Parking identification method, device, system and equipment and storage medium

Abstract

The invention relates to a parking identification method, device and system, equipment and a storage medium. The method comprises the following steps: acquiring a vehicle event output by monitoring equipment; if the vehicle event is a leaving event that the target vehicle leaves the parking lot, vehicle information generated in the parking process of the target vehicle and parking lot information of the parking lot are acquired; fusing the vehicle information and the parking lot information into a preset instruction file to obtain a target instruction file; wherein the preset instruction file comprises a plurality of preset parking analysis instructions; the target instruction file is submitted to a big language model, the big language model is acquired to analyze the vehicle information and the parking lot information according to each parking analysis instruction in the target instruction file, a parking result of the target vehicle is obtained, and the parking result at least comprises to-be-paid fee information corresponding to the target vehicle. The method can improve the efficiency of determining the to-be-paid fee for vehicle parking.

Description

Technical Field

[0001] This application relates to the field of parking management technology, and in particular to a parking identification method, device, system, equipment and storage medium. Background Technology

[0002] With the continuous improvement of my country's economic development level, the urban population has surged, and the number of motor vehicles has risen rapidly. As a result, parking spaces have been set up on many roads.

[0003] In related technologies, inspectors scan parking spaces one by one along the parking route to obtain information on the parking status of each space, thereby determining the fees to be paid for the vehicles.

[0004] However, the method of determining vehicle fees based on manual scanning results using relevant technologies is inefficient. Summary of the Invention

[0005] Therefore, it is necessary to provide a parking identification method, device, system, equipment, and storage medium that can improve the efficiency of determining the parking fees to be paid for vehicles, in order to address the above-mentioned technical problems.

[0006] Firstly, this application provides a parking recognition method. The method includes:

[0007] Acquire vehicle events output by the monitoring equipment;

[0008] If the vehicle event is a departure event of the target vehicle leaving the parking lot, then obtain the vehicle information generated during the parking process of the target vehicle and the parking lot information.

[0009] Vehicle information and parking information are integrated into a preset instruction file to obtain a target instruction file; the preset instruction file includes multiple preset parking analysis instructions.

[0010] The target instruction file is submitted to the large language model. The large language model analyzes the vehicle information and parking lot information based on the parking analysis instructions in the target instruction file to obtain the parking result of the target vehicle. The parking result includes at least the information of the outstanding fees corresponding to the target vehicle.

[0011] In one embodiment, there are multiple vehicle information and parking lot information; the vehicle information and parking lot information are merged into a preset instruction file to obtain a target instruction file, including:

[0012] Based on the type of each vehicle information, fill each vehicle information into the first information position in the preset instruction file;

[0013] Based on the type of information for each parking lot, fill the second information position in the preset instruction file with the information for each parking lot;

[0014] Use the filled-in preset instruction file as the target instruction file.

[0015] In one embodiment, obtaining vehicle information includes:

[0016] The system acquires multiple first vehicle information and vehicle images output by the monitoring equipment; wherein each first vehicle information is acquired by the monitoring equipment when it detects a target vehicle entering the parking lot.

[0017] Based on the vehicle image, determine the license plate number and vehicle characteristics of the target vehicle;

[0018] Based on the target vehicle's license plate number, retrieve the target vehicle's historical parking information from a pre-set database;

[0019] Use at least one of the following as vehicle information: first vehicle information, target vehicle license plate number, vehicle characteristics, and target vehicle historical parking information.

[0020] In one embodiment, acquiring multiple first vehicle information items and vehicle images output by the monitoring device includes:

[0021] If the vehicle event is a departure event where the target vehicle leaves the parking lot, the entry event associated with the target vehicle's license plate number is determined from multiple candidate entry events based on the license plate number of the target vehicle associated with the departure event.

[0022] Obtain information and images of each first vehicle associated with the entry event.

[0023] In one embodiment, obtaining parking lot information includes:

[0024] Obtain basic parking information from a pre-set knowledge base; parking information includes at least one of the following: speed limit information of the road where the parking lot is located, lane information of the road where the parking lot is located, congestion time information of the road where the parking lot is located, current number of cars parked in the parking lot, number of parking spaces in the parking lot, and parking fee rules;

[0025] Vehicle images are identified and analyzed to determine pedestrian and vehicle traffic information at the location of the parking lot on the road.

[0026] The basic information of the parking lot, pedestrian flow information, and vehicle flow information are used as parking lot information.

[0027] In one embodiment, the parking result also includes the probability that the target vehicle is parked in the parking lot; wherein, the parking analysis instruction includes a parking probability analysis instruction, and the large language model determines the probability that the target vehicle is parked in the parking lot based on the vehicle information under the instruction of the parking probability analysis instruction.

