Non-contact road transportation inspection system and method

Abstract

The invention discloses a non-contact road transportation inspection system and method, and relates to the technical field of traffic safety inspection, and the system comprises a personnel image collection module which obtains the face images of vehicle personnel entering a preset inspection area, the vehicle digital authentication module is used for acquiring a digital authentication certificate or vehicle characteristics corresponding to the vehicle based on the Internet of Vehicles, and the verification module is used for judging the high-risk vehicle by performing fusion verification on the digital authentication certificate, the vehicle image characteristics and the face image of the person in the vehicle. According to the invention, the authenticity and reliability of vehicle identity verification are ensured by using the vehicle digital authentication certificate and the block chain technology, the multi-modal feature verification is combined with the vehicle image features and the in-vehicle personnel biological feature information, the comprehensive judgment of the vehicle and personnel identities is realized, and the verification precision is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of traffic safety inspection technology, specifically a non-contact road transport inspection system and method. Background Technology

[0002] In the current road transport safety management system, the inspection of vehicles and their occupants still mainly relies on contact-based manual verification and non-contact single-point device identification. Traditional manual verification requires law enforcement officers to manually check the vehicle information, driver's license, and the identities of accompanying personnel sequentially. This process is cumbersome, slow, and can easily cause traffic congestion and delays during peak hours. Furthermore, it is susceptible to human error, leading to the risk of misjudgments and omissions, making efficient supervision difficult.

[0003] Conventional contactless methods mostly focus on reading and verifying single-dimensional data such as license plate recognition and RFID tag scanning. These verification methods are prone to inaccuracy when vehicle information is forged, obscured, or the tag is invalid, and they cannot effectively identify people inside the vehicle. Summary of the Invention

[0004] 1) Technical problems to be solved This invention provides a non-contact road transport inspection system and method that can verify and inspect vehicles and their occupants without human intervention.

[0005] (ii) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a non-contact road transport inspection system, comprising: The personnel image acquisition module counts the number of people inside the vehicle and acquires the facial images of the people inside the vehicle when the vehicle enters the preset inspection area. The vehicle digital authentication module uses a blockchain consensus mechanism to perform digital identity authentication on the vehicle when it enters the preset inspection area. If the digital identity authentication is successful, a digital authentication certificate is obtained; if the digital identity authentication fails, vehicle image features including license plate number, vehicle model and body color are obtained. The verification module includes a first sub-channel and a second sub-channel of a parallel processing architecture; wherein: The first sub-channel is used to receive the successfully authenticated digital certificate, and confirm the legal identity of the corresponding vehicle based on the hash signature verification method nested in the blockchain consensus mechanism, and generate the corresponding identity code; if the vehicle is not successfully authenticated, the vehicle image feature analysis channel is started, the multimodal vehicle image feature information is extracted using image depth features, uniformly encoded, and compared with the vehicle identity database before outputting the vehicle verification result. The second sub-channel uses its facial recognition network to extract the facial features of the people inside the vehicle, compares them with the registered identity information database, and outputs the personnel verification result; when there is at least one verification result marked as abnormal, the vehicle is marked as a high-risk vehicle. The passage strategy module sends a release command to a vehicle and synchronizes its passage status when the vehicle entering the preset inspection area is not marked as a high-risk vehicle by the verification module; when the vehicle is marked as a high-risk vehicle, the passage of the vehicle is blocked, triggering a manual re-inspection, and the vehicle's identification code, vehicle image features, and facial images of the people inside the vehicle are recorded and uploaded.

[0006] Furthermore, when the vehicle enters the preset inspection area, the personnel image acquisition module counts the number of people in the vehicle using infrared sensors set up in the preset inspection area, and verifies the number of passengers in the vehicle by combining the camera and facial recognition detection algorithm. The camera in the personnel image acquisition module uses autofocus technology to capture facial images of people inside the vehicle. After image enhancement and facial localization, the captured facial image data is extracted, encrypted, and synchronized to the verification module.

[0007] Furthermore, when a vehicle enters the preset inspection area, the vehicle digital authentication module initiates an identity authentication request to the vehicle via a communication terminal, and performs digital identity authentication of the vehicle using a blockchain consensus mechanism; specifically: If the vehicle has been registered and authenticated in the vehicle network digital identity system, its digital identity certificate registered on the blockchain is extracted as a credential of vehicle legitimacy, the vehicle identity is confirmed and the digital authentication result is recorded. If the vehicle identification fails or the vehicle is not registered, the vehicle image recognition process is triggered, and the following steps are performed: The camera set up in the preset inspection area captures the vehicle's exterior image, extracts vehicle image features including license plate number, vehicle model and body color, and synchronizes them to the verification module after encryption.

[0008] Furthermore, the personnel image acquisition module, vehicle digital authentication module, and verification module adopt a unified data exchange protocol; The verification module receives structured facial image data output by the personnel image acquisition module in real time via an internally configured data bus, and receives digital authentication certificates and vehicle image feature data output by the vehicle digital authentication module in real time.

