Motor vehicle wrong fuel adding prevention early warning method and system based on image recognition and insurance big data

Abstract

The invention discloses an image recognition and insurance big data-based motor vehicle wrong fuel adding prevention early warning method and system, and relates to the intelligent traffic and gas station safety control cross technical field, and the system comprises a cloud processor (1), a camera module (2), a storage arithmetic unit (3) and a voice player (4); the camera module (2) is connected with the input end of the storage arithmetic unit (3), the output end of the storage arithmetic unit (3) is connected with the voice player (4), and the camera module and the voice player are in two-way communication with the cloud processor (1) through a wired or wireless network; the cloud processor (1) stores a high-risk vehicle feature database constructed by claim settlement data of an insurance company. The camera module (2) collects vehicle images at two positions, the storage arithmetic unit (3) extracts features, uploads the features for matching, pre-generates an instruction after high risk is judged, and triggers voice early warning after secondary verification is correct. According to the system, manual and physical protection blind areas are made up by means of big data accurate identification and active pre-warning before refueling, fuel mistake accidents are reduced from the source, insurance risk reduction is achieved, the system structure is simple, and the deployment cost is low.

Description

A method and system for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data. Technical Field

[0001] This invention relates to the field of intelligent transportation and gas station safety management, specifically to a method and system for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data. Background Technology

[0002] According to statistics, as of the end of September 2025, my country's motor vehicle ownership had exceeded 465 million, with over 6 billion refueling trips per year. Gas stations are facing increasing pressure in ensuring safe fuel supply, leading to frequent incidents of incorrect fuel being dispensed. These incidents primarily involve three core issues: gasoline engines being mistakenly filled with diesel, diesel engines being mistakenly filled with gasoline, and gasoline vehicles being filled with fuel of the wrong octane rating. Submitting the wrong fuel can directly damage a vehicle's engine and fuel system, with repair costs reaching tens of thousands of yuan. It can also cause disputes between car owners, gas stations, and insurance companies, delay normal vehicle use, and increase gas station complaint rates and insurance company claims costs.

[0003] Currently, domestic gas stations mainly rely on two traditional methods to prevent the misfilling of fuel: First, gas station attendants manually ask drivers what type of fuel their vehicles require or check the fuel tank cap markings. This method depends entirely on human responsibility and experience, and is prone to oversight, forgetfulness, or communication errors, especially during peak refueling hours when traffic is heavy, significantly increasing the probability of misjudgment. Second, they use a physical nozzle interface design that differentiates between gasoline and diesel fillers by nozzle diameter. This method can only prevent the misfilling of certain types of fuel for some vehicle models, and cannot effectively prevent the misfilling of different grades of gasoline of the same type or the misfilling of certain high-risk vehicle models. Furthermore, some older or modified vehicles can still bypass the physical interface limitations to complete the misfilling, indicating a significant limitation in the prevention effect.

[0004] Among existing technologies, some gas stations use license plate recognition to bind refueling information and fuel cards to fuel types for control. These solutions can only achieve the binding and control of fixed vehicles and fixed fuel types, and cannot identify temporary vehicles entering the station or social vehicles without binding information. Furthermore, they do not combine actual data on incorrect fuel refueling incidents for risk classification, resulting in a lack of targeted early warning. In addition, a few image recognition-based refueling assistance systems only achieve basic vehicle model recognition and do not rely on the massive historical claims data of the insurance industry to mine patterns of high-risk vehicle models. They cannot dynamically update the risk database, resulting in low accuracy of early warnings and a high false alarm rate, making it difficult to achieve risk prevention and control at the source.

[0005] In summary, existing technologies generally suffer from technical shortcomings such as the inability to proactively identify high-risk vehicles, lagging early warning mechanisms, insufficient data support, lack of dynamic updates, and high false alarm rates. These shortcomings fail to meet the needs of efficient and safe operation at gas stations and hinder cooperation with insurance companies in managing the risk of incorrect fuel dispensing. Therefore, developing a system for preventing incorrect fuel dispensing based on real-time image recognition, leveraging insurance big data to accurately locate high-risk vehicle models, providing proactive early warnings, and adapting to all vehicles entering gas stations has become a pressing technical challenge in this field. Summary of the Invention

[0006] To address the aforementioned issues, this invention provides a method and system for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data. It innovates and improves upon these issues from two levels: hardware system architecture and methodological process control. Through a dual architecture of local real-time image acquisition and recognition combined with cloud-based insurance big data linkage and matching, coupled with a tiered warning and secondary verification mechanism, it achieves automatic identification of high-risk vehicle models entering the station, proactive voice warnings, and cross-station synchronous database updates. This reduces the incidence of incorrect fuel filling accidents from the source, while also minimizing multi-party disputes, achieving a dual benefit of improved gas station operation quality and reduced insurance risks.