[0028] Secondly, this application also provides a parking recognition device. The device includes:

[0029] The event acquisition module is used to acquire vehicle events output by the monitoring equipment;

[0030] The information acquisition module is used to acquire vehicle information generated during the parking process of the target vehicle and parking information of the parking lot if the vehicle event is a departure event of the target vehicle leaving the parking lot.

[0031] The information fusion module is used to fuse the vehicle information and the parking lot information into a preset instruction file to obtain a target instruction file; wherein, the preset instruction file includes multiple preset parking analysis instructions;

[0032] The result analysis module is used to submit the target instruction file to the large language model, obtain the parking result of the target vehicle by the large language model analyzing the vehicle information and the parking lot information according to each parking analysis instruction in the target instruction file, and the parking result includes at least the payment information of the target vehicle.

[0033] In one embodiment, there are multiple vehicle information and parking lot information; the information fusion module is specifically used for:

[0034] Based on the type of each vehicle information, fill each vehicle information into the first information position in the preset instruction file;

[0035] Based on the type of information for each parking lot, fill the second information position in the preset instruction file with the information for each parking lot;

[0036] Use the filled-in preset instruction file as the target instruction file.

[0037] In one embodiment, the information acquisition module is specifically used for:

[0038] The system acquires multiple first vehicle information and vehicle images output by the monitoring equipment; wherein each first vehicle information is acquired by the monitoring equipment when it detects a target vehicle entering the parking lot.

[0039] Based on the vehicle image, determine the license plate number and vehicle characteristics of the target vehicle;

[0040] Based on the target vehicle's license plate number, retrieve the target vehicle's historical parking information from a pre-set database;

[0041] Use at least one of the following as vehicle information: first vehicle information, target vehicle license plate number, vehicle characteristics, and target vehicle historical parking information.

[0042] In one embodiment, the information acquisition module is specifically used for:

[0043] If the vehicle event is a departure event where the target vehicle leaves the parking lot, the entry event associated with the target vehicle's license plate number is determined from multiple candidate entry events based on the license plate number of the target vehicle associated with the departure event.

[0044] Obtain information and images of each first vehicle associated with the entry event.

[0045] In one embodiment, the information acquisition module is specifically used for:

[0046] Obtain basic parking information from a pre-set knowledge base; parking information includes at least one of the following: speed limit information of the road where the parking lot is located, lane information of the road where the parking lot is located, congestion time information of the road where the parking lot is located, current number of cars parked in the parking lot, number of parking spaces in the parking lot, and parking fee rules;

[0047] Vehicle images are identified and analyzed to determine pedestrian and vehicle traffic information at the location of the parking lot on the road.

[0048] The basic information of the parking lot, pedestrian flow information, and vehicle flow information are used as parking lot information.

[0049] In one embodiment, the parking result also includes the probability that the target vehicle is parked in the parking lot; wherein, the parking analysis instruction includes a parking probability analysis instruction, and the large language model determines the probability that the target vehicle is parked in the parking lot based on the vehicle information under the instruction of the parking probability analysis instruction.

[0050] Thirdly, this application also provides a computer device, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described in any of the first aspects above.

[0051] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects above.

[0052] Fifthly, this application also provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of the method described in any one of the first aspects above.

[0053] The aforementioned parking identification method, device, system, equipment, and storage medium acquire vehicle events output by monitoring equipment. If the vehicle event is a departure event of a target vehicle leaving the parking lot, it acquires vehicle information generated during the target vehicle's parking process and parking lot information. The vehicle information and parking lot information are then integrated into a preset instruction file to obtain a target instruction file. This preset instruction file includes multiple preset parking analysis instructions. The target instruction file is submitted to a large language model, which analyzes the vehicle and parking lot information based on the parking analysis instructions in the target instruction file to obtain the parking result for the target vehicle. The parking result includes at least the outstanding payment information for the target vehicle. In this way, by automatically capturing vehicle information using monitoring equipment and then using a large language model to perform multi-dimensional data analysis on the vehicle and parking lot information, parking results are quickly obtained, enabling automatic license plate recording and billing. Compared to manual scanning, the efficiency of obtaining parking results is greatly improved. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a diagram illustrating the application environment of the parking recognition method in one embodiment;

[0056] Figure 2 This is a flowchart illustrating a parking recognition method in one embodiment;

[0057] Figure 3 This is a flowchart illustrating the process of obtaining the target instruction file in one embodiment;

[0058] Figure 4 This is a schematic diagram of the process for obtaining vehicle information in one embodiment;

[0059] Figure 5 This is a schematic diagram of the process for obtaining parking information in one embodiment;

[0060] Figure 6 This is a schematic diagram of the structure of an edge computing device in one embodiment;

[0061] Figure 7 This is a schematic diagram of the parking recognition process in one embodiment;

[0062] Figure 8 This is a structural block diagram of a parking recognition device in one embodiment;

[0063] Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0064] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that many specific details are set forth in the following description in order to provide a full understanding of this application, but this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0066] With the continuous improvement of my country's economic development level, the urban population has surged, and the number of motor vehicles has risen rapidly. As a result, parking spaces have been set up on many roads.