[0009] Furthermore, the vehicle's digital certificate includes the vehicle's unique identifier, the certificate issuing authority's public key, the validity period, and a digital signature; After the verification module obtains the vehicle's digital certification certificate in real time, the first sub-channel first uses the built-in public key of the certificate issuing authority to perform hash signature verification on the certificate, and compares it with the issuance record and revocation record on the blockchain, and confirms that the certificate is currently in a valid state after comparing it with the valid time interval. If the hash signature verification passes and the certificate is currently valid, the vehicle's identity is confirmed as legitimate. The verification module, based on the hash signature structure nested in the blockchain consensus mechanism, combines the certificate hash with the number of the current query node to generate a unique identity code for the vehicle in this verification process.

[0010] Furthermore, when the verification module obtains the vehicle image features, it indicates that the vehicle has not been successfully authenticated. The vehicle image feature analysis channel in the verification module is then activated. The feature fusion algorithm is used to uniformly encode the license plate number, vehicle model, and body color information into a vehicle feature vector. This vector is then compared with the vehicle identity database, and the similarity threshold matching result is output to verify the legality of the vehicle identity.

[0011] Furthermore, when the verification module acquires the facial images of the people inside the vehicle in real time, a facial recognition network is set up in the second sub-channel to extract the facial feature vectors corresponding to the people inside the vehicle. The facial feature vectors of the people inside the vehicle are compared with the registered identity information database in the public security transportation supervision system. People whose comparison results match are marked as legitimate passengers. If the comparison results show blacklisted, untrustworthy, or illegal personnel, abnormal personnel information is output and the corresponding vehicle is marked as a high-risk vehicle.

[0012] Furthermore, when a vehicle entering the preset inspection area is not identified as a high-risk vehicle by the verification module, the passage strategy module sends a release instruction to the vehicle and simultaneously uploads the vehicle's passage status, including the verification time node, identity code, and passage result, to the central database. When a vehicle entering the preset inspection area is verified as a high-risk vehicle by the module, the passage strategy module sends a delayed release signal to block the vehicle from passing, triggers a manual interception and re-inspection process, and pushes a high-risk passage warning to adjacent inspection areas, including sending the vehicle's identification code or vehicle image features, as well as facial images of the people inside the vehicle.

[0013] A non-contact road transport inspection method includes the following steps: Multiple preset inspection areas are set up. When a vehicle enters any of the preset inspection areas, the number of people in the vehicle is counted and the facial images of the people in the vehicle are collected. While collecting the facial images, the vehicle is digitally authenticated using a blockchain consensus mechanism. If the digital identity authentication is successful, its digital authentication certificate is obtained; if the digital identity authentication fails, vehicle image features including license plate number, vehicle model, and body color are obtained. Upon receiving the successfully authenticated digital certificate, the system confirms the legal identity of the corresponding vehicle based on the hash signature verification method nested in the blockchain consensus mechanism and generates a corresponding identity code. If the vehicle fails authentication, the system extracts multimodal vehicle image feature information using image depth features, performs unified encoding, compares it with the vehicle identity database, and outputs the vehicle verification result. The system also extracts the facial features corresponding to the occupants of the vehicle, compares them with the registered identity information database, and outputs the occupant verification result. If at least one verification result is marked as abnormal, the vehicle is marked as a high-risk vehicle. When a vehicle entering the preset inspection area is not marked as a high-risk vehicle, a release instruction is sent to the vehicle and its passage status is synchronized; when it is marked as a high-risk vehicle, the vehicle's passage is blocked, triggering a manual re-inspection, and the vehicle's identification code, vehicle image features, and facial images of the people inside the vehicle are recorded and uploaded.

[0014] (iii) Beneficial effects: Compared with the prior art, this invention has the following beneficial effects: The system of this invention can automatically complete the process of personnel image acquisition, vehicle identity authentication, high-risk judgment and release after a vehicle enters the inspection area. It can perform the core verification process of rapid processing and comparison within the inspection area without human intervention, reducing the burden on front-line law enforcement and reducing human error.

[0015] By employing a blockchain consensus mechanism, vehicles entering the inspection area are digitally identified, ensuring the authentication results cannot be tampered with and supporting traceability, thus enhancing the credibility of the verification. Furthermore, by combining and comparing the acquired license plate number, vehicle model, and body color image features, alternative verification methods can be used even in the absence of digital authentication, ensuring the flexibility of the system's authentication process. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a non-contact road transport inspection system provided in an embodiment of the present invention; Figure 2 A schematic diagram of the operation flow of each module in a non-contact road transport inspection system provided in an embodiment of the present invention; Figure 3 This is a flowchart illustrating a non-contact road transport inspection method provided in an embodiment of the present invention. Figure 4A schematic diagram illustrating the connection relationship and data processing between the personnel image acquisition module, the vehicle digital authentication module, and the verification module in a non-contact road transport inspection system provided in an embodiment of the present invention. Figure 5 This is a schematic diagram of the process for verifying vehicle information in the verification module of a non-contact road transport inspection system provided in an embodiment of the present invention. In the picture: 100. Personnel Image Acquisition Module; 200. Vehicle Digital Authentication Module; 300. Verification Module; 400. Traffic Strategies Module. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0018] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0019] Furthermore, the terms "first" and "second" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0020] It should be noted that, where there is no conflict, the features in the embodiments of the present invention can be combined with each other.

