A dual-end passenger identity authentication method for autonomous driving vehicles

Abstract

The present invention belongs to the field of identity authentication for autonomous driving, and specifically discloses a two-end passenger identity authentication method for autonomous driving vehicles, including the following steps: pre-processing the passenger's facial information for verification, the pre-processing process includes facial information collection, detection and verification; saving the original passenger face image to the server; facial reasoning framework service, converting the newly uploaded passenger face photo in real time into facial vector data, returning and saving it to the database; after the passenger successfully places an order, the server pushes the order and the passenger's facial vector data to the vehicle system of the corresponding order-receiving vehicle; and performing identity verification on the passenger after the passenger gets on the vehicle. The present invention can accurately identify the passenger's true identity through facial recognition, prevent the inconsistency between the taxi-takers and the passengers, and has a fast facial recognition speed, which improves the convenience of user operation and user experience.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving identity authentication, and in particular to a dual-end passenger identity authentication method for an autonomous driving vehicle. Background Art

[0002] Identity authentication technology has been developing continuously in recent years, especially the technology based on human's inherent biometrics to verify user identity has been widely used in various fields. As a new taxi-hailing model for online ride-hailing, safety officers and self-driving vehicles are being tested on the road in the experimental area. Passenger identity recognition is a key existence for vehicle safety.

[0003] At present, autonomous driving online ride-hailing vehicles are still in the research and development stage. They cannot ensure rapid identification of users on complex public transportation roads, that is, they cannot achieve convenient, safe and efficient passenger identity authentication processing. Therefore, a dual-end passenger identity authentication method for autonomous driving vehicles is needed, which can accurately complete the identity authentication and identification processing, ensuring efficiency while also ensuring the accuracy of identification. Summary of the invention

[0004] The purpose of the present invention is to provide a dual-end passenger identity authentication method for an autonomous driving vehicle to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a dual-end passenger identity verification method for an autonomous driving vehicle, comprising the steps of:

[0006] S1: Verify and preprocess the passenger's facial information. The preprocessing process includes facial information collection, detection and verification;

[0007] S2: Save the original passenger face image to the server;

[0008] S3: Face reasoning framework service, which converts newly uploaded passenger face photos into face vector data and saves it to the database;

[0009] S4: After the passenger successfully places an order, the server pushes the order and the passenger's face vector data to the vehicle system of the corresponding vehicle that receives the order;

[0010] S5: After the passenger gets on the bus, the passenger's identity is verified.

[0011] Preferably, S1 specifically includes:

[0012] S101: calling the mobile phone camera to take photos, and obtaining multiple facial photos of passengers;

[0013] S102: Obtain a face photo and use a multi-task cascade convolutional neural network algorithm to perform face detection;

[0014] S103: Perform liveness check on passengers based on MINIFASNET liveness silent algorithm to ensure real person operation.

[0015] Preferably, when S2 is sent to the server, a signature string sign is added to the message to increase security. The signature algorithm uses SHA256, and the image is converted into a binary byte stream and transmitted to the server, and then saved in the database.

[0016] Preferably, data push in S4 adopts a signature algorithm based on SHA256.

[0017] Preferably, S5 specifically includes:

[0018] S501: Taking photos and collecting them through the vehicle computer and / or an externally installed camera to obtain a facial photo of the passenger boarding the vehicle;

[0019] S502: Use a multi-task cascade convolutional neural network algorithm to detect the face of the passenger boarding the bus;

[0020] S503: Performing liveness detection to ensure real-person operation;

[0021] S504: Extracting mobile terminal facial feature strings based on MobileFaceNets algorithm;

[0022] S505: Collect and extract the passenger's current feature vector, compare and verify it with the face collected on the mobile side and the feature vector calculated by the server, and complete the final verification of the passenger's identity.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] The present invention introduces face recognition into the field of identity recognition in autonomous driving. Through face recognition, the real identity of the passenger can be accurately identified to prevent the taxi-hailing person and the passenger from being inconsistent. The face recognition speed is fast, which improves the convenience of user operation and user experience. Moreover, the method logic of the present invention is that the original face image of the passenger is transmitted through a vector, which greatly enhances security and is not easy to leak user private information. The passenger's face is stored in the online taxi-hailing platform, which can facilitate subsequent supervision and real-name riding. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 The figure is a flow logic block diagram of the method of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0027] See also Figure 1 The present invention provides a technical solution: a dual-end passenger identity authentication method for an autonomous driving vehicle, comprising the steps of:

[0028] S1: Verify and preprocess the passenger's facial information. The preprocessing process includes facial information collection, detection and verification;

[0029] S2: Save the original passenger face image to the server;

[0030] S3: Face reasoning framework service, which converts newly uploaded passenger face photos into face vector data and saves it to the database;

[0031] S4: After the passenger successfully places an order, the server pushes the order and the passenger's face vector data to the vehicle system of the corresponding vehicle that receives the order;

[0032] S5: After the passenger gets on the bus, the passenger's identity is verified.