[0067] In related technologies, inspectors scan parking spaces one by one along the parking route to obtain information on the parking status of each space, thereby determining the fees to be paid for the vehicles.

[0068] For example, inspectors scan parking spaces by bicycle, take photos with handheld PPS terminals, and manually enter the license plate and start / end time. A single inspector can only cover a maximum of 40 parking spaces per hour, with a missed rate of ≥15%, and efficiency drops by 50% in rainy or snowy weather. Furthermore, locations with installed geomagnetic sensors only provide "car present / absent" pulses, requiring additional RFID or PDAs to record license plates. The hardware cost per parking space is ≥800 yuan, and batteries need to be replaced every two years, with maintenance costs accounting for 25% of the total CAPEX. Rate calculations use a fixed tiered table published by the municipality, unrelated to real-time availability, traffic speed, weather, or other dynamic variables, resulting in zero price elasticity during peak congestion periods and a consistently high arrears rate (above 8%). Toll evasion audits rely on inspectors visually comparing "whitelists," with the central platform performing batch matching afterward. The average audit cycle is 7 days, making on-the-spot interception impossible, resulting in a malicious arrears recovery rate of <60%. Heterogeneous interfaces across subsystems and hardware stacking add an extra 15% energy consumption and 90ms network jitter, limiting rapid city-wide replication.

[0069] Therefore, the method of determining vehicle fees based on manual scanning results using relevant technologies is inefficient.

[0070] In view of this, this application provides a parking identification method, device, system, equipment, and storage medium. The method can acquire vehicle events output by monitoring equipment; if the vehicle event is a departure event of a target vehicle leaving a parking lot, it acquires vehicle information generated during the target vehicle's parking process and parking lot information; it integrates the vehicle information and parking lot information into a preset instruction file to obtain a target instruction file; wherein the preset instruction file includes multiple preset parking analysis instructions; it submits the target instruction file to a large language model, and obtains the parking result of the target vehicle by analyzing the vehicle information and parking lot information according to each parking analysis instruction in the target instruction file. The parking result includes at least the outstanding payment information corresponding to the target vehicle. In this way, by automatically capturing vehicle information using monitoring equipment and then using a large language model to perform multi-dimensional data comprehensive analysis of the vehicle information and parking lot information, the parking result can be quickly obtained, achieving automatic license plate recording and billing. Compared with manual scanning, the efficiency of obtaining parking results is greatly improved.

[0071] The parking recognition method provided in this application can be applied to, for example... Figure 1 The application environment is shown. In this environment, the edge computing device and the monitoring device can communicate. The data storage system can store the data that the edge computing device needs to process. The data storage system can be integrated on a server, or it can be located in the cloud or on other network servers. The edge computing device is used to execute the parking recognition method. The edge computing device can be a personal computer, laptop, smartphone, tablet, or server, etc. The server can be a standalone server or a server cluster composed of multiple servers. There can be two monitoring devices, which can be integrated radar-view cameras. One monitoring device is set at the entrance of the parking lot to capture vehicle information and generate a vehicle entry event when a vehicle enters the parking lot; the other monitoring device is set at the exit of the parking lot to capture vehicle information and generate a vehicle departure event when a vehicle leaves the parking lot. The parking lot can be, for example, a street parking lot with multiple standardized parking spaces along the street, or other types of parking lots.

[0072] In one embodiment, such as Figure 2 As shown, a parking recognition method is provided, which can be applied to... Figure 1 Taking edge computing devices as an example, the explanation includes the following steps:

[0073] Step 201: Obtain vehicle events output by the monitoring equipment.

[0074] As mentioned above, a monitoring device is installed at both the parking lot entrance and exit. These devices monitor in real time whether vehicles are entering or leaving the parking lot. Specifically, the vehicle event generated when a vehicle enters is called an "entry event," and the vehicle event generated when a vehicle leaves is called a "departure event."

[0075] Optionally, the edge computing device includes a license plate recording service module, an edge intelligent agent module, a post-processing service module, a RAG knowledge base module, and an MCP tool module. Each module can be implemented as part of the edge computing device through software, hardware, or a combination of both. Optionally, the license plate recording service module may receive vehicle events output by the monitoring equipment.

[0076] Step 202: If the vehicle event is a departure event of the target vehicle leaving the parking lot, then obtain the vehicle information generated during the parking process of the target vehicle and the parking lot information.

[0077] When the monitoring equipment detects a vehicle, it generates vehicle information and vehicle events based on the captured vehicle images. The vehicle information and vehicle events are then sent together to the edge computing device.