[0021] Combination Figures 1 to 5 This invention discloses a non-contact road transport inspection system and method. The system combines vehicle digital authentication certificates and image features with personnel biometric information. Through multi-dimensional verification methods, it accurately identifies vehicles and personnel, achieving high-precision identity verification and status recognition of vehicles and accompanying personnel without human intervention. The system can automatically classify risk levels based on verification results, send real-time warnings to high-risk targets, and link subsequent inspection areas, forming a coordinated response and closed-loop tracking intelligent supervision model to prevent abnormal vehicles from being missed or lost to oversight.

[0022] First, refer to Figure 1The preset inspection area in this embodiment refers to the physical space where the system is actually deployed and data is collected. Typically, this type of preset inspection area is located before or after a highway or expressway toll station or ETC lane. Within these common deployment areas, existing ETC devices or video surveillance facilities can be integrated to facilitate data collection from vehicles and images of occupants when vehicles slow down or make brief stops. Combining with existing monitoring and data collection equipment also ensures convenient deployment and controllable costs.

[0023] For reference here Figure 1 and Figure 2 First, there is the personnel image acquisition module 100 and the vehicle digital authentication module 200, which acquire vehicle and personnel data.

[0024] The main task of the personnel image acquisition module 100 is to automatically identify the number of people inside the vehicle when the vehicle enters the preset inspection area, and to acquire and extract high-quality facial images of each person, converting them into feature vectors. This enables a non-contact, automated, real-time, edge-processing image acquisition and processing flow, providing reliable input data for the downstream verification module 300, namely the number of people inside the vehicle plus their corresponding facial features.

[0025] Specifically, the personnel image acquisition module 100 includes devices deployed within the preset inspection area, including but not limited to infrared wide-angle cameras and facial recognition cameras. The infrared wide-angle camera is typically deployed in front of the preset detection area to scan the entire vehicle cabin, suitable for nighttime and other complex lighting conditions. In some embodiments of the invention, assuming the preset inspection area is an ETC checkpoint, a high-resolution infrared wide-angle camera is installed at the vehicle passage lane of the ETC checkpoint to cover all passenger areas inside the vehicle. The facial recognition camera is used to focus on the window area to obtain clear facial images. The selection of this device needs to consider its autofocus and imaging capabilities in low-light environments to ensure clear facial image acquisition under different lighting conditions.

[0026] Regarding the triggering mechanism for activating the personnel image acquisition module 100 to acquire images, in some embodiments of the present invention, the vehicle entering the area can be determined by geomagnetic detection, inductive loop, or visual detection, thereby triggering image acquisition.

[0027] Before acquiring face images for facial feature extraction, it is necessary to count the number of people in the vehicle. A lightweight face detection and recognition model, such as MobileNetV3, can be used to detect the faces of all people in the vehicle in real time, thereby counting the number of people in the vehicle and collecting face images of each person.

[0028] The facial image data of the people inside the vehicle is pre-processed locally in the personnel image acquisition module 100, and then the facial features of each image are extracted, encrypted, and synchronized to the verification module 300 of the system for subsequent verification.

[0029] More specifically, the preprocessing operations performed on each captured face image include, but are not limited to, image enhancement and face localization and cropping. Image enhancement is suitable for dealing with scenes with uneven lighting and backlighting, while face localization and cropping are for the purpose of facilitating the input of a unified and standardized model in the future.

[0030] After extracting facial features, a recognition model is used to extract feature vectors from each standardized facial image. For privacy protection, the collected facial images are encrypted in real time, such as with AES encryption, to ensure security during transmission and storage.

[0031] Based on the overall implementation process of the personnel image acquisition module 100 described above, let's take an example: Suppose a truck enters an ETC checkpoint. The personnel image acquisition module 100 detects two people in the front row. An infrared camera captures the approximate outline of each person, and an RGB camera acquires high-resolution images of the people. The personnel image acquisition module 100 successfully identifies the two faces using a lightweight model based on the MobileNetV3 and Blaze Face architecture. Arc Face extracts 128-dimensional features from each face image. After image preprocessing and angle correction, the generated features are synchronized to the subsequent verification module 300 via an edge encryption module for subsequent correlation analysis.

[0032] This combination Figure 3 The inspection method shown can be understood as follows: S10 is performed by the personnel image acquisition module 100: multiple preset inspection areas are set up, and when a vehicle enters any preset inspection area, the number of people in the vehicle is counted and the facial images of the people in the vehicle are acquired.

[0033] This step can be understood as follows: when the vehicle enters the inspection area, the personnel image acquisition module 100 is activated. The camera acquires a panoramic image of the vehicle interior, automatically detects and crops the facial image of each passenger, and uploads the acquired facial image to the system's verification module 300 for identity comparison.

[0034] In summary, it can be understood that the personnel image acquisition module 100 can quickly count the number of people and acquire images to ensure the accuracy of subsequent identity verification. No manual operation is required. It automatically detects the number of people and collects features. Furthermore, the personnel feature recognition mentioned in the above embodiment uses a lightweight module, so it can achieve data processing at the second level. The feature data is processed locally and then encrypted before transmission to avoid image leakage.

