' service attitudes during ride-hailing operations, achieving scientific monitoring of service attitude violations throughout the entire ride-hailing journey. Furthermore, based on real-time driver voice and text information during ride-hailing operations, and combined with intelligent search algorithms and standard stored keywords of standard violations and service attitudes in the voice and text information, the algorithm performs deep and reliable analysis of driver service attitude violations during ride-hailing operations, achieving intelligent and accurate supervision of service attitude violations throughout the entire ride-hailing journey based on the fusion of visual and voice multimodal data. This enables intelligent and safe monitoring of ride-hailing order driving risks from multiple perspectives, including driver legitimacy, driving behavior, and service attitude.

[0017] Third, based on the legality of drivers, their driving behavior and service attitude, and combined with data analysis, a scientific and comprehensive assessment of the driving risks of ride-hailing orders is conducted to achieve efficient and safe supervision of ride-hailing order driving risks based on multi-data fusion; based on ride-hailing orders, ride-hailing driver driving characteristic information, driver legality, driving behavior and service attitude, and combined with data processing, summary information on ride-hailing driving risks is obtained in a timely and efficient manner; at the same time, in conjunction with the ride-hailing management platform and display screen, the operation of pushing ride-hailing order driving risk information is carried out safely and efficiently, so as to achieve safe and efficient feedback of ride-hailing order driving risks, improve the safety of ride-hailing travel and passenger service experience. Attached Figure Description

[0018] Figure 1 A schematic diagram of a vehicle-based order risk management system provided by the present invention; Figure 2 The flowchart illustrates a vehicle-based order risk management method provided by this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] An example of a vehicle-based order risk management method and system is as follows: Example 1: Please refer to Figures 1-2 A method for risk management of vehicle-based orders, comprising the following steps: The system collects ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information. Based on the ride-hailing order information, it searches and processes the facial images of legally registered drivers to obtain facial image information of registered drivers. It analyzes the legality of driver identities during ride-hailing operations to obtain legality analysis information for drivers. It identifies and processes drivers' illegal driving behaviors during ride-hailing operations to obtain illegal driving behavior identification information for drivers. Finally, it uses visual detection to identify drivers' illegal service behavior during ride-hailing operations. The system analyzes and processes information to obtain visual information on drivers' poor service attitude during ride-hailing operations; it also analyzes and processes voice information on drivers' poor service attitude during ride-hailing operations; and it assesses the driving risks of drivers during ride-hailing order execution to obtain ride-hailing order driving risk assessment information. When it is safe, the system continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed; when it is dangerous, the system constructs a summary of ride-hailing driving risks, performs ride-hailing order driving risk push operations, and continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed.

[0021] For further details, please refer to Figures 1-2 The process involves collecting ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information. Based on the ride-hailing order information, the facial images of legitimate drivers registered for ride-hailing orders are searched and processed to obtain the facial image information of drivers registered for ride-hailing orders, including the following steps: Step 11: Obtain ride-hailing order information confirmed by the driver online through the ride-hailing management platform, and generate ride-hailing order text information. The ride-hailing order text information includes the driver's identity information, ride-hailing vehicle information, passenger's identity information, passenger's origin and destination information, and ride-hailing route information. The ride-hailing management platform includes any one of Didi Chuxing, T3 Mobility, and Cao Cao Mobility. During the execution of the ride-hailing order, collect the driver's frontal facial and body movement images online through the mobile terminal's camera, and generate real-time driver driving characteristic image information. During the execution of the ride-hailing order, collect the driver's driving communication voice text information online through the mobile terminal's voice module, and generate real-time driver voice text information. The mobile terminal includes any one of smartphones and in-vehicle computers. Step 12: Based on the ride-hailing order text information, perform facial image search and processing on the legitimate drivers registered for ride-hailing orders to obtain facial image information of drivers registered for ride-hailing orders.

[0022] The process of searching for and processing the facial images of legitimate drivers registered for ride-hailing services based on ride-hailing order text information to obtain the facial image information of drivers registered for ride-hailing services includes the following steps: Step 121: Obtain ride-hailing order text information; Step 122: Using the KD tree nearest neighbor search algorithm, search for the facial image information of the ride-hailing driver who has legally registered the ride-hailing order on the ride-hailing management platform based on the ride-hailing order text information, and construct the facial image information of the driver registered for the ride-hailing order.

[0023] By leveraging the ride-hailing management platform, mobile terminal cameras, and mobile terminal voice modules, real-time and efficient acquisition of ride-hailing order text information, real-time driver driving characteristic images, and real-time driver voice and text information is achieved, providing real data support for subsequent ride-hailing order driving risk supervision. Based on the ride-hailing order text information and in conjunction with the ride-hailing management platform, efficient and accurate search and processing of facial images of legally registered drivers for ride-hailing orders is performed, enabling efficient and accurate search of facial images of legally registered drivers for ride-hailing services and improving the reliability of ride-hailing order driving risk supervision.