[0078] Therefore, the license plate recording service module can receive vehicle information and vehicle events, and then send the vehicle information and vehicle events to the edge agent module. Optionally, the edge agent module can be used to specifically determine whether the vehicle event is a departure event of the target vehicle leaving the parking lot, and to obtain vehicle information.

[0079] Vehicle information is generated during the parking process of the target vehicle. Specifically, it refers to any vehicle-related information generated during the entire process from when the vehicle enters the parking lot, parks (or simply passes through the street) to when the vehicle leaves the parking lot.

[0080] For vehicle information generated during the parking process of a target vehicle by edge computing devices, some can be directly generated by the monitoring equipment, while others can be obtained by the edge computing device through recognition based on vehicle images acquired by the monitoring equipment. All information related to the vehicle itself can be used as vehicle information, such as vehicle characteristics, license plate number, vehicle speed, etc., not all of which are illustrated here.

[0081] Step 203: Integrate vehicle information and parking lot information into a preset instruction file to obtain the target instruction file.

[0082] The preset instruction file includes multiple preset parking analysis instructions.

[0083] In an optional embodiment of this application, parking information may be pre-stored in the edge computing device. Optionally, the parking information may be pre-stored in the RAG knowledge base module.

[0084] Therefore, edge computing devices can obtain vehicle information and parking information. The parking information includes details related to the parking lot, such as its location, billing rules, number of parking spaces, and available parking spaces; however, no security examples are provided here.

[0085] Optionally, the edge computing device can obtain all parking lot information, or it can obtain only the information related to parking billing to calculate the unpaid fees in the parking results.

[0086] The preset instruction file is pre-generated and deployed on the edge computing device. Optionally, the preset instruction file is pre-deployed in the edge agent. The preset instruction file includes multiple preset parking analysis instructions.

[0087] In optional embodiments of this application, the parking analysis instructions can be preset language instructions, such as Chinese instructions, etc., which are not fully exemplified here. Each parking analysis instruction is used to instruct the large language model to analyze vehicle information and parking lot information to obtain the parking result of the target vehicle.

[0088] Optionally, a file format that the large language model can recognize can be determined, and a preset instruction file in that format can be deployed in the edge computing device. For example, the preset instruction file is a JSON file. Of course, other format types that the large language model can recognize can also be used; these are not fully exemplified here.

[0089] Vehicle information and parking information are merged and added to a preset instruction file to form a target instruction file.

[0090] Step 204: Submit the target instruction file to the large language model, and obtain the parking result of the target vehicle by analyzing the vehicle information and parking lot information according to the parking analysis instructions in the target instruction file.

[0091] The parking results should include at least the information on the outstanding fees for the target vehicle.

[0092] For example, a large language model refers to a deep learning network based on Transformer or a hybrid architecture with ≥10B parameters; it has cross-modal understanding and generation capabilities, and in the embodiments of this application, it is used to perform unified semantic encoding and analysis on natural language toll requests, image license plates, and voice disputes to generate parking results.

[0093] The target instruction file is submitted to the large language model, which reads each parking analysis instruction sequentially and responds to each instruction, performing analysis based on vehicle and parking lot information. By responding to each parking analysis instruction, the final parking result is gradually obtained.

[0094] Parking results should at least include information on the outstanding fees for the target vehicle. Parking results may also include other information, such as parking duration and parking end time, etc., but not all examples are provided here.

[0095] Understandably, the preset instruction file can be updated based on the required parking results. That is, a new set of parking analysis instructions is determined that can be used to instruct the large language model analysis to obtain the desired parking results. These new parking analysis instructions are then added to the preset instruction file, replacing all or some of the original parking analysis instructions, resulting in an updated preset instruction file. This improves the flexibility of determining parking results.

[0096] The aforementioned parking identification method acquires vehicle events output by monitoring equipment. If the vehicle event is a departure event of a target vehicle leaving the parking lot, it acquires vehicle information generated during the target vehicle's parking process and parking lot information. The vehicle information and parking lot information are then integrated into a preset instruction file to obtain a target instruction file. This preset instruction file includes multiple preset parking analysis instructions. The target instruction file is submitted to a large language model, which analyzes the vehicle and parking information based on the parking analysis instructions in the target instruction file to obtain the parking result for the target vehicle. The parking result includes at least the outstanding payment information for the target vehicle. In this way, by automatically capturing vehicle information using monitoring equipment and then using a large language model to perform multi-dimensional data analysis on the vehicle and parking information, parking results are quickly obtained, enabling automatic license plate recording and billing. Compared to manual scanning, the efficiency of obtaining parking results is greatly improved.