[0035] Regarding the vehicle digital authentication module 200, its task is to automatically acquire the vehicle's legitimate identity information when the vehicle enters the preset inspection area. It prioritizes a blockchain-based digital identity authentication mechanism. If digital authentication fails, the vehicle digital authentication module 200 automatically switches to image acquisition and feature recognition methods to provide reliable and structured vehicle identity information for subsequent verification. Regarding digital identity authentication, it is important to understand that it pertains to vehicle information equipped with vehicle-to-everything (V2X) technology. These vehicles are equipped with onboard units (OBUs) that store digital authentication certificates. These certificates include the vehicle's unique digital ID, public key, signature, and other information. The authenticity and validity of the certificate are verified using a blockchain consensus mechanism, such as PBFT. Specifically, this involves checking whether the certificate signature matches and verifying whether the certificate has expired or been revoked.

[0036] The reason why this invention chooses a vehicle digital certificate as the primary means of vehicle identity verification is that when each vehicle registers with the vehicle network, a unique identity identifier, such as the vehicle identification number (VIN) and manufacturer's signature, is written into the vehicle network's digital identity system. This unique identifier generates a corresponding digital certificate, which is stored in encrypted form on the blockchain and possesses the characteristics of being tamper-proof, traceable, and uniformly trustworthy. This avoids the problem of relying on easily forged features such as license plate numbers for verification.

[0037] Digital certificates are registered on a blockchain network and can be shared across multiple inspection areas without requiring duplicate registration or local verification logic. This improves the efficiency of identity reuse in cross-regional transportation scenarios and ensures high consistency. Once a vehicle is registered on the blockchain, the system can automatically verify the certificate's legitimacy without manual intervention.

[0038] Furthermore, it's understandable that verifying a digital certificate only requires extracting the certificate's public key, signature, and hash, and then checking the signature on a consensus node based on a blockchain consensus mechanism. This process is a typical non-interactive verification, requiring no back-and-forth communication and typically completing verification within milliseconds. If the digital certificate verification is successful and valid, the vehicle can be directly considered legitimate and registered. This unique, automated digital authentication is particularly suitable for edge nodes—that is, individual checkpoint nodes that can quickly invoke the verification module and perform millisecond-level verification.

[0039] In other embodiments of the present invention, considering the need to improve inspection efficiency and the fact that image acquisition and digital verification are not structurally conflicting, the acquisition of these two processes is performed in parallel. Specifically, the image acquisition module 100 begins acquiring images of the staff and the vehicle's exterior as soon as the vehicle enters the inspection area, while the vehicle digital authentication module 200 simultaneously attempts to acquire the vehicle's digital identity. If the digital authentication certificate is successfully verified, the vehicle image features are cached and not further processed for identification; they are simply archived. If digital authentication fails, the vehicle image features are then processed in the subsequent feature fusion analysis process as a basis for identification. The advantage of this approach is that if the digital authentication certificate is damaged or the network is temporarily interrupted, the vehicle image channel can serve as a fallback channel; even if the certificate verification is successful, the image can still be archived for audit playback, anti-fraud analysis, etc. Most importantly, these two processes are processed in parallel, without mutual waiting, and the system's response time remains within seconds, enabling rapid operation of the inspection system.

[0040] More specifically, when a vehicle enters a pre-defined inspection area, the vehicle digital authentication module 200 first sends a digital identity request through the vehicle's onboard communication terminal. A lightweight blockchain consensus mechanism, such as PBFT, is used to quickly verify this request. If the vehicle has already registered in the vehicle-to-everything (V2X) digital identity system and the digital identity authentication is successful, the system will extract its bound vehicle digital certificate (VDC) from the blockchain as a trusted identity identifier.

[0041] This combination Figure 3 The following steps illustrate S20: While collecting facial images, the vehicle is digitally authenticated using the consensus mechanism of the blockchain. If the digital identity authentication is successful, the corresponding digital authentication certificate is obtained; if the digital identity authentication fails, the vehicle image features, including the license plate number, vehicle model, and body color, are obtained.

[0042] When a vehicle enters the preset inspection area, the vehicle entry detection system can detect the vehicle's entry in real time using a geomagnetic sensor or a camera. At this time, the system activates the digital authentication module and sends an authentication request, that is, it initiates communication between the ETC inspection station and the vehicle's OBU, with the vehicle actively sending its digital authentication certificate.

[0043] The inspection area receives the certificate sent by the vehicle and verifies its digital signature and encryption integrity. The certificate content is then submitted to the blockchain network for consensus mechanism verification of its authenticity and validity, specifically checking the credibility of the certificate issuing authority and verifying whether the certificate has been revoked or expired. If the certificate verification is successful, the vehicle's identity is confirmed, and its digital certification is obtained for subsequent use.

[0044] If the vehicle entering the inspection area is not equipped with a vehicle-to-everything (V2X) system, then it is necessary to obtain the vehicle's characteristics through cameras or other means deployed within the inspection area.

[0045] Specifically, vehicle images are captured from cameras in the inspection area, and feature information, including license plate number, vehicle model, and body color, is extracted from them.