[0024] For further details, please refer to Figures 1-2The process involves three main steps: analyzing the legality of the driver's identity during ride-hailing operations to obtain driver legitimacy analysis information; identifying and processing the driver's illegal driving behavior during ride-hailing operations to obtain illegal driving behavior identification information; analyzing and processing the driver's poor service attitude during ride-hailing operations based on visual detection to obtain poor service attitude analysis information from the driver's visual perspective; and analyzing and processing the driver's poor service attitude during ride-hailing operations based on voice recognition to obtain poor service attitude analysis information from the driver's voice perspective. Step 21: Based on the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order, perform an analysis and processing of the legality of the driver's identity during the ride-hailing process to obtain legality analysis information of the ride-hailing driver. Step 22: Based on the real-time driver driving feature image information and the standard illegal driving behavior image information matrix of the ride-hailing driver, identify the illegal driving behavior of the driver during the ride-hailing process to obtain the illegal driving behavior identification information of the ride-hailing driver. Step 23: Based on the real-time driving feature image information of the ride-hailing driver and the standard violation service attitude image information matrix of the ride-hailing driver, perform visual detection to analyze and process the driver's violation service attitude during the ride-hailing process, and obtain the visual violation service attitude analysis information of the ride-hailing driver; based on the real-time voice text information of the ride-hailing driver and the standard violation service attitude voice text keyword matrix of the ride-hailing driver, perform voice recognition to analyze and process the driver's violation service attitude during the ride-hailing process, and obtain the voice violation service attitude analysis information of the ride-hailing driver.

[0025] Based on real-time driver feature image information and facial image information of drivers registered for ride-hailing orders, the legality analysis of the identity of drivers during ride-hailing operations is performed to obtain legality analysis information of ride-hailing drivers, including the following steps: Step 211: Obtain real-time driving feature image information of the ride-hailing driver and facial image information of the driver registered for the ride-hailing order; Step 212: Use the SURF image feature matching algorithm to match the real-time driving feature image information of the ride-hailing driver with the facial image information of the driver registered in the ride-hailing order, and generate legal analysis information of the ride-hailing driver object based on the image feature matching results; When the real-time driving feature image information of the ride-hailing driver is successfully matched with the facial image information of the driver registered in the ride-hailing order, it indicates that the current ride-hailing driver is a driver who has been legally registered for the ride-hailing order. In this case, the legality analysis information of the ride-hailing driver object is output as legal. If the real-time driving feature image information of the ride-hailing driver fails to match the facial image information of the driver registered in the ride-hailing order, it means that the current ride-hailing driver is not a driver legally registered for the ride-hailing order. In this case, the legality analysis information of the ride-hailing driver will be output as "illegal".