[0097] In one embodiment, such as Figure 3 A flowchart illustrating the process of obtaining the target instruction file is shown. Vehicle information and parking information are merged into a preset instruction file to obtain the target instruction file, which includes:

[0098] Step 301: According to the type of each vehicle information, fill each vehicle information into the first information position in the preset instruction file.

[0099] The vehicle information includes multiple entries. Pre-defined instruction files have reserved spaces for each type of vehicle information; these entries must be filled in to ensure accurate analysis by the large language model.

[0100] For example, the default instruction file includes the following:

[0101] "You are an intelligent parking management system. Please determine whether a vehicle is truly parked and decide whether to charge based on the following information. Note: Briefly slowing down or waiting at a red light does not count as parking. If a vehicle stops near the entrance / exit without entering a parking space, it is also not considered parking. If a vehicle remains stationary in the middle area for an extended period and meets the characteristics of parking, it is considered parking."

[0102] "License plate number: XXX, entry time: XXX, departure time: XXX, stay time: XXX"

[0103] In the above text, "XXX" is the first information position that needs to be filled. It is understood that the default instruction file includes more first information positions, which are not fully exemplified here.

[0104] Step 302: According to the type of each parking lot information, fill the second information position of each parking lot information into the preset instruction file.

[0105] There are multiple parking lot information entries. Similarly, the preset instruction file has reserved spaces for filling in various types of parking lot information, which need to be filled in accordingly to ensure the accuracy of the large language model analysis.

[0106] For example, the preset instruction file also includes the following:

[0107] "XX Road, speed limit XX km / h, current parking lot has XX parking spaces, current average pedestrian traffic is XX people, current parking fee is XX, peak hours are XX."

[0108] In the above text, "XX" refers to the location of the second piece of information that needs to be filled. It is understood that the default instruction file includes more locations for this second piece of information, which are not fully illustrated here.

[0109] Step 303: Use the filled preset instruction file as the target instruction file.

[0110] In this way, by accurately filling in the information of each vehicle and parking lot, it can be ensured that the information of each vehicle and parking lot is used effectively, thereby ensuring the accuracy and efficiency of parking result analysis.

[0111] In one embodiment, the parking analysis instructions in the preset instruction file include at least one of the following:

[0112] "You are an intelligent parking management system. Please determine whether a vehicle is truly parked based on the following information and decide whether to charge. Note: If it is only a brief deceleration or waiting for a red light, it is not considered parking. If the vehicle stops near the entrance / exit but does not enter a parking space, it is also not considered parking. If the vehicle remains stationary in the middle area for a long time and meets the characteristics of parking, it is considered parking."

[0113] "Check the parking records for this vehicle."

[0114] "Check vehicles that have exceeded the time limit."

[0115] "Analyze whether the vehicle might have simply passed through quickly, or whether there was a possibility of 'lingering for an extended period without stopping.' Provide the results and reasons regarding whether or not to charge the vehicle."

[0116] “Conduct a problem diagnosis.”

[0117] Please provide a solution.

[0118] Please provide the confidence level for this parking result.

[0119] "Do you need to ask any further questions?"

[0120] Please answer in Chinese, and keep it professional and concise.

[0121] Of course, many other parking analysis commands can be included, but not all are listed here.

[0122] The process of obtaining vehicle information is explained below.

[0123] In one embodiment, such as Figure 4 A schematic diagram of a process for obtaining vehicle information is shown. Obtaining vehicle information includes:

[0124] Step 401: Obtain multiple first vehicle information and vehicle images output by the monitoring equipment.

[0125] In some embodiments, acquiring multiple first vehicle information items and vehicle images output by the monitoring device includes:

[0126] If the vehicle event is an entry event of a target vehicle entering a parking lot, the monitoring equipment outputs multiple first vehicle information and vehicle images associated with the entry event. Optionally, the multiple first vehicle information and vehicle images associated with the entry event are stored.

[0127] The first vehicle information is acquired by the monitoring equipment when it detects a target vehicle entering the parking lot. The vehicle image is a frame containing the target vehicle captured by the monitoring equipment when it detects the target vehicle entering the parking lot.

[0128] For example, the first vehicle information includes at least one of the following:

[0129] The time the vehicle entered the parking lot;

[0130] The vehicle's license plate number;

[0131] Vehicle model;

[0132] The speed at which the vehicle enters.

[0133] In some embodiments, acquiring multiple first vehicle information items and vehicle images output by the monitoring device includes:

[0134] If the vehicle event is a departure event of the target vehicle leaving the parking lot, based on the license plate number of the target vehicle associated with the departure event, determine the entry event associated with the license plate number of the target vehicle from multiple candidate entry events; obtain the first vehicle information and vehicle image associated with each entry event.

[0135] Each entry event is stored together with the associated first vehicle information and vehicle image, and is identified by the license plate number. Therefore, the edge computing device stores multiple candidate entry events.