[0046] Digital authentication uses blockchain for rapid identity verification, while image feature capture provides a supplementary means for authentication failures, ensuring that every vehicle can be effectively verified. This ensures comprehensive identity verification for every vehicle, effectively identifying vehicles even if digital authentication fails and avoiding missed detections.

[0047] Based on the two data acquisition modules described above, it can be understood that these modules achieve real-time data collection of vehicles and personnel through automated processes, avoiding the inefficiency and inaccuracy of manual operations. By combining facial image acquisition with vehicle digital authentication, multi-dimensional data on vehicles and passengers is covered, ensuring the completeness and accuracy of verification. Furthermore, real-time encrypted transmission and storage of personnel and vehicle data protect privacy and prevent data leakage.

[0048] After collecting the information of vehicles and personnel, it is sent to the verification module 300. The verification module 300 of this system adopts a parallel dual-channel architecture design, which corresponds to the vehicle identity verification and in-vehicle personnel identity verification tasks respectively, to ensure that the system still has efficient recognition capabilities in a high-passage density information verification environment.

[0049] Specifically, when a vehicle enters the preset inspection area, if the vehicle has been connected to the vehicle-to-everything (V2X) digital identity system and completed registration, the system will receive the digital authentication certificate it carries. This certificate includes the vehicle's unique identifier, the certificate issuing authority's public key, the validity period, and the digital signature.

[0050] The first sub-channel in the 300-parallel dual-channel architecture of the verification module employs a hash signature verification mechanism. It utilizes blockchain nodes to retrieve the registered public key of the vehicle on the chain, performs hash calculations (e.g., SHA-256) on the certificate content, verifies the signature, and if the verification passes, the vehicle is considered legitimate. The system generates a Vehicle Identity Code (VID) for subsequent recording and scheduling. This first sub-channel's verification path boasts high processing efficiency, avoiding redundant image recognition calculations, thus enabling automated verification.

[0051] If the digital certificate is missing, expired, or fails to be verified, the system will automatically initiate the vehicle image recognition process to complete the vehicle identity verification.

[0052] Specifically, the vehicle exterior image is captured by calling the camera in the personnel image acquisition module 100, and preprocessing operations such as image enhancement, occlusion removal, and angle correction are performed after image acquisition. For example, in some embodiments of the present invention, YOLOv5 is used to detect the license plate frame, then the license plate number is extracted, the vehicle model is identified based on a lightweight convolutional neural network, and HSV color analysis and clustering are performed on the main color tone of the vehicle body. The extracted results are encoded into structured feature vectors, matched and identified with the vehicle database, and the legality judgment result is output. It can be understood that this process serves as an automatic supplementary path for certificate verification, ensuring the system's scene adaptability.

[0053] Regarding the identity verification of occupants, it is done through the second sub-channel in the parallel dual-channel architecture of the verification module 300. This channel is responsible for collecting and comparing the identity information of each passenger in the vehicle.

[0054] The second sub-channel is equipped with a facial recognition network to extract facial feature vectors corresponding to occupants of the vehicle. These vectors are then compared with the registered identity information database in the public security transportation supervision system. Those matching the comparison are marked as legitimate passengers. If the comparison results show individuals on a blacklist, untrustworthy list, or illegal personnel, abnormal personnel information is output, and the corresponding vehicle is marked as a high-risk vehicle, thus identifying the risk and triggering subsequent strategies. Furthermore, in some embodiments of this invention, this sub-channel supports simultaneous identification and processing of multiple individuals and has a supplementary data collection mechanism. That is, when data collection fails or the quality is substandard, the system automatically delays supplementary collection or switches the camera perspective to ensure the integrity and reliability of the verification data.

[0055] In summary, it is important to understand that the two sub-channels in the verification module 300 operate synchronously and in parallel. Therefore, in order to ensure the continuity of data reception and processing, the personnel image acquisition module 100, the vehicle digital authentication module 200, and the verification module 300 adopt a unified data exchange protocol. The verification module 300 receives the structured face image data output by the personnel image acquisition module 100 in real time through an internally set data bus, and also receives the digital authentication certificate and vehicle image feature data output by the vehicle digital authentication module 200 in real time.

[0056] For reference here Figure 4 The personnel image acquisition module 100 and the vehicle digital authentication module 200 serve as front-end information acquisition units, communicating via a unified lightweight communication protocol, such as... Figure 4 The diagram shows a protocol connection between the MQTT protocol and the multimodal feature verification module 300, ensuring the real-time performance and security of data during transmission.

[0057] After the vehicle enters the preset inspection area, these two acquisition modules complete their acquisition tasks and transmit the information to the verification module 300 in the form of structured data packets. In conjunction with the above, the personnel image acquisition module 100 uses the infrared camera and facial recognition camera deployed inside the vehicle to count the number of occupants in real time based on a lightweight model. It then extracts the feature vector of each occupant and encapsulates it into a data packet containing vehicle identification, timestamp, and facial vector fields. This packet is sent as input to the second sub-channel of the verification module 300 for personnel identity verification.