[0026] Based on the real-time driving characteristic image information of ride-hailing drivers and the standard illegal driving behavior image information matrix of ride-hailing drivers, the illegal driving behavior of drivers during the ride-hailing process is identified and processed to obtain the illegal driving behavior identification information of ride-hailing drivers, including the following steps: Step 221: Establish a matrix of image information on standard illegal driving behaviors of ride-hailing drivers. ,in Indicates the first The standard image information of ride-hailing drivers corresponding to different types of illegal driving behavior includes driving with eyes closed, driving while talking on the phone, driving while eating, driving with hands off the steering wheel, and driving while picking up items. The standard image information of ride-hailing drivers represents the standard image information of drivers' driving behavior during ride-hailing operations, set for different types of illegal driving behavior. Step 222: Combine the real-time driving characteristic image information of ride-hailing drivers with the standard illegal driving behavior image information matrix of ride-hailing drivers. Image information of standard illegal driving behavior of ride-hailing drivers The process involves performing image feature matching and generating identification information for illegal driving behavior of ride-hailing drivers based on the matching results. The specific steps for generating this information are as follows: Step 2221, Initialization Phase: Update the maximum number of iterations Image information matrix of standard illegal driving behaviors of ride-hailing drivers Initialize and update driving behavior within the search space to search for the location of the female spider wasp; the formula for initializing and updating the location is: ,in Indicates the first A female spider wasp searching for driving behavior in a spatial dimension of [missing information]. Image information matrix of standard illegal driving behaviors of ride-hailing drivers The initial position in the search space; Indicates the value random function, and These represent the image information matrix of standard illegal driving behaviors of female spiders and wasps in ride-hailing vehicles. The upper and lower boundaries of the search space; Step 2222, Search Phase: Searching for images of female spiders and wasps exhibiting constant step lengths while driving in a ride-hailing vehicle, based on a matrix of images depicting standard driver violations. Randomly search the search space to find standard illegal driving behavior images of ride-hailing drivers that match the real-time driving characteristic image information of ride-hailing drivers. Simultaneously, the matrix of image information on standard driving violations by female spiders and wasps in ride-hailing vehicles was updated. The search position is determined within the search space, where the formula for updating the position of the female spider wasp during the search phase, based on driving behavior, is: ,in and These represent the search phase, number 1 and 2 respectively. After the nth iteration and the th After the nth iteration A matrix of images showing driving behavior, including female spiders and wasps, and standard driving violations by drivers in ride-hailing services. The location within the search space; and These represent the search phase, number 1 and 2 respectively. After the nth iteration, the first one is randomly selected. and the A matrix of images showing driving behavior, including female spiders and wasps, and standard driving violations by drivers in ride-hailing services. The location within the search space; This indicates the constant step size for searching female spider wasps based on their driving behavior. ,in and They represent the values ​​respectively. A random function; Step 2223, Follow and Hunt Phase: Searching for female spiders using random step sizes to observe standard illegal driving behavior image information matrix of ride-hailing drivers. Randomly search the search space to find standard illegal driving behavior images of ride-hailing drivers that match the real-time driving characteristic image information of ride-hailing drivers. The search for prey involves female spiders and wasps that target images of drivers whose driving characteristics best match those of ride-hailing drivers, identifying images of drivers exhibiting standard violations of driving rules. The prey direction is moved and the position is updated; the formula for updating the position of the female spider wasp during the following and hunting phases is as follows: ,in This indicates the first stage of the following and hunting phase. After the nth iteration A matrix of images showing driving behavior, including female spiders and wasps, and standard driving violations by drivers in ride-hailing services. The location within the search space; This indicates the first stage of the following and hunting phase. After the next iteration, the image information matrix of female spider wasps searching for standard illegal driving behaviors of drivers in ride-hailing vehicles was obtained. Search the search space to find the standard illegal driving behavior images of ride-hailing drivers that best match the real-time driving characteristic image information of ride-hailing drivers. The optimal position; This represents the random step size used to search for female spider wasps based on their driving behavior. Indicates the value A random function; Step 2224: When the maximum number of iterations is met, output the standard illegal driving behavior image information of the ride-hailing driver that best matches the real-time driving feature image information of the ride-hailing driver. ; Step 2225: Based on the real-time driving characteristic image information of the ride-hailing driver and the standard illegal driving behavior image information of the ride-hailing driver. Image feature matching results are used to generate identification information on illegal driving behavior of ride-hailing drivers; When ride-hailing vehicles are in motion, real-time driver's driving characteristic image information is compared with standard and illegal driving behavior image information of the ride-hailing vehicle. Image feature matching was successful, indicating that the current ride-hailing driver has a [missing information - likely a typo or incomplete sentence]. If a certain type of illegal driving behavior is detected, the system will output the illegal driving behavior identification information of the ride-hailing driver, indicating that an illegal driving behavior has occurred, and simultaneously output the third type of illegal driving behavior. Text information on various types of illegal driving behavior; When ride-hailing vehicles are in motion, real-time driver's driving characteristic image information is compared with standard and illegal driving behavior image information of the ride-hailing vehicle. If no matching of image features is found, it indicates that the current ride-hailing driver has not committed any illegal driving behavior. Therefore, the output of the ride-hailing driver illegal driving behavior identification information is "no illegal driving behavior".