[0136] Once a departure event is obtained, the license plate numbers of each candidate entry event are compared sequentially with the license plate numbers of the target vehicle associated with the departure event to determine the entry events related to the departure event, i.e., the departure and entry events of the target vehicle. This results in multiple pieces of first vehicle information and vehicle images of the target vehicle.

[0137] In an optional embodiment of this application, the edge computing device stores each entry event together with the associated first vehicle information and vehicle image in the RAG knowledge base.

[0138] Step 402: Determine the license plate number and vehicle characteristics of the target vehicle based on the vehicle image.

[0139] In an optional embodiment of this application, a multimodal large model is deployed in the edge computing device. Vehicle images are input into the multimodal large model to obtain the license plate number and vehicle features output by the model. The multimodal large model (MLM) can simultaneously receive modalities such as text, images, speech, and structured sensor data, and outputs a unified vector representation through a shared semantic space mapping layer, enabling one-time joint inference of license plate images.

[0140] Alternatively, the monitoring equipment can identify and obtain the vehicle's license plate number on its own.

[0141] Vehicle features can be, for example, multiple local features of the vehicle.

[0142] Step 403: Obtain the historical parking information of the target vehicle from the preset database based on the vehicle's license plate number.

[0143] Understandably, information about all vehicles associated with this parking lot can be stored in a pre-set database. That is, each time a vehicle parks in the parking lot, parking information is recorded once.

[0144] In an optional embodiment of this application, the large language model interfaces with the MCP tool module in the edge computing device. The large language model can call the MCP tool to obtain the historical parking information of the target vehicle from a preset database.

[0145] Optionally, the default database can be set in the edge computing device or in other servers.

[0146] Historical parking information includes, for example, the number of times a vehicle has parked, the duration of a vehicle's parking, the time of departure, the time of entry, etc., but no specific limits are set here.

[0147] Step 404: Use at least one of the following as vehicle information: first vehicle information, target vehicle license plate number, vehicle characteristics, and target vehicle historical parking information.

[0148] In one embodiment, such as Figure 5 A flowchart illustrating the process of obtaining parking information is shown. Obtaining parking information includes:

[0149] Step 501: Obtain basic information about the parking lot from the preset knowledge base.

[0150] The basic information includes at least one of the following: speed limit information of the road where the parking lot is located, lane information of the road where the parking lot is located, congestion time information of the road where the parking lot is located, current number of cars parked in the parking lot, number of parking spaces in the parking lot, and parking lot billing rules.

[0151] The congestion information for the road where the parking lot is located can refer to peak hours.

[0152] Step 502: Recognize and analyze the vehicle images to determine the pedestrian and vehicle traffic information of the road where the parking lot is located.

[0153] In an optional embodiment of this application, vehicle images are input into a multimodal large model for analysis to obtain pedestrian and vehicle traffic information output by the multimodal large model.

[0154] Step 503: Use the basic information of the parking lot, pedestrian flow information, and vehicle flow information as parking lot information.

[0155] In one embodiment, the output of the large language model is a file in a preset format, such as a JSON file. The JSON file is parsed using a post-processing service in an edge computing device to obtain parking results, and both vehicle information and parking results are written to a preset database as historical data for the target vehicle.

[0156] In one embodiment, the parking result also includes the probability that the target vehicle is parked in the parking lot.

[0157] Among them, the parking analysis instruction includes a parking probability analysis instruction. Under the guidance of the parking probability analysis instruction, the large language model determines the probability that the target vehicle will park in the parking lot based on the vehicle information.

[0158] Optionally, the large language model can be instructed to first analyze the probability of a vehicle currently parking based on vehicle information (current vehicle information and historical parking information). If the parking probability is low (i.e., the vehicle only passes through this road), the parking result analysis is not continued. Conversely, if the parking probability is high, the large language model is instructed to continue analyzing the vehicle's parking result based on the vehicle information to obtain information including parking fees to be paid, etc.

[0159] This improves the efficiency of the model's analysis results when only vehicles pass by, avoiding unnecessary cost analysis. It also enhances the flexibility of large language models in performing analyses.

[0160] In one embodiment, the parking result also includes a confidence level of the parking result.

[0161] For ease of understanding, the following describes the application of the above parking recognition method to an edge computing device using a complete embodiment.

[0162] This method is used in edge computing devices. License plate recognition cameras are deployed at street entrances and exits to acquire vehicle information such as license plate number, vehicle model, and speed, and send this information to the edge computing device. The edge computing device has a built-in intelligent agent with a large model that analyzes the current vehicle parking behavior based on the license plate recognition information, vehicle information, vehicle speed, and external environment to achieve vehicle registration and billing. The structure of the edge computing device is as follows: Figure 6 As shown.