[0058] Meanwhile, the vehicle digital authentication module 200 will first attempt to query and verify the vehicle's digital identity certificate (VDC) through the blockchain network. If the authentication is successful, the extracted certificate information, including the DID identifier, public key, signature, and hash field, will be sent to the first sub-channel of the verification module 300. This sub-channel uses a hash signature verification mechanism to confirm the legitimacy of the vehicle's identity.

[0059] If authentication fails, the image recognition process is initiated. A high-definition front-facing camera system captures images of the vehicle's exterior, extracting image features such as the license plate number, vehicle model, and body color. These features are then structured, packaged into a data packet, and sent to the image feature analysis channel of verification module 300 as a supplementary verification path for vehicle identity. It's important to understand that the image feature analysis channel is a backup channel for the first sub-channel.

[0060] In summary, it can be understood that in the verification module 300, the first sub-channel and the image feature analysis channel form a primary and backup vehicle verification branch, while the second sub-channel independently undertakes the personnel verification task. The verification module 300 employs a parallel processing architecture, asynchronously verifying personnel and vehicle information separately, and ultimately summarizing the verification results. When an abnormal matching result occurs for either personnel or vehicle, the system marks the vehicle as a high-risk vehicle and synchronizes the processing instructions to the subsequent processing strategy module.

[0061] Combine the verification module 300 Figure 3 Let's understand S30. S30 involves: using the vehicle's digital certificate, vehicle image features, and facial images of occupants to independently verify both the vehicle and the occupants; if the digital certificate is valid, the vehicle's identity is directly confirmed and the verification result is recorded; if no digital certificate exists, the vehicle image features, including the license plate number, vehicle model, and body color, are fused and compared with records in the vehicle database to verify the vehicle's legitimacy; facial recognition technology is used to compare each facial image with identity information in the personnel database to verify if there are any flagged abnormal individuals; if at least one abnormal verification result is found, the vehicle is marked as a high-risk vehicle.

[0062] Specifically, when a vehicle enters the preset inspection area, refer to the following: Figure 5 The execution process shown is as follows: the verification module 300 first obtains the vehicle's digital authentication certificate and verifies it with the records of the blockchain network, checking information such as signature matching and certificate status (whether it has been revoked or expired). If the authentication is successful, the vehicle's identity is directly confirmed and the verification result is recorded, that is, the image feature verification is skipped.

[0063] For vehicles without digital authentication certificates, vehicle features such as license plate number, vehicle model, and body color are extracted from the vehicle images. A feature fusion algorithm is then used to match the multimodal features to the database to determine whether the vehicle is legitimate.

[0064] Next, facial features are extracted from the face image and compared with the registered personnel information in the personnel database to verify the person's identity. This is to check for any abnormal personnel.

[0065] Finally, a risk assessment is performed based on the two verification results. If both vehicle and personnel verifications pass, the vehicle is marked as normal. If any verification result is abnormal, the vehicle is marked as high-risk, and the specific abnormal information is recorded. The vehicle's risk level and verification results are then output for use by the subsequent traffic strategy module 400.

[0066] Continuing with the example above, a minivan enters an ETC checkpoint (preset inspection area). There are four passengers inside. The camera captures a panoramic image of the vehicle's interior, detecting four faces. The minivan provides a valid digital authentication certificate, and verification is successful. After comparing the facial images of the four passengers with the personnel database, no abnormalities are found. Therefore, the system confirms the vehicle is a legitimate vehicle, records the passage information, and allows it to pass.

[0067] Now, let's assume an older truck enters an ETC checkpoint without a digital certificate. The station's camera captures an image of the truck, extracting the license plate number "X-XYZ456," vehicle type "truck," and color "blue." This information is compared to the vehicle management office's database to verify consistency. However, there are two passengers inside, one of whom is blacklisted in the personnel database. Therefore, the system flags the vehicle as high-risk and notifies the subsequent processing module for manual inspection. This situation falls under the category of a vehicle without a certificate and with suspicious personnel.

[0068] Now, suppose a car enters an ETC checkpoint. The car's digital certificate verification fails, and the station's camera captures an image of the vehicle. The image feature extraction results show the license plate number "X-ABC123," the vehicle type "sedan," and the color "white," which do not match the "SUV" registered in the database. Therefore, the system will also mark this vehicle as high-risk. This situation involves a forged vehicle despite no suspicious individuals present.

[0069] In summary, it is understandable that by combining digital authentication with multimodal feature fusion verification, multi-dimensional features of vehicles and personnel are covered. Even if a vehicle lacks a digital authentication certificate, image feature verification can ensure data integrity, avoid verification gaps, accurately identify high-risk vehicles or abnormal personnel, mark them, and trigger subsequent warning and inspection processes, reducing potential security risks. Furthermore, the automated verification process shortens inspection time, reduces manual intervention, and improves inspection efficiency.

[0070] Finally, there's the passage strategy module 400. This module executes S40: when a high-risk vehicle is marked, it sends a warning signal to the adjacent preset inspection area and transmits the vehicle's digital authentication certificate, vehicle image features, and facial images of the occupants, triggering a manual inspection process. Essentially, the passage strategy module 400 dynamically determines the vehicle's handling strategy based on the verification results from the verification module 300, including allowing passage, issuing a warning, or conducting a manual inspection, to ensure both passage efficiency and safety.