[0027] Based on real-time driver driving feature image information and a matrix of standard service attitude violation images of ride-hailing drivers, visual detection is used to analyze and process the driver's service attitude violations during ride-hailing operations, resulting in visual service attitude violation analysis information. Based on real-time driver voice text information and a matrix of standard service attitude violation voice text keywords, speech recognition is used to analyze and process the driver's service attitude violations during ride-hailing operations, resulting in voice service attitude violation analysis information. This includes the following steps: Step 231: Establish image information matrices of standard violations and service attitudes of ride-hailing drivers. ,in Indicates the first The system includes standard image information of ride-hailing drivers exhibiting different service attitudes, corresponding to various types of such attitudes. These attitude types include contempt, anger, disdain, resentment, verbal abuse, and intimidation. The system also includes standard facial and body language images of drivers during ride-hailing operations, tailored to different service attitude types. Finally, a keyword matrix of voice and textual descriptions of standard service attitude violations by ride-hailing drivers is established. ,in Indicates the first The standard voice text keywords for ride-hailing drivers corresponding to different types of service attitude violations are defined as the standard voice text keywords for drivers' communication during ride-hailing operations, which are set for different types of service attitude violations. The standard voice text keywords for ride-hailing drivers' service attitude violations include: get lost, try it, you don't deserve it, are you sick, I can't be bothered with you, useless, are you tired of living? Step 232: Use the FLANN image feature matching algorithm to combine the real-time driving feature image information of ride-hailing drivers with the standard violation service attitude image information matrix of ride-hailing drivers. Image information on the standard violations and service attitude of ride-hailing drivers in China Perform image feature matching, and generate analysis information on the driver's visual behavior regarding service attitude violations based on the image feature matching results; When ride-hailing vehicles are in motion, real-time driver's driving characteristics image information and ride-hailing vehicle driver's standard violation service attitude image information are used. Image feature matching was successful, indicating that the current ride-hailing driver has a [missing information - likely a typo or incomplete sentence]. If a certain type of service attitude is found to be non-compliant, the system will output the analysis information of the non-compliant service attitude from the driver's visual perspective, indicating that a non-compliant service attitude exists. Simultaneously, the system will output the third type of service attitude analysis. Textual information regarding unacceptable service attitude; When ride-hailing vehicles are in motion, real-time driver's driving characteristics image information and ride-hailing vehicle driver's standard violation service attitude image information are used. If no matching of image features is found, it indicates that the current ride-hailing driver does not have any illegal service attitude. Therefore, the output of the ride-hailing driver's visual illegal service attitude analysis information is "no illegal service attitude". Step 233: Use the Rabin-Karp text search algorithm to combine the real-time voice text information of ride-hailing drivers with the keyword matrix of standard violations and service attitude voice text of ride-hailing drivers. Chinese ride-hailing drivers' standards, violations, service attitude, voice text keywords Perform voice-text information matching, and generate analysis information on the illegal service attitude of ride-hailing drivers based on the voice-text information matching results; When a ride-hailing vehicle is in motion, real-time driver voice and text messages are exchanged for keywords related to the driver's standard driving behavior, violations, service attitude, and other issues. Successful matching of voice and text information indicates that the current ride-hailing driver has a [number] [unclear - possibly related to a previous sentence or paragraph]. If a driver exhibits a poor service attitude, the system will output "Analysis of Poor Service Attitude" from the driver's voice input device, indicating that a poor service attitude exists. Simultaneously, it will output the next... Textual information regarding unacceptable service attitude; When a ride-hailing vehicle is in motion, real-time driver voice and text messages are exchanged for keywords related to the driver's standard driving behavior, violations, service attitude, and other issues. If neither voice nor text information is successfully matched, it indicates that the current ride-hailing driver has not engaged in any service misconduct. Therefore, the output of the ride-hailing driver's voice service misconduct analysis information will be "no service misconduct."

[0028] By leveraging real-time driver image information and facial images of drivers registered for ride-hailing orders, combined with intelligent image search algorithms, the system autonomously and efficiently analyzes the legality of drivers' identities during ride-hailing operations. This enables dynamic and legal supervision of drivers throughout the entire ride-hailing journey, improving the compliance and regulatory oversight of ride-hailing order driving risks. Furthermore, by combining real-time driver image information with artificial intelligence algorithms and images of standard driving violations from big data storage, the system intelligently identifies drivers' illegal driving behaviors during ride-hailing operations, achieving intelligent supervision of illegal driving behaviors throughout the entire ride-hailing journey. Finally, the system utilizes real-time driver image information combined with intelligent image recognition algorithms... This system utilizes visual detection to precisely monitor drivers' service attitude violations during ride-hailing operations, combining legal and scientifically stored image information of drivers' standard service attitude violations with visual detection. Furthermore, it employs real-time driver voice and text information, combined with intelligent search algorithms and standard stored keywords related to drivers' standard service attitude violations, to perform deep and reliable voice recognition analysis of drivers' service attitude violations during ride-hailing operations. This enables intelligent and accurate supervision of service attitude violations throughout the entire ride-hailing process based on the fusion of visual and voice multimodal data, achieving multi-perspective intelligent safety monitoring of ride-hailing order driving risks based on driver legitimacy, driving behavior, and service attitude.