[0007] To achieve the above objectives, the present invention provides the following technical solution.

[0008] A method and system for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data includes: a cloud processor (1), a camera module (2), a storage and processing unit (3), and a voice player (4).

[0009] The storage processor (3) is connected to the camera module (2) and the voice player (4) via a data cable. At the same time, the storage processor (3) establishes a two-way communication link with the cloud processor (1) through a wireless or wired communication network to realize real-time interaction between local image acquisition data and cloud data, and return of risk judgment instructions. The cloud processor (1) is deployed on a cloud server and has a built-in data storage unit. Its core is used to store and maintain a database of characteristics of vehicles with high risk of incorrect fuel filling. This database is built based on the statistical analysis of historical fuel filling claims data from insurance companies. It includes the brand, model and corresponding fuel information of vehicles with high incidence of incorrect fuel filling. It can be dynamically updated according to the real-time new claims data from insurance companies to ensure the timeliness and accuracy of the warning data. The cloud processor (1) can receive vehicle feature query requests sent by the storage processor (3) in real time, quickly complete data matching and comparison, and synchronously return the risk judgment results to the local storage processor (3).

[0010] The camera module (2) adopts a high-definition explosion-proof image acquisition component, which is mainly installed at the entrance of the gas station to collect images of the appearance of vehicles waiting to be refueled in the gas station area. The camera module (2) includes a high-definition camera and an image sensor, which can clearly capture the core features such as brand and model identification on the front and rear of the vehicle. It is also equipped with an adaptive supplementary lighting device to automatically supplement and brighten the image in low light environments such as night and cloudy days, ensuring the clarity of image acquisition in different environments.

[0011] The storage processor (3) is a local industrial control processing unit, which is fixedly installed inside the control room of the gas station. It has a built-in mature image recognition processing unit. This unit is subdivided into a vehicle brand recognition module and a model recognition module from a functional perspective. The brand recognition module completes the accurate identification of the brand based on the fixed features of the vehicle's front grille and logo. The model recognition module completes the specific model lock based on the overall appearance outline of the vehicle and the model identification features of the rear. It can quickly and finely analyze the vehicle images collected by the camera module (2), extract the core features of the vehicle brand and model, and upload them to the cloud processor (1) to complete big data matching. At the same time, it receives the risk judgment results transmitted back from the cloud, generates voice broadcast instructions for high-risk models, and does not output instructions for non-high-risk models to avoid invalid warnings interfering with the normal operation order of the gas station.

[0012] The voice player (4) is fixedly installed in a prominent and easily audible position on the pillar of the gas station canopy to ensure that gas station attendants and car owners can clearly receive the prompt voice; after receiving the control command issued by the storage and calculation unit (3), it immediately plays a targeted voice prompt for the risk of adding the wrong oil, so as to realize the proactive reminder in advance and avoid the risk of adding the wrong oil from the source.

[0013] Furthermore, the cloud processor (1) can be networked and linked with the storage processors (3) of multiple gas stations to update the high-risk vehicle feature database in real time based on the newly added wrong oil claim data of the insurance company, and push the updated database to all networked stations in a synchronized manner to realize cross-site sharing of high-risk vehicle risk information.

[0014] A method and system for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data, characterized by the following steps:

[0015] The first step is vehicle image fixed-point acquisition: when a vehicle to be refueled enters the preset entrance control area of ​​the gas station, the camera module (2) at the entrance is triggered to start. The matching adaptive supplementary lighting device automatically adjusts the brightness according to the ambient light and acquires a complete and clear image of the vehicle's appearance in real time. After the acquisition is completed, the image data is synchronously transmitted to the local storage and processing unit (3) to complete the local image acquisition process.

[0016] The second step is to refine the vehicle features: After receiving the image data, the storage processor (3) calls the built-in image recognition processing unit. Through the division of labor and cooperation between the brand recognition module and the model recognition module, the vehicle image is extracted in multiple dimensions, irrelevant background interference information is removed, and the unique brand and model information of the vehicle are accurately locked, thus completing the digital extraction and preprocessing of vehicle features.

[0017] The third step is cloud-based big data matching and judgment: the storage processor (3) transmits the pre-processed vehicle brand and model feature data to the cloud processor (1) through a wireless or wired communication network; the cloud processor (1) calls the high-risk vehicle feature database in the data storage unit, compares the received feature data with the information in the database one-to-one, and quickly determines whether the target vehicle belongs to the high-risk model of the wrong fuel. After the judgment is completed, the result is sent back to the local storage processor (3) in real time.

