On-board information processing method, device, medium and equipment for environmentally friendly vehicles

Abstract

The present application provides a method, device, medium, and equipment for processing onboard information of an environmentally friendly vehicle, belonging to the field of data processing technology. The method includes: obtaining an onboard list image to be identified; identifying a corresponding first vehicle identifier from an identification code in the onboard list image; performing multimodal resolution enhancement processing on the onboard list image to generate at least two enhanced images, calling at least one OCR model to extract text information in the onboard list image from the at least two enhanced images; identifying corresponding first environmental protection key information based on the extracted text information; and determining whether the corresponding environmentally friendly vehicle has passed verification based on a comparison of the first vehicle identifier, the first environmental protection key information, and known vehicle information. The present application can improve the accuracy of environmental protection verification.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device, medium and equipment for processing on-board information of an environmentally friendly vehicle. Background Art

[0002] With the continuous improvement of the national environmental protection policy system and the continuous strengthening of supervision, the control of environmental protection information of vehicles entering and leaving factories and parks has become an important part of environmental governance. According to current regulations, vehicles that do not meet the environmental protection information standards will be prohibited from entering the factory to ensure the green compliance of production activities. In the current commonly used control process, applicants need to fill in environmental protection information through a dedicated information system and upload pictures of the vehicle list (including photos of paper lists or electronic documents). Only after manual review can they obtain authorization to enter the factory. However, this model has significant flaws: because the style of the vehicle list is affected by changes in years and regional policy differences, it presents diverse characteristics. Manual review is prone to problems such as information omissions and misjudgments, making it difficult to guarantee the efficiency and accuracy of the review.

[0003] To address these issues, the industry has introduced optical character recognition (OCR) technology. By automatically identifying the content of vehicle inventory images uploaded via mobile devices, it assists manual review and improves efficiency. However, this technology still faces numerous challenges in practical application. Firstly, paper vehicle inventory documents are prone to physical damage such as wrinkles and tears during circulation. Secondly, user photography often results in insufficient image resolution and blurred content, often due to limitations in equipment performance, operating procedures, and environmental conditions. These factors significantly reduce the accuracy of OCR technology in recognizing environmental information, leading to erroneous review results and compromising the reliability and effectiveness of the management and control process. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, medium and equipment for processing on-board information of an environmentally friendly vehicle to solve at least one of the above-mentioned technical problems.

[0005] In a first aspect, the present application provides a method for processing onboard information of an environmentally friendly vehicle, the method comprising:

[0006] Get the vehicle list image to be identified;

[0007] Identify the corresponding first vehicle identifier from the identification code in the vehicle list image;

[0008] Performing multimodal resolution enhancement processing on the vehicle list image to generate at least two enhanced images, and extracting text information from the vehicle list image from the at least two enhanced images by calling at least one OCR model;

[0009] Identifying the corresponding first environmental protection key information according to the extracted text information;

[0010] Based on the comparison between the first vehicle identification, the first environmental protection key information and known vehicle information, it is determined whether the corresponding environmental protection vehicle has passed the verification.

[0011] Optionally, comparing the first vehicle identifier, the first environmental protection key information, and known vehicle information to determine whether the corresponding environmentally friendly vehicle has passed verification includes:

[0012] comparing the first vehicle identifier and the second vehicle identifier in the known vehicle information to determine whether they are completely consistent;

[0013] Verifying the authenticity of the corresponding on-board manifest based on the first environmental protection key information;

[0014] When the vehicle identification is completely consistent and the accompanying vehicle list is a genuine list, the first environmental protection key information is compared with the second environmental protection key information in the known vehicle information to see if they match. If so, the verification is determined to be successful.

[0015] If the vehicle identification is not completely consistent, or the accompanying list is not a real list, or the key environmental protection information does not match, the verification will be deemed to have failed.

[0016] Optionally, verifying the authenticity of the corresponding on-board manifest based on the first environmental protection key information includes:

[0017] Extracting a corresponding first key field and first field content corresponding to the first key field from the first environmental protection key information;

[0018] Identify the overlap rate between the extracted key fields and the standard key fields in the standard key field set required by the vehicle manifest;

[0019] Identify whether the content of each first field complies with the field content specification of the corresponding overlapping standard key field, and calculate the corresponding content standardization degree based on the standardization result;

[0020] When the coincidence rate exceeds the corresponding coincidence rate threshold, and the content standardization exceeds the corresponding standardization threshold, the on-board manifest is determined to be a true manifest;

[0021] When the overlap rate does not exceed the corresponding overlap rate threshold, or the content standardization does not exceed the corresponding standardization threshold, it is determined that the on-board manifest is not a true manifest.