[0029] For further details, please refer to Figures 1-2 The process of assessing and processing the driving risks of drivers during the execution of ride-hailing orders yields ride-hailing order driving risk assessment information. When it is safe, the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed; when it is dangerous, a summary of ride-hailing driving risk information is constructed, the ride-hailing order driving risk push operation is executed, and the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed. This includes the following steps: Step 31: Based on the legality analysis information of ride-hailing drivers, the identification information of illegal driving behavior of ride-hailing drivers, the analysis information of illegal service attitude of ride-hailing drivers from the visual end, and the analysis information of illegal service attitude of ride-hailing drivers from the voice end, conduct a driving risk assessment of drivers during the execution of ride-hailing orders to obtain ride-hailing order driving risk assessment information; When the legality analysis information of the ride-hailing driver is legal, the identification information of the ride-hailing driver's illegal driving behavior is that there is no illegal driving behavior, the analysis information of the ride-hailing driver's visual service attitude is that there is no illegal service attitude, and the analysis information of the ride-hailing driver's voice service attitude is that there is no illegal service attitude, it means that there is no driver driving risk in the current ride-hailing order during the execution of the ride-hailing order. Then the ride-hailing order driving risk assessment information is output as safe. At this time, the ride-hailing order risk supervision operation continues until the ride-hailing order is completed. When the legality analysis information of the ride-hailing driver is illegal, or the identification information of the ride-hailing driver's illegal driving behavior is illegal, or the analysis information of the ride-hailing driver's visual service attitude is illegal, or the analysis information of the ride-hailing driver's voice service attitude is illegal, it indicates that there is a driver's driving risk in the current ride-hailing order during the order execution process, and the ride-hailing order driving risk assessment information is output as dangerous. Step 32: When the ride-hailing order driving risk assessment information is dangerous, the ride-hailing order text information, real-time driver driving feature image information, real-time driver voice text information, ride-hailing order registration driver facial image information, ride-hailing driver object legality analysis information, ride-hailing driver illegal driving behavior identification information, ride-hailing driver visual illegal service attitude analysis information, and ride-hailing driver voice illegal service attitude analysis information are combined and labeled to construct a ride-hailing driving risk summary information; Step 33: Push the summary information of ride-hailing driving risks to the ride-hailing management platform through the mobile communication network and output the screen flashing prompts to execute the ride-hailing order driving risk push operation. After the ride-hailing order driving risk push operation is completed, continue to execute the ride-hailing order risk supervision operation until the ride-hailing order is completed.

[0030] Based on driver legitimacy, driving behavior, and service attitude, and combined with data analysis, a comprehensive and scientific assessment of ride-hailing order driving risks is conducted. This enables efficient and safe supervision of ride-hailing order driving risks through multi-data fusion. It also allows for timely and efficient acquisition of summary information on ride-hailing driving risks based on ride-hailing orders, driver characteristics, driver legitimacy, driving behavior, and service attitude, combined with data processing. Furthermore, it works in conjunction with the ride-hailing management platform and display screens to safely and efficiently push ride-hailing order driving risk information, achieving safe and efficient feedback of ride-hailing order driving risks and improving the safety of ride-hailing travel and passenger service experience.