[0018] The fourth step is to execute the graded warning command: After receiving the cloud judgment result, the storage arithmetic unit (3) executes the graded warning logic; if it is determined to be a high-risk vehicle, it immediately generates the corresponding voice broadcast control command and does not trigger the broadcast for the time being, and waits for the subsequent secondary verification to pass before issuing it; if it is determined to be a non-high-risk vehicle, it directly terminates the subsequent command output, and the whole system remains in a silent standby state, without interfering with the normal refueling operation process of the gas station.

[0019] Step 5, Secondary verification and data synchronization: After the vehicle stops at the designated work station on the refueling island, the camera module (2) installed above the refueling island starts secondary image acquisition. The storage and processing unit (3) cross-compares and verifies the two feature recognition results collected at the entrance and at the refueling island. If the verification is consistent, the voice broadcast command pre-generated in step 4 is issued to officially trigger the final voice warning. If the verification is inconsistent, the pre-generated command is invalidated to prevent false alarms. At the same time, the cloud processor (1) synchronizes the latest claims data of the insurance company in the background, updates the database of high-risk vehicles, and sends the updated content to all networked gas stations to ensure that the warning data of the wireless or wired communication network is unified and accurate.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] 1. Proactive early warning: The vehicle image is automatically collected and the features are identified by the camera module (2). Combined with insurance big data, the risk is accurately determined. The risk can be proactively warned before refueling, making up for the blind spots and deficiencies of manual inquiry and physical fuel gun protection, and the warning efficiency is greatly improved.

[0022] 2. Data-driven, precise early warning with an extremely low false alarm rate: The system's core high-risk vehicle database comes from real historical claims data from insurance companies, accurately identifying high-risk vehicle models and corresponding error-prone fuels. Combined with a dual-module identification and secondary verification mechanism, the early warning is highly targeted, effectively compensating for the lack of experience among gas station attendants and preventing the occurrence of fuel-incorrect incidents from the root.

[0023] 3. The database is dynamically updated. The cloud processor (1) can connect to the network to synchronize the latest high-risk vehicle data, realize cross-regional risk information sharing, and quickly include newly added high-risk vehicles in the early warning scope to maintain the effectiveness of the plan;

[0024] 4. Significant benefits to all parties: By intervening in advance to reduce the occurrence of accidents, insurance risks are reduced, insurance companies' claims costs are lowered, and disputes and complaints at gas stations are reduced, thereby improving operational safety.

[0025] 5. The system has a simple structure and is easy to deploy. It adopts a combination of mature technologies and can be installed and upgraded at low cost on the basis of existing gas stations without large-scale site changes. It is compatible with various types of gas stations and is easy to promote and apply on a large scale. Attached Figure Description

[0026] Figure 1 is a schematic diagram of the structure of the present invention.

[0027] Among them: (1) cloud processor; (2) camera module; (3) storage processor; (4) voice player. Detailed Implementation

[0028] The present invention will be further described below with reference to Figure 1. It should be noted that the present invention does not improve the existing image recognition algorithm itself. It can be implemented by using mature vehicle image recognition technology in this field. The overall system and method have complete practicality and engineering feasibility.

[0029] As shown in Figure 1, the hardware architecture of the system of the present invention is as follows: the cloud processor (1) is deployed on the cloud server and establishes stable two-way communication with the storage processor (3) distributed in each gas station through a wireless or wired communication network; the camera module (2) is installed at the gas station entrance and the canopy above the gas station island respectively, and the signal output end is connected to the signal input end of the storage processor (3) through a data cable; the control output end of the storage processor (3) is connected to the voice player (4) through a data cable to complete signal transmission and command control.

[0030] The collaborative workflow of the method and system of this invention: When a vehicle to be refueled enters the entrance control area of ​​the gas station, the camera module (2) collects a clear image of the vehicle's appearance in real time with the assistance of an adaptive supplementary lighting device, and transmits the image data synchronously to the storage processor (3); the storage processor (3) analyzes the image data through its built-in image recognition processing unit, extracts the core feature information of the brand and model, and uploads it to the cloud processor (1) with encryption; after receiving the feature information, the cloud processor (1) quickly matches it with the internal high-risk vehicle feature database to determine whether the vehicle is a high-risk model for refueling the wrong vehicle, and transmits the matching result back to the storage processor (3) in real time. If the returned result is a high-risk vehicle, the storage processor (3) does not directly trigger an early warning. After the vehicle stops at the refueling position, the camera module (2) above the refueling island collects the image a second time to complete the verification. After the verification is consistent, a voice broadcast command is generated to control the voice player (4) to play the preset risk warning voice, reminding the gas station attendant and the car owner to repeatedly check the fuel; if the vehicle is not a high-risk vehicle, the system remains silent throughout the process and does not interfere with the normal refueling process of the gas station.