[0022] Optionally, the first enhanced image uses nearest neighbor interpolation to upsample the vehicle list image to the target resolution and then performs Gaussian filtering, the second enhanced image uses bicubic interpolation to upsample the image to the target resolution; the third enhanced image uses bilinear interpolation to upsample the image to the target resolution;

[0023] The calling of at least one OCR model to extract text information in the vehicle list image from at least two enhanced images includes: calling at least one preset OCR model to perform text recognition on each enhanced image, each OCR model outputting a character recognition result for each enhanced image, the OCR model including a Tesseract-OCR model; and determining the text information in the vehicle list image based on each character recognition result.

[0024] Optionally, each character recognition result includes the recognized character and the original confidence of the corresponding character; and determining the text information in the vehicle list image based on each character recognition result includes:

[0025] The original confidence output by each OCR model is calibrated across models to obtain the standard confidence;

[0026] Character selection is performed based on the recognized characters in each character recognition result and the corresponding standard confidence level;

[0027] The text information in the vehicle list picture is formed based on the selected characters.

[0028] Optionally, the character selection based on the recognized characters in each character recognition result and the corresponding standard confidence level includes: for each character recognition result, when the recognized characters belonging to the same position are inconsistent, performing a weighted voting calculation on each character in the same position based on the standard confidence level, and selecting a candidate character from a plurality of characters based on the weighted voting calculation result;

[0029] The forming of text information in the vehicle list picture based on the selected characters includes:

[0030] forming initial text information based on the selected candidate characters;

[0031] Performing a vehicle information standardization check on the initial text information, and identifying candidate characters in the initial text information that do not conform to the vehicle information standardization as characters to be corrected;

[0032] Correcting the character to be corrected based on each character recognition result to obtain a corrected character corresponding to the character to be corrected;

[0033] The text information in the vehicle list picture is formed according to the initial text information and the corrected characters.

[0034] Optionally, performing cross-model calibration on the original confidence output by each OCR model to obtain a standard confidence includes: mapping the original confidence c0 to the corresponding standard confidence c1 according to the following piecewise linear function:

[0035] ,

[0036] Among them, k1, k2, b1, b2, and c2 are preset confidence calibration parameters.

[0037] In a second aspect of the present application, a vehicle-mounted information processing device for an environmentally friendly vehicle is provided, the device comprising:

[0038] Image acquisition module, used to obtain the vehicle list image to be identified;

[0039] an identification extraction module, configured to identify a corresponding first vehicle identification from the identification code in the vehicle list image;

[0040] an OCR recognition module configured to perform multimodal resolution enhancement processing on the vehicle manifest image to generate at least two enhanced images, invoke at least one OCR model to extract text information from the vehicle manifest image from the at least two enhanced images, and identify corresponding first environmental protection key information based on the extracted text information;

[0041] The verification module is used to compare the first vehicle identifier, the first environmental protection key information and known vehicle information to determine whether the corresponding environmental protection vehicle has passed the verification.

[0042] In a third aspect of the present application, a computer-readable storage medium is provided, on which executable instructions are stored. When the executable instructions are executed by a processor, the processor executes the method described in any embodiment of the present application.

[0043] In a fourth aspect of the present application, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the method as described in any one of the embodiments of the present application.

[0044] The on-board information processing method, device, medium and equipment of an environmentally friendly vehicle in the present application generates at least two enhanced images by identifying a first vehicle identification from an on-board list image and performing multimodal resolution enhancement processing on the on-board list image, calling at least one OCR model to extract text information in the on-board list image from the at least two enhanced images, extracting first environmental key information from the text information, and performing environmental verification based on the first vehicle identification and the first environmental key information by comparing them with known vehicle information, which can improve the efficiency of environmental verification. In addition, by extracting the text information through at least one OCR model, the accuracy of the extracted text information can be improved in low-resolution on-board list images, thereby improving the accuracy of the subsequent extraction of the first environmental key information, and ultimately improving the accuracy of environmental verification. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope of the present application.

[0046] Figure 1 1 is a flow chart of a method for processing on-board information of an environmentally friendly vehicle in one embodiment;

[0047] Figure 2 1. A flowchart of an embodiment of comparing a first vehicle identifier, first environmental protection key information, and known vehicle information to determine whether a corresponding environmentally friendly vehicle has passed verification.

[0048] Figure 3 1. A schematic diagram of a process for verifying the authenticity of a corresponding vehicle manifest based on first environmental protection key information in one embodiment;

[0049] Figure 4 A schematic diagram of a process for extracting text information from the vehicle list image from at least two enhanced images by calling at least one OCR model in one embodiment;

[0050] Figure 5 A schematic diagram of a process for determining text information in a vehicle list image based on each character recognition result in one embodiment;

[0051] Figure 6 A schematic diagram of a process for forming text information in a vehicle list image based on selected characters in one embodiment;

[0052] Figure 7 Schematic diagram of the structure of an on-board information processing device for an environmentally friendly vehicle in one embodiment;

[0053] Figure 8FIG. 1 is a schematic structural diagram of an electronic device in an embodiment. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0055] All terms (including technical and scientific terms) used in this application have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0056] For example, the terms "first," "second," etc. used in this application may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish a first element from another element.