[0031] Example 2: Please see Figures 1-2A vehicle-based order risk management system is used to implement a vehicle-based order risk management method. The system includes a ride-hailing order service information acquisition module, a ride-hailing order driving safety monitoring module, and a ride-hailing order driving safety early warning module. The ride-hailing order service information acquisition module includes a ride-hailing order information collection unit, a ride-hailing order registration driver facial image search unit, a ride-hailing real-time driver driving feature image collection unit, and a ride-hailing real-time driver voice and text information collection unit. The system includes: a ride-hailing order information collection unit, which collects ride-hailing order text information through the ride-hailing management platform; a ride-hailing order registration driver facial image search unit, which searches for and processes the facial images of legally registered drivers based on the ride-hailing order text information and in conjunction with the ride-hailing management platform; a ride-hailing real-time driver driving feature image collection unit, which collects real-time driver driving feature image information through a mobile terminal camera; and a ride-hailing real-time driver voice and text information collection unit, which collects real-time driver voice and text information through a mobile terminal voice module. The ride-hailing order driving safety monitoring module includes a ride-hailing driver object legality analysis unit, a ride-hailing driver standard illegal driving behavior image storage unit, a ride-hailing driver illegal driving behavior identification unit, a ride-hailing driver standard illegal service attitude image storage unit, a ride-hailing driver visual illegal service attitude analysis unit, a ride-hailing driver standard illegal service attitude voice text keyword storage unit, and a ride-hailing driver voice illegal service attitude analysis unit; The system comprises the following components: a ride-hailing driver legitimacy analysis unit, which analyzes the legitimacy of driver identities during ride-hailing operations based on real-time driver feature images and facial images of drivers registered for ride-hailing orders; a standard driver violation image storage unit, which stores images of standard driver violation behaviors; a driver violation identification unit, which identifies driver violations during ride-hailing operations based on real-time driver feature images and standard driver violation images; and a standard driver service attitude image storage unit, which stores images of standard driver service attitude violations. The system includes: a visual image analysis unit for ride-hailing drivers, which analyzes and processes drivers' service attitude violations during ride-hailing operations based on real-time driver driving feature images and standard service attitude violation images, to obtain visual service attitude violation analysis information; a voice text keyword storage unit for standard service attitude violations, used to store voice text keywords of standard service attitude violations; and a voice-based service attitude violation analysis unit for ride-hailing drivers, which analyzes and processes drivers' service attitude violations during ride-hailing operations based on real-time driver voice text and standard service attitude violation voice text keywords, to obtain voice-based service attitude violation analysis information. The ride-hailing order driving safety early warning module includes a ride-hailing order driving risk assessment unit, a ride-hailing order driving risk information construction unit, and a ride-hailing order driving risk information push unit; The ride-hailing order driving risk assessment unit assesses the driving risks of drivers during the ride-hailing order execution process based on data analysis of driver legitimacy, driver violation behavior identification, driver visual and voice-based service attitude violation analysis, and other data. The ride-hailing order driving risk information construction unit constructs a summary of ride-hailing driving risks based on ride-hailing order text information, real-time driver driving characteristic image information, real-time driver voice text information, driver facial image information registered for the ride-hailing order, driver legitimacy analysis, driver violation behavior identification, driver visual and voice-based service attitude violation analysis, and other data processing. The ride-hailing order driving risk information push unit pushes ride-hailing order driving risk information based on the summary of driving risks and in conjunction with the ride-hailing management platform and display screen.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for risk management of vehicle-based orders, characterized in that, The method includes the following steps: The system collects ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information. Based on the ride-hailing order information, it searches and processes the facial images of legally registered drivers to obtain the facial image information of drivers registered for ride-hailing orders. The legitimacy of the identity of the ride-hailing driver during the ride-hailing process is analyzed and processed to obtain the legitimacy analysis information of the ride-hailing driver. Identify and process the illegal driving behavior of drivers during the ride-hailing process to obtain identification information of illegal driving behavior of ride-hailing drivers; Based on visual detection, the analysis and processing of drivers' non-compliant service attitude during ride-hailing operations are carried out to obtain the visual analysis information of non-compliant service attitude of ride-hailing drivers. Based on speech recognition, the analysis and processing of drivers' non-compliant service attitude during ride-hailing operations are performed to obtain analysis information on non-compliant service attitude of ride-hailing drivers via voice input. The system assesses and processes the driving risks of drivers during the execution of ride-hailing orders, obtaining ride-hailing order driving risk assessment information. When it is safe, the system continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed. When it is dangerous, the system constructs a summary of ride-hailing driving risks, performs ride-hailing order driving risk push operations, and continues to perform ride-hailing order risk monitoring until the ride-hailing order is completed.

2. The method for risk management of vehicle-based orders according to claim 1, characterized in that: The process involves collecting ride-hailing order text information, real-time driver driving feature image information, and real-time driver voice text information. Based on the ride-hailing order information, facial images of legitimate drivers registered for ride-hailing orders are searched and processed to obtain the facial image information of drivers registered for ride-hailing orders, including the following steps: The system obtains ride-hailing order information confirmed by drivers online through the ride-hailing management platform and generates ride-hailing order text information; it also collects images of drivers' frontal faces and body movements online during the ride-hailing order execution through the mobile terminal's camera and generates real-time driver driving characteristic image information; and it collects drivers' driving communication voice text information online during the ride-hailing order execution through the mobile terminal's voice module and generates real-time driver voice text information. Based on the ride-hailing order text information, a search and processing of the facial images of legitimate drivers registered for ride-hailing orders is performed to obtain the facial image information of drivers registered for ride-hailing orders.

3. The method for risk management of vehicle-based orders according to claim 2, characterized in that: Based on the ride-hailing order text information, the process of searching for and processing the facial images of legitimate drivers registered for ride-hailing orders to obtain the facial image information of drivers registered for ride-hailing orders includes the following steps: Obtain the ride-hailing order text information; The KD-tree nearest neighbor search algorithm is used to search for facial image information of ride-hailing drivers who have legally registered ride-hailing orders on the ride-hailing management platform based on the ride-hailing order text information, and to construct the facial image information of drivers registered for ride-hailing orders.