[0031] As a preferred embodiment, the prompt voice content of the voice player (4) can be customized according to the actual operating needs of the gas station, such as "Please note that this model is prone to being filled with the wrong diesel, please check the fuel filler cap label" and "According to the insurance data, it is recommended to fill this vehicle with the corresponding grade of gasoline", adapting to the personalized usage scenarios of different gas stations.

Claims

1. A motor vehicle fuel filling error prevention early warning system based on image recognition and insurance big data, characterized in that: The system includes a cloud processor (1), a camera module (2), a storage processor (3), and a voice player (4). The output of the camera module (2) is connected to the input of the storage processor (3) via a data cable, and the output of the storage processor (3) is connected to the input of the voice player (4) via a data cable. The storage processor (3) establishes a bidirectional communication connection with the cloud processor (1) via a wired or wireless communication network. The cloud processor (1) uses cloud server hardware and has a built-in dedicated data storage unit for storing and maintaining a database of characteristics of vehicles with incorrect fuel filling. This database is based on historical fuel filling claims from insurance companies. The data is updated dynamically on a regular basis; the camera module (2) is used to collect images of the appearance of vehicles waiting to be refueled in the gas station area in real time; the storage processor (3) is a local industrial control processing unit with a built-in image recognition processing unit, which is used to analyze the vehicle images collected by the camera module (2) in real time, extract the vehicle brand and model feature information, transmit the feature information to the cloud processor (1) to complete the matching comparison, and generate the corresponding voice control command according to the matching result returned by the cloud processor (1); the voice player (4) is used to receive the control command issued by the storage processor (3) and play the voice warning of the risk of refueling the wrong vehicle for high-risk vehicles.

2. The motor vehicle fuel filling error prevention early warning system based on image recognition and insurance big data according to claim 1, characterized in that: The camera module (2) is installed at the entrance of the gas station and is equipped with a supplementary lighting device for automatic supplementary lighting and brightening in nighttime and low-light environments to ensure the clarity of vehicle image acquisition; the storage and processing unit (3) is installed inside the control room of the gas station, and the voice player (4) is fixedly installed in a conspicuous and easily listenable position on the pillar of the gas station canopy.

3. The motor vehicle fuel filling error warning system based on image recognition and insurance big data according to claim 1, characterized in that: The camera module (2) is also installed in the canopy above the refueling island to collect vehicle images a second time after the vehicle stops at the refueling position. The storage and processing unit (3) is equipped with a comparison and verification unit to compare and verify the two recognition results collected at the entrance and the refueling island. The voice warning is triggered only after the verification is consistent to avoid misidentification and misbroadcasting.

4. The motor vehicle fuel filling error prevention early warning system based on image recognition and insurance big data according to claim 1, characterized in that: The cloud processor (1) is networked with the storage and computing units (3) of multiple gas stations, and can dynamically update the high-risk vehicle feature database based on the newly added wrong oil claim data of the insurance company, and push the updated database to each networked gas station in a synchronized manner, so as to realize the sharing of risk information of high-risk models across stations.

5. A method for preventing incorrect fuel filling in motor vehicles based on image recognition and insurance big data, applied to the early warning system described in any one of claims 1 to 4, characterized in that, Includes the following steps: When a vehicle waiting to refuel enters the gas station entrance control area, the camera module (2) at the entrance captures an image of the vehicle's exterior with the assistance of a supplementary lighting device, and transmits the image data to the storage processor (3). The storage processor (3) performs refined analysis of the vehicle image through its built-in image recognition processing unit, removes interference information, and extracts the core features of the vehicle's brand and model. The storage processor (3) encrypts and uploads the feature information to the cloud processor (1) via a wired or wireless communication network. The cloud processor (1) matches and compares the received feature information with its internal high-risk vehicle feature database to determine whether the vehicle is a high-risk vehicle that has been refueled incorrectly. The system identifies the vehicle model and sends the results back to the storage processor (3). If the vehicle is identified as a high-risk vehicle, the storage processor (3) pre-generates a voice broadcast command but does not trigger it. After the vehicle stops at the refueling island, the system collects the vehicle image a second time through the camera module (2) above the refueling island. The storage processor (3) cross-verifies the two identification results. If the verification is consistent, the pre-generated command is issued and the voice warning is activated. If the verification is inconsistent, the command is invalidated. If the vehicle is identified as a non-high-risk vehicle, the system remains silent. The cloud processor (1) periodically synchronizes the latest claims data from the insurance company, updates the high-risk vehicle feature database, and synchronizes it to each networked gas station.

Images