[0163] License plate recording service: The system uses a radar-guided camera to capture events of vehicles entering and leaving the street, as well as the vehicle model, speed of entry and exit, and analyzes the current traffic flow information based on images of entering vehicles, then sends the information to the intelligent agent.

[0164] Edge Agent: A data analysis agent based on a large language model and a multimodal large model. After receiving vehicle information, the agent reads parking information about the street from the RAG knowledge base (e.g., street vehicle speed limits, street lane information, street pedestrian flow information, peak hours, street parking space information, current street charging rules, etc.). It uses the MCP tool to read the current vehicle's parking history and the total number of vehicles that have been on the street for more than 30 minutes. The agent integrates the data read by the MCP tool, the RAG knowledge base information, and the context of the latest information into a prompt (preset instruction file) and sends it to the large model. The large model then determines the probability of the current vehicle parking and collects parking fees.

[0165] Post-processing service: The output of the intelligent agent is a JSON file. The JSON file is parsed to obtain vehicle parking behavior data and billing information, which is then written to an SQL database.

[0166] MCP tool: Provides real-time information on vehicles currently entering the parking lot and historical parking records.

[0167] RAG Knowledge Base: Stores street parking information (street vehicle speed limits, street lane information, street pedestrian flow information, peak hours, street parking space information, current street charging rules, etc.).

[0168] The parking recognition process of edge computing devices is as follows Figure 7 As shown in the complete diagram. The entry event is a driving-in event, and relevant information is sent to the agent. When the agent receives the exit event (i.e., the driving-out event), it uses a large language model to analyze the parking result and obtain billing information as to whether or not payment is required.

[0169] This method replaces the original three-tiered architecture of "license plate recognition + rule-based fee rates + manual auditing" with a single large model that operates end-to-end. It inputs multimodal data, including entry / exit events, parking duration, vehicle type, speed, historical parking records, real-time street speed limits, and peak traffic flow, into the large model at once. This directly regresses to generate a dynamic fee rate tensor and completes millisecond-level deductions. Simultaneously, incremental enhancement and fine-tuning shorten the deployment cycle for new scenarios, completely eliminating rule table maintenance and manual auditing, ensuring the efficiency and reliability of parking result acquisition.

[0170] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0171] Based on the same inventive concept, this application also provides a parking recognition device for implementing the parking recognition method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more parking recognition device embodiments provided below can be found in the limitations of the parking recognition method described above, and will not be repeated here.

[0172] In one embodiment, such as Figure 8 As shown, a parking recognition device is provided, including: an event acquisition module, an information acquisition module, an information fusion module, and a result analysis module, wherein:

[0173] The event acquisition module is used to acquire vehicle events output by the monitoring equipment;

[0174] The information acquisition module is used to acquire vehicle information generated during the parking process of the target vehicle and parking information of the parking lot if the vehicle event is a departure event of the target vehicle leaving the parking lot.

[0175] The information fusion module is used to merge vehicle information and parking lot information into a preset instruction file to obtain a target instruction file; the preset instruction file includes multiple preset parking analysis instructions.

[0176] The results analysis module is used to submit the target instruction file to the large language model, and obtain the parking results of the target vehicle by analyzing the vehicle information and parking lot information based on the parking analysis instructions in the target instruction file. The parking results include at least the information of the outstanding fees corresponding to the target vehicle.

[0177] In one embodiment, there are multiple vehicle information and parking lot information; the information fusion module is specifically used for:

[0178] Based on the type of each vehicle information, fill each vehicle information into the first information position in the preset instruction file;

[0179] Based on the type of information for each parking lot, fill the second information position in the preset instruction file with the information for each parking lot;

[0180] Use the filled-in preset instruction file as the target instruction file.

[0181] In one embodiment, the information acquisition module is specifically used for:

[0182] The system acquires multiple first vehicle information and vehicle images output by the monitoring equipment; wherein each first vehicle information is acquired by the monitoring equipment when it detects a target vehicle entering the parking lot.

[0183] Based on the vehicle image, determine the license plate number and vehicle characteristics of the target vehicle;

[0184] Based on the target vehicle's license plate number, retrieve the target vehicle's historical parking information from a pre-set database;

[0185] Use at least one of the following as vehicle information: first vehicle information, target vehicle license plate number, vehicle characteristics, and target vehicle historical parking information.

[0186] In one embodiment, the information acquisition module is specifically used for:

[0187] If the vehicle event is a departure event where the target vehicle leaves the parking lot, the entry event associated with the target vehicle's license plate number is determined from multiple candidate entry events based on the license plate number of the target vehicle associated with the departure event.

[0188] Obtain information and images of each first vehicle associated with the entry event.