4. The method for risk management of vehicle-based orders according to claim 3, characterized in that: The legitimacy of the identity of the ride-hailing driver during the ride-hailing process is analyzed and processed to obtain the legitimacy analysis information of the ride-hailing driver. Identify and process the illegal driving behavior of drivers during the ride-hailing process to obtain identification information of illegal driving behavior of ride-hailing drivers; Based on visual detection, the analysis and processing of drivers' non-compliant service attitude during ride-hailing operations are carried out to obtain the visual analysis information of non-compliant service attitude of ride-hailing drivers. The analysis of drivers' poor service attitude during ride-hailing operations based on speech recognition includes the following steps: Based on the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order, the legality analysis of the identity of the driver during the ride-hailing process is performed to obtain legality analysis information of the ride-hailing driver. Based on the real-time driver driving feature image information and the standard illegal driving behavior image information matrix of the ride-hailing driver, the illegal driving behavior of the driver during the ride-hailing process is identified and processed to obtain the illegal driving behavior identification information of the ride-hailing driver. Based on the real-time driving feature image information of the ride-hailing driver and the standard violation service attitude image information matrix of the ride-hailing driver, the driver's violation service attitude during the ride-hailing process is analyzed and processed using visual detection to obtain the driver's visual violation service attitude analysis information; based on the real-time voice text information of the ride-hailing driver and the standard violation service attitude voice text keyword matrix of the ride-hailing driver, the driver's violation service attitude during the ride-hailing process is analyzed and processed using speech recognition to obtain the driver's voice violation service attitude analysis information.

5. The method for risk management of vehicle-based orders according to claim 4, characterized in that: Based on the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order, the legality analysis of the driver's identity during the ride-hailing process is performed to obtain the legality analysis information of the ride-hailing driver, including the following steps: Obtain the real-time driving feature image information of the ride-hailing driver and the facial image information of the driver registered for the ride-hailing order; The SURF image feature matching algorithm is used to match the real-time driving feature image information of the ride-hailing driver with the facial image information of the driver registered in the ride-hailing order. Based on the image feature matching results, legal analysis information of the ride-hailing driver object is generated. When the real-time driving feature image information of the ride-hailing driver is successfully matched with the facial image information of the driver registered in the ride-hailing order, the legality analysis information of the ride-hailing driver object is output as legal. If the real-time driving feature image information of the ride-hailing driver fails to match the facial image information of the driver registered in the ride-hailing order, the legality analysis information of the ride-hailing driver object will be output as invalid.

6. The method for risk management of vehicle-based orders according to claim 5, characterized in that: Based on the real-time driver driving feature image information and the standard illegal driving behavior image information matrix of the ride-hailing driver, the illegal driving behavior identification process of the driver during the ride-hailing process is performed to obtain the illegal driving behavior identification information of the ride-hailing driver, including the following steps: Establish a matrix of image information on standard driving violations by ride-hailing drivers. The include ;in Indicates the first Image information of standard illegal driving behavior of ride-hailing drivers corresponding to various types of illegal driving behavior; The real-time driver's driving feature image information of the ride-hailing vehicle is compared with the... The above Image feature matching is performed, and information on illegal driving behavior of ride-hailing drivers is generated based on the image feature matching results. The specific steps for generating the information on illegal driving behavior of ride-hailing drivers are as follows: Step 2221, Initialization Phase: Update the maximum number of iterations In the Initialize and update driving behavior within the search space to search for the location of the female spider wasp; Step 2222, Search Phase: The driving behavior of the female spider wasp searches for female spiders at a constant step size in the... Randomly search the search space to find the images that match the real-time driving feature information of the ride-hailing driver. Meanwhile, the driving behavior search for female spider wasps in the above... The search location within the search space; Step 2223, Follow and Hunt Phase: The driver searches for the female spider wasp at random step lengths. Randomly search the search space to find the images that match the real-time driving feature information of the ride-hailing driver. The prey, the female spider wasp searching for driving behavior, moves towards the image information that best matches the real-time driving characteristics of the ride-hailing driver. The prey moves in a direction and its position is updated. Step 2224: When the maximum number of iterations is met, output the image that best matches the real-time driver's driving feature image information of the ride-hailing vehicle. ; Step 2225: Based on the real-time driver's driving feature image information of the ride-hailing vehicle and the... Image feature matching results are used to generate identification information on illegal driving behavior of ride-hailing drivers; When the ride-hailing vehicle is in motion, the real-time driver's driving feature image information is compared with the above... Image feature matching was successful, and the first result was output. Text information on various types of illegal driving behavior; When the ride-hailing vehicle is in motion, the real-time driver's driving feature image information is compared with the above... If no matching of image features is found, the output of the ride-hailing driver's violation identification information is "no violation of driving behavior".