[0071] Specifically, for vehicles that pass the verification by module 300, the system generates a release command, sends a release notification through the vehicle terminal, and updates the vehicle's passage status. For vehicles marked as high-risk, the system generates a warning message, notifies adjacent inspection areas, and triggers a manual inspection process to further confirm the risk.

[0072] More specifically, when the passage strategy module 400 receives the verification result (normal or high-risk) from the verification module 300, for normal vehicles, the system generates a release instruction, notifies the driver via the onboard terminal, allowing the vehicle to pass normally, and synchronizes the vehicle's passage status (such as time and location) to the central database. For vehicles marked as high-risk, the passage strategy module 400 generates a warning message, including detailed vehicle characteristics such as the extracted license plate number, vehicle type, body color, and the determined risk category, and synchronizes it to adjacent inspection areas and the traffic management platform, triggering a more refined manual inspection process, with the manual inspection results recorded in real time.

[0073] The key to the aforementioned early warning and linkage mechanism lies in the information interconnection and early warning linkage between various preset inspection areas. Specifically, when an area determines that a vehicle is a high-risk target, the system immediately pushes an early warning message to downstream or surrounding relevant inspection areas. The message content includes the license plate number, vehicle characteristics, and facial images of people inside the vehicle. Each area shares the vehicle's digital identity record and behavior log through a blockchain network, which can prevent tampering and ensure the traceability of verification data. Each authentication operation will generate a record on the chain, facilitating subsequent area verification of identity consistency or identification of abnormal behavior.

[0074] In some embodiments, the system can utilize an IoT architecture to synchronize detailed information about high-risk vehicles—including the extracted license plate number, vehicle model, body color, and determined risk category—in real time to adjacent inspection areas and the traffic management platform. Upon receiving the high-risk vehicle information, adjacent inspection areas prioritize monitoring the vehicle and further verify the risk. Finally, all processing results and verification records are stored in a central database for subsequent strategy optimization and risk prediction.

[0075] Now, suppose a verified private car receives a clearance instruction. The system notifies the driver via the onboard terminal: "Verification successful, please continue." Simultaneously, the central database records the vehicle's passage time, location, and verification result. This allows legitimate vehicles to pass quickly without human intervention, improving traffic efficiency.

[0076] In the example above, an older truck marked as a high-risk vehicle was found to be an anomaly during personnel verification. The system then sends a warning to adjacent inspection areas and the traffic management platform, extracting the license plate number "X-XYZ456," vehicle type "truck," color "blue," and the reason for the anomaly: "blacklisted person boarded the vehicle." Simultaneously, a manual inspection process is triggered, with personnel further verifying the truck's cargo and the identity of the driver. This ensures that high-risk vehicles are promptly intercepted and dealt with.

[0077] In summary, it can be understood that the inspection system in this embodiment of the invention can achieve high-precision identity verification and status recognition of vehicles and accompanying personnel without the need for manual intervention, improve the overall information processing efficiency and response speed of the system, reduce the impact on road traffic, and provide a reliable and traceable data chain to ensure that every release or warning has technical support and data basis.

[0078] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A non-contact road transport inspection system, characterized in that, include: The personnel image acquisition module counts the number of people inside the vehicle and acquires the facial images of the people inside the vehicle when the vehicle enters the preset inspection area. The vehicle digital authentication module uses a blockchain consensus mechanism to perform digital identity authentication on the vehicle when it enters the preset inspection area. If the digital identity authentication is successful, it obtains the digital authentication certificate. If the digital identity is not successfully authenticated, vehicle image features including license plate number, vehicle model, and body color will be obtained. The verification module includes a first sub-channel and a second sub-channel of a parallel processing architecture; wherein: The first sub-channel is used to receive the successfully authenticated digital certificate, and confirm the legal identity of the corresponding vehicle based on the hash signature verification method nested in the blockchain consensus mechanism, and generate the corresponding identity code; if the vehicle is not successfully authenticated, the vehicle image feature analysis channel is started, the multimodal vehicle image feature information is extracted using image depth features, uniformly encoded, and compared with the vehicle identity database before outputting the vehicle verification result. The second sub-channel uses its facial recognition network to extract the facial features of the people inside the vehicle, compares them with the registered identity information database, and outputs the personnel verification result; when there is at least one verification result marked as abnormal, the vehicle is marked as a high-risk vehicle. The passage strategy module sends a release command to a vehicle and synchronizes its passage status when the vehicle entering the preset inspection area is not marked as a high-risk vehicle by the verification module; when the vehicle is marked as a high-risk vehicle, the passage of the vehicle is blocked, triggering a manual re-inspection, and the vehicle's identification code, vehicle image features, and facial images of the people inside the vehicle are recorded and uploaded.

2. The non-contact road transport inspection system according to claim 1, characterized in that, When a vehicle enters the preset inspection area, the personnel image acquisition module counts the number of people in the vehicle using infrared sensors set up in the preset inspection area, and verifies the number of passengers in the vehicle by combining the camera and facial recognition detection algorithm. The camera in the personnel image acquisition module uses autofocus technology to capture facial images of people inside the vehicle. After image enhancement and facial localization, the captured facial image data is extracted, encrypted, and synchronized to the verification module.