[0189] In one embodiment, the information acquisition module is specifically used for:

[0190] Obtain basic parking information from a pre-set knowledge base; parking information includes at least one of the following: speed limit information of the road where the parking lot is located, lane information of the road where the parking lot is located, congestion time information of the road where the parking lot is located, current number of cars parked in the parking lot, number of parking spaces in the parking lot, and parking fee rules;

[0191] Vehicle images are identified and analyzed to determine pedestrian and vehicle traffic information at the location of the parking lot on the road.

[0192] The basic information of the parking lot, pedestrian flow information, and vehicle flow information are used as parking lot information.

[0193] In one embodiment, the parking result also includes the probability that the target vehicle is parked in the parking lot; wherein, the parking analysis instruction includes a parking probability analysis instruction, and the large language model determines the probability that the target vehicle is parked in the parking lot based on the vehicle information under the instruction of the parking probability analysis instruction.

[0194] Each module in the aforementioned parking recognition device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0195] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores parking recognition data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a parking recognition method.

[0196] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0197] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0198] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0199] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0200] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0201] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0202] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0203] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A parking recognition method, characterized in that, The method includes: Acquire vehicle events output by the monitoring equipment; If the vehicle event is a departure event of the target vehicle leaving the parking lot, then obtain the vehicle information generated during the parking process of the target vehicle and the parking lot information. The vehicle information and the parking lot information are merged into a preset instruction file to obtain a target instruction file; wherein, the preset instruction file includes multiple preset parking analysis instructions; The target instruction file is submitted to a large language model, and the large language model analyzes the vehicle information and the parking lot information according to the parking analysis instructions in the target instruction file to obtain the parking result of the target vehicle. The parking result includes at least the payment information of the target vehicle.

2. The method according to claim 1, characterized in that, The vehicle information and the parking lot information are multiple; the step of fusing the vehicle information and the parking lot information into a preset instruction file to obtain a target instruction file includes: According to the type of each vehicle information, each vehicle information is filled into the first information position in the preset instruction file; According to the type of each parking lot information, fill each parking lot information into the second information position of the preset instruction file; The pre-defined instruction file after filling in the information is used as the target instruction file.

3. The method according to claim 1, characterized in that, Obtaining the vehicle information includes: The monitoring device outputs multiple first vehicle information and vehicle images; wherein each first vehicle information is obtained by the monitoring device when it detects the target vehicle entering the parking lot. Based on the vehicle image, determine the license plate number and vehicle characteristics of the target vehicle; Based on the license plate number of the target vehicle, retrieve the historical parking information of the target vehicle from a preset database; The vehicle information is defined as at least one of the following: the first vehicle information, the license plate number of the target vehicle, the vehicle characteristics, and the historical parking information of the target vehicle.

4. The method according to claim 3, characterized in that, The acquisition of multiple first vehicle information and vehicle images output by the monitoring device includes: If the vehicle event is a departure event of the target vehicle leaving the parking lot, the entry event associated with the license plate number of the target vehicle is determined from multiple candidate entry events based on the license plate number of the target vehicle associated with the departure event. Obtain information and images of each first vehicle associated with the entry event.

5. The method according to claim 3, characterized in that, Obtaining the parking lot information includes: Obtain basic information about the parking lot from a preset knowledge base; the basic information includes at least one of the following: speed limit information of the road where the parking lot is located, lane information of the road where the parking lot is located, congestion time information of the road where the parking lot is located, current number of cars parked in the parking lot, number of parking spaces in the parking lot, and parking lot billing rules. The vehicle images are identified and analyzed to determine the pedestrian and vehicle traffic information at the location of the parking lot on the road. The basic information of the parking lot, the pedestrian flow information, and the vehicle flow information are used as the parking lot information.

6. The method according to any one of claims 1 to 5, characterized in that, The parking result also includes the probability that the target vehicle is parked in the parking lot; The parking analysis instruction includes a parking probability analysis instruction, and the large language model, under the instruction of the parking probability analysis instruction, determines the probability that the target vehicle will park in the parking lot based on the vehicle information.

7. A parking recognition device, characterized in that, The device includes: The event acquisition module is used to acquire vehicle events output by the monitoring equipment; The information acquisition module is used to acquire vehicle information generated during the parking process of the target vehicle and parking information of the parking lot if the vehicle event is a departure event of the target vehicle leaving the parking lot. The information fusion module is used to fuse the vehicle information and the parking lot information into a preset instruction file to obtain a target instruction file; wherein, the preset instruction file includes multiple preset parking analysis instructions; The result analysis module is used to submit the target instruction file to the large language model, obtain the parking result of the target vehicle by the large language model analyzing the vehicle information and the parking lot information according to the parking analysis instructions in the target instruction file, and the parking result includes at least the payment information of the target vehicle.

8. A parking recognition system, characterized in that, The parking recognition system includes an edge computing device and a monitoring device. The edge computing device is used to execute the parking recognition method as described in any one of claims 1 to 6; the monitoring device is used to monitor vehicle events.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images