7. The method for risk management of vehicle-based orders according to claim 6, characterized in that: Based on the real-time driving feature image information of the ride-hailing driver and the standard violation service attitude image information matrix of the ride-hailing driver, the driver's violation service attitude during the ride-hailing process is analyzed and processed using visual detection to obtain the visual violation service attitude analysis information of the ride-hailing driver. Based on the real-time voice and text information of the ride-hailing driver and the keyword matrix of the driver's standard violation service attitude voice and text during the ride-hailing process, the analysis and processing of the driver's violation service attitude during the ride-hailing process is carried out based on voice recognition. The analysis information of the driver's violation service attitude on the voice terminal includes the following steps: Establish separate image information matrices for ride-hailing drivers' standard violations and service attitudes. The include ;in Indicates the first Image information of standard service attitude violations by ride-hailing drivers corresponding to different types of service attitude violations; and the establishment of a keyword matrix of voice and text messages regarding standard service attitude violations by ride-hailing drivers. The include ;in Indicates the first Keywords in the voice text of ride-hailing drivers' standard non-compliant service attitudes corresponding to different types of non-compliant service attitudes; The FLANN image feature matching algorithm is used to match the real-time driving feature image information of the ride-hailing driver with the... The above Perform image feature matching, and generate analysis information on the driver's visual behavior regarding service attitude violations based on the image feature matching results; When the ride-hailing vehicle is in motion, the real-time driver's driving feature image information is compared with the above... If image feature matching is successful, the output of the ride-hailing driver's visual terminal violation service attitude analysis information is "violation service attitude exists," and the output of the first... Textual information regarding unacceptable service attitude; When the ride-hailing vehicle is in motion, the real-time driver's driving feature image information is compared with the above... If no matching of image features is found, the output of the ride-hailing driver's visual terminal violation service attitude analysis information is "no violation service attitude". The Rabin-Karp text search algorithm was used to combine the real-time voice and text information of the ride-hailing driver with the... The above Perform voice-text information matching, and generate analysis information on the illegal service attitude of ride-hailing drivers based on the voice-text information matching results; When the ride-hailing vehicle is in motion, the real-time driver's voice and text information is consistent with the above. If the voice-to-text information matching is successful, the output of the analysis information on the ride-hailing driver's voice terminal for non-compliant service attitude is "there is a non-compliant service attitude," and the output of the first... Textual information regarding unacceptable service attitude; When the ride-hailing vehicle is in motion, the real-time driver's voice and text information is consistent with the above. If neither voice nor text information is successfully matched, the output of the analysis information on the non-compliant service attitude of the ride-hailing driver's voice terminal will be "no non-compliant service attitude".

8. The method for risk management of vehicle-based orders according to claim 7, characterized in that: The process of assessing and processing the driving risks of drivers during the execution of ride-hailing orders yields ride-hailing order driving risk assessment information. When it is safe, the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed; when it is dangerous, a summary of ride-hailing driving risk information is constructed, the ride-hailing order driving risk push operation is executed, and the ride-hailing order risk monitoring operation continues until the ride-hailing order is completed. This includes the following steps: Based on the legality analysis information of the ride-hailing driver, the identification information of the ride-hailing driver's illegal driving behavior, the analysis information of the ride-hailing driver's visual illegal service attitude, and the analysis information of the ride-hailing driver's voice illegal service attitude, the driving risk assessment of the driver during the ride-hailing order execution process is carried out to obtain the ride-hailing order driving risk assessment information. When the legality analysis information of the ride-hailing driver is legal, the identification information of the ride-hailing driver's illegal driving behavior is that there is no illegal driving behavior, the analysis information of the ride-hailing driver's visual service attitude is that there is no illegal service attitude, and the analysis information of the ride-hailing driver's voice service attitude is that there is no illegal service attitude, then the ride-hailing order driving risk assessment information is output as safe. At this time, the ride-hailing order risk supervision operation continues until the ride-hailing order is completed. When the legality analysis information of the ride-hailing driver is illegal, or the illegal driving behavior identification information of the ride-hailing driver indicates illegal driving behavior, or the illegal service attitude analysis information of the ride-hailing driver's visual terminal indicates illegal service attitude, or the illegal service attitude analysis information of the ride-hailing driver's voice terminal indicates illegal service attitude, then the ride-hailing order driving risk assessment information is output as dangerous. When the ride-hailing order driving risk assessment information is deemed dangerous, the ride-hailing order text information, the real-time driver driving feature image information, the real-time driver voice text information, the ride-hailing order registration driver facial image information, the ride-hailing driver object legality analysis information, the ride-hailing driver illegal driving behavior identification information, the ride-hailing driver visual illegal service attitude analysis information, and the ride-hailing driver voice illegal service attitude analysis information are combined and identified to construct a ride-hailing driving risk summary information; The summarized information on ride-hailing driving risks is pushed to the ride-hailing management platform via the mobile communication network, and a flashing prompt is displayed on the screen to execute the ride-hailing order driving risk push operation. After the ride-hailing order driving risk push operation is completed, the ride-hailing order risk supervision operation continues until the ride-hailing order is completed.

9. A vehicle-based order risk management system, used to implement the vehicle-based order risk management method according to any one of claims 1-8, characterized in that: The system includes a ride-hailing order service information acquisition module, a ride-hailing order driving safety monitoring module, and a ride-hailing order driving safety early warning module.

Images