[0057] For example, the terms "include", "comprising", etc. used in this application indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0058] This application provides a method for processing vehicle information of an environmentally friendly vehicle. Figure 1 As shown, the method includes:

[0059] Step 110: Obtain the vehicle list image to be identified.

[0060] In this embodiment, the vehicle accompanying list image is generated by photographing or scanning the vehicle's environmental information accompanying list. The environmental information accompanying list is a document that records detailed environmental information about the vehicle. It covers environmental data for various types of vehicles, including gasoline, diesel, and electric vehicles. It is a key procedure in the vehicle registration process and includes, but is not limited to, fields related to basic vehicle information, emission standards, fuel efficiency, key environmental configurations, and the content within the corresponding fields.

[0061] The on-board list image can be an image obtained by photographing a paper environmental protection information on-board list using the client, or an image converted from a scanned copy of the paper environmental protection information on-board list or an electronic version of the list in a PDF format. For example, a user can photograph a paper environmental protection information on-board list using the client to create an image, upload the image to an electronic device, and the electronic device will receive the image and use it as the on-board list image.

[0062] Step 120 : Identify the corresponding first vehicle identification from the identification code in the vehicle list image.

[0063] In this embodiment, the vehicle list image includes the above-mentioned fields and the environmental protection information of the field contents, and also includes the identification code of the corresponding environmentally friendly vehicle, which is loaded with the vehicle identification of the corresponding environmentally friendly vehicle. The identification code can be a barcode or a QR code.

[0064] Upon receiving the image of the accompanying vehicle list, the electronic device invokes the corresponding identification code recognition module to scan and identify the identification code, reading the content contained therein as the first vehicle identification. The vehicle identification is used to uniquely identify the vehicle and can be composed of a preset number of digits, letters, or other special characters. Specifically, the vehicle identification can be the VIN code of an environmentally friendly vehicle.

[0065] Step 130 : Perform multimodal resolution enhancement processing on the vehicle list image to generate at least two enhanced images, and call at least one OCR model to extract text information in the vehicle list image from the at least two enhanced images.

[0066] In this embodiment, the electronic device is preset with one or more OCR models, which may include but are not limited to: one or more of the Tesseract-OCR model, the Transformer-based OCR model, and the end-to-end OCR model. Each OCR model includes an image preprocessing module, an image enhancement module, and an OCR recognition framework. The image preprocessing module performs grayscale, binarization, denoising, tilt correction, and other operations on the input image; the image enhancement module is used to perform multimodal resolution enhancement processing on the preprocessed image, thereby obtaining at least one resolution-enhanced image to improve the recognition accuracy of the text information output by the subsequent OCR recognition framework. For example, a variety of image enhancement processing methods can be used to obtain corresponding multiple enhanced images (for example, a variety of image enhancement processing methods such as nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation can be used), and text information recognition is performed independently for each enhanced image to obtain corresponding multiple text information recognition results.

[0067] Specifically, for each OCR model, the electronic device pre-prepares a corresponding training dataset and test dataset. Each OCR model is iteratively trained based on the training dataset. After the iterative training is completed, the test dataset is used to test and verify each iteratively trained OCR model. If the verification passes, the trained OCR model is obtained. If the verification fails, the training and verification of the failed OCR model continues until the verification passes. Based on each trained OCR model, the vehicle list image is recognized to identify the corresponding text information from the vehicle list image.

[0068] The training and test datasets may include a large number of historical vehicle inventory images and the corresponding text recognition results. The images in the training and test datasets have different resolutions, and some images are randomly added with a small amount of Gaussian noise. Other images are randomly rotated (±3°), tilted (±2°), and subjected to perspective distortion (0.05-0.1) to simulate the slight tilt effect caused by scanning or photographing images.

[0069] Taking the Tesseract-OCR model as an example, Tesseract-OCR uses the long short-term memory (LSTM) network to train the tessdata_fast network in the Tesseract-OCR model. This model can automatically recognize and extract key information from vehicle manifests. If the recognition effect is not satisfactory during use, you can further optimize the model using the tessdata_best network (available in the Tesseract repository). This is a model that continues to be trained and optimized to improve recognition accuracy and reliability.