3. The non-contact road transport inspection system according to claim 1, characterized in that, When a vehicle enters the preset inspection area, the vehicle digital authentication module initiates an identity authentication request to the vehicle via a communication terminal, and performs digital identity authentication of the vehicle using a blockchain consensus mechanism; specifically: If the vehicle has been registered and authenticated in the vehicle network digital identity system, its digital identity certificate registered on the blockchain is extracted as a credential of vehicle legitimacy, the vehicle identity is confirmed and the digital authentication result is recorded. If the vehicle identification fails or the vehicle is not registered, the vehicle image recognition process is triggered, and the following steps are performed: The camera set up in the preset inspection area captures the vehicle's exterior image, extracts vehicle image features including license plate number, vehicle model and body color, and synchronizes them to the verification module after encryption.

4. The non-contact road transport inspection system according to claim 1, characterized in that, The personnel image acquisition module, vehicle digital authentication module, and verification module use a unified data exchange protocol. The verification module receives structured facial image data output by the personnel image acquisition module in real time via an internally configured data bus, and receives digital authentication certificates and vehicle image feature data output by the vehicle digital authentication module in real time.

5. A non-contact road transport inspection system according to claim 1, characterized in that, The vehicle's digital certificate includes the vehicle's unique identifier, the certificate issuing authority's public key, the validity period, and a digital signature; After the verification module obtains the vehicle's digital certification certificate in real time, the first sub-channel first uses the built-in public key of the certificate issuing authority to perform hash signature verification on the certificate, and compares it with the issuance record and revocation record on the blockchain, and confirms that the certificate is currently in a valid state after comparing it with the valid time interval. If the hash signature verification passes and the certificate is currently valid, the vehicle's identity is confirmed as legitimate. The verification module, based on the hash signature structure nested in the blockchain consensus mechanism, combines the certificate hash with the number of the current query node to generate a unique identity code for the vehicle in this verification process.

6. A non-contact road transport inspection system according to claim 5, characterized in that, When the verification module obtains the vehicle image features, it indicates that the vehicle has not been successfully authenticated. The vehicle image feature analysis channel in the verification module is then activated. The feature fusion algorithm is used to uniformly encode the license plate number, vehicle model, and body color information into a vehicle feature vector. This vector is then compared with the vehicle identity database, and the similarity threshold matching result is output to verify the legality of the vehicle identity.

7. A non-contact road transport inspection system according to claim 4, characterized in that, When the verification module acquires the facial images of the people inside the vehicle in real time, a facial recognition network is set up in the second sub-channel to extract the facial feature vectors corresponding to the people inside the vehicle. The facial feature vectors of the people inside the vehicle are compared with the identity information database registered in the public security transportation supervision system. People whose comparison results match are marked as legitimate passengers. If the comparison results show blacklist, blacklist, or illegal personnel, abnormal personnel information is output and the corresponding vehicle is marked as a high-risk vehicle.

8. A non-contact road transport inspection system according to claim 1, characterized in that, When a vehicle entering the preset inspection area is not identified as a high-risk vehicle by the verification module, the passage strategy module sends a release instruction to the vehicle and simultaneously uploads the vehicle's passage status, including the verification time node, identity code, and passage result, to the central database. When a vehicle entering the preset inspection area is verified as a high-risk vehicle by the module, the passage strategy module sends a delayed release signal to block the vehicle from passing, triggers a manual interception and re-inspection process, and pushes a high-risk passage warning to adjacent inspection areas, including sending the vehicle's identification code or vehicle image features, as well as facial images of the people inside the vehicle.

9. A non-contact road transport inspection method, characterized in that, Includes the following steps: Multiple preset inspection areas are set up. When a vehicle enters any of the preset inspection areas, the number of people in the vehicle is counted and the facial images of the people in the vehicle are collected. While collecting the facial images, the vehicle is digitally authenticated using a blockchain consensus mechanism. If the digital identity authentication is successful, its digital authentication certificate is obtained. If the digital identity is not successfully authenticated, vehicle image features including license plate number, vehicle model, and body color will be obtained. Upon receiving the successfully authenticated digital certificate, the system confirms the legitimate identity of the corresponding vehicle based on the hash signature verification method nested in the blockchain consensus mechanism and generates the corresponding identity code. If the vehicle fails to authenticate, the multimodal vehicle image feature information is extracted using image depth features, uniformly encoded, and compared with the vehicle identity database to output the vehicle verification result; the facial features corresponding to the occupants are extracted, compared with the registered identity information database, and the occupant verification result is output; when there is at least one verification result marked as abnormal, the vehicle is marked as a high-risk vehicle. When a vehicle entering the preset inspection area is not marked as a high-risk vehicle, a release instruction is sent to the vehicle and its passage status is synchronized; when it is marked as a high-risk vehicle, the vehicle's passage is blocked, triggering a manual re-inspection, and the vehicle's identification code, vehicle image features, and facial images of the people inside the vehicle are recorded and uploaded.

Images