[0033] In this embodiment, S1 specifically includes:

[0034] S101: calling the mobile phone camera to take photos, and obtaining multiple facial photos of passengers;

[0035] S102: Obtain a face photo and use a multi-task cascade convolutional neural network algorithm to perform face detection;

[0036] S103: Perform liveness check on passengers based on MINIFASNET liveness silent algorithm to ensure real person operation.

[0037] In this embodiment, when S2 is sent to the server, a signature string sign is added to the message to increase security. The signature algorithm uses SHA256, and the image is converted into a binary byte stream and transmitted to the server, and then saved in the database.

[0038] In this embodiment, the data push in S4 adopts a signature algorithm based on SHA256.

[0039] In this embodiment, S5 specifically includes:

[0040] S501: Taking photos and collecting them through the vehicle computer and / or an externally installed camera to obtain a facial photo of the passenger boarding the vehicle;

[0041] S502: Use a multi-task cascade convolutional neural network algorithm to detect the face of the passenger boarding the bus;

[0042] S503: Performing liveness detection to ensure real-person operation;

[0043] S504: Extracting mobile terminal facial feature strings based on MobileFaceNets algorithm;

[0044] S505: Collect and extract the passenger's current feature vector, compare and verify it with the face collected on the mobile side and the feature vector calculated by the server, and complete the final verification of the passenger's identity.

[0045] In this embodiment, the MobileFaceNets used in S504 has the advantage of being more than twice as fast as other models in the same environment compared to other CNN architectures with high efficiency, such as MobilenetV1, ShuffleNet, and MobileNetV2. When the single size is below 5MB, the analysis accuracy of MobileFaceNets on LFW and AgeDB 30 can reach more than 95%, and the calculation time on the mobile phone can be as short as 20 milliseconds, and finally a face vector with 128 values ​​is obtained.

[0046] In this embodiment, the facial feature vector similarity comparison algorithm uses a cosine similarity algorithm to perform comparison to obtain a score after comparison. The higher the score, the higher the similarity, and the two can be regarded as the same person, and the identity verification is completed.

[0047] The cosine of the angle β between two vectors can be expressed as:

[0048] Cosβ=a×b / |a|×|b|

[0049] In this embodiment, assuming two vectors a=[2,2,3] and b=[3,3,4], the cosine similarity of the angle β between the two vectors is calculated as follows. Substituting vectors a and b into the formula:

[0050]

[0051] =0.998

[0052] The numerical range of cosine similarity is the range of cosine value, that is, [-1, 1]. The higher the value, the greater the similarity. Generally, it is considered to be the same person if it is greater than 0.7. A vector of actual face comparison contains 128 items of data.

[0053] The embodiment of the present invention uses multi-task cascade convolutional neural network MTCN + liveness detection MINIFASNET + feature string extraction MobileFaceNets, and finally uses cosine similarity algorithm for comparison, forming the face recognition architecture system of the present invention. At the same time, it opens up the process from collecting faces on the mobile phone, placing orders, to identifying passengers on the unmanned vehicle.

[0054] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A dual-end passenger identity verification method for an autonomous driving vehicle, characterized in that: Includes steps: S1: Verify and preprocess the passenger's facial information. The preprocessing process includes facial information collection, detection and verification; S1 specifically includes: S101: calling the mobile phone camera to take photos, and obtaining multiple facial photos of passengers; S102: Obtain a face photo and use a multi-task cascade convolutional neural network algorithm to detect face information; S103: Perform liveness check on passengers based on MINIFASNET liveness silent algorithm to ensure real-person operation; S2: Save the original photo of the passenger's face to the server; When the original photo of the passenger's face is saved to the server in S2, a signature string sign is added to the message to increase security. The signature algorithm uses SHA256, and the image is converted into a binary byte stream and transmitted to the server, and then saved to the database; S3: Face reasoning framework service, which converts newly uploaded passenger face photos into face vector data and saves it to the database; S4: After the passenger successfully places an order, the server pushes the order and the passenger's face vector data to the vehicle system of the corresponding vehicle that receives the order; S5: After the passenger gets on the bus, the passenger's identity is verified; S5 specifically includes: S501: Taking photos and collecting them through the vehicle computer and / or an externally installed camera to obtain a facial photo of the passenger boarding the vehicle; S502: Use a multi-task cascade convolutional neural network algorithm to detect the face of the passenger boarding the bus; S503: Performing liveness detection to ensure real-person operation; S504: Extracting mobile terminal facial feature strings based on MobileFaceNets algorithm; S505: Collect and extract the passenger's current feature vector, compare and verify it with the face collected on the mobile terminal and the feature vector calculated by the server, and complete the final verification of the passenger's identity.

2. The dual-end passenger identity verification method for an autonomous driving vehicle according to claim 1, characterized in that: The data push in S4 adopts a signature algorithm based on SHA256.

Images