[0070] Specifically, the Tesseract-OCR model can be based on the Tesseract 5.3.3 open source engine, replacing the original LSTM network with a bidirectional LSTM network to add an attention mechanism layer. The Transformer-based OCR model can be a CNN + Transformer attention integrated model. The end-to-end OCR model can be an LSTM-CTC end-to-end model. Its feature extraction layer can use a convolutional layer with 7 convolutional blocks + 3 2×2 max pooling layers. The sequence processing layer includes a bidirectional LSTM network and a Bahdanau-based attention mechanism.

[0071] Each OCR model's training mode can include fine-tuning and training from scratch. For fine-tuning, an existing trained OCR model can be started with the aforementioned training dataset as additional data for iterative training to fine-tune the model, resulting in a fully trained and verified OCR model. By incorporating the environmental knowledge graph, during the fine-tuning process, samples containing environmental-related terms in the recognized images are given greater weight (e.g., by applying an attention weight shift) and a reward for matching environmental terms is added to the loss function, thereby improving the recognition accuracy of vehicle inventory images.

[0072] If the OCR model obtained by fine-tuning training fails verification, you can cut off the top layer (or any number of layers) from the network of the corresponding OCR model and retrain a new top layer using the new dataset until a trained OCR model is obtained.

[0073] If it still does not work, you can start training from scratch by increasing the number of iterations (for example, using 800 epochs of training iterations) and the number of training samples, so that you can eventually obtain trained OCR models.

[0074] By integrating a trained OCR model into the information system, automatic recognition of vehicle manifests is achieved. Users simply upload an image of the manifest, and the corresponding OCR model outputs the text information contained in the image. Each OCR model can output text recognition results for one or more corresponding vehicle manifest images. The electronic device then combines these multiple text recognition results to ultimately generate the corresponding text information. For example, the characters contained in each text recognition result can be referenced, and a voting mechanism can be used to select the most accurately recognized characters. These selected characters are then combined to form the corresponding text information.

[0075] In this embodiment, resolution recognition is performed on a captured image of the vehicle manifest. If the recognized resolution falls below a preset resolution threshold, the aforementioned multiple resolution enhancement processes are applied. Each resolution enhancement process outputs an enhanced image, thereby converting the vehicle manifest image into multiple enhanced images with the target resolution. For each enhanced image, one or more of the aforementioned OCR models are invoked for OCR recognition. Each OCR model outputs a corresponding character recognition result for each enhanced image. Text information is then determined from these multiple character recognition results, thereby improving the accuracy of OCR recognition of low-resolution vehicle manifest images.

[0076] Among the multiple OCR models, one of the OCR models can be used as a default OCR model, for example, the Tesseract-OCR model can be used as the default OCR model; among the multiple resolution enhancement processing modes, one of the resolution enhancement processing modes can be used as the default processing mode, for example, the bicubic interpolation image enhancement processing can be used as the default processing mode.

[0077] When the recognized resolution exceeds a preset resolution threshold, only one of the resolution enhancement processes (e.g., the default processing mode) may be used to output an enhanced image. One of the OCR models (e.g., the default OCR model) is then invoked to perform OCR on the enhanced image. The text information is then determined based on the recognized character recognition results, thereby improving OCR recognition efficiency. The resolution threshold may be a preset threshold value based on actual conditions, such as 75dpi, 100dpi, 120dpi, 150dpi, 200dpi, or any other suitable value.

[0078] Step 140: Identify the corresponding first environmental protection key information based on the extracted text information.

[0079] In this embodiment, the first environmental protection key information refers to information that requires environmental protection verification, which may include the field content in each environmental protection field that requires review and verification, such as the specific content corresponding to the vehicle basic information, emission standards, environmental protection key configuration, validity period, fuel efficiency and other fields (i.e., the first field content below). These contents constitute the corresponding first environmental protection key information.

[0080] Specifically, semantic alignment and data verification can be performed on the extracted text information. During the semantic alignment process, based on the keyword library in the field of motor vehicle environmental protection information, the fields identified by OCR (such as "emission standards") are mapped to standardized fields, so that the fields to which each text content belongs and the contents under the specific fields can be known. During the data verification process, the compliance of the specific content of the field can be verified in real time to intercept invalid data with format errors or missing fields, thereby improving the accuracy of the extracted first environmental protection key information. It can be understood that the field content (specific content) corresponding to different fields has certain data specification requirements. Based on this data specification requirement, the specific content can be verified.

[0081] Step 150 : Based on the comparison of the first vehicle identification, the first environmental protection key information and the known vehicle information, it is determined whether the corresponding environmental protection vehicle has passed the verification.

[0082] In this embodiment, the identified first vehicle identifier and first environmental protection key information can be compared with known vehicle information. Specifically, based on the first vehicle identifier, a vehicle search can be performed in a vehicle information database to identify a target vehicle requiring environmental protection verification. The target vehicle's information stored in the vehicle information database is then compared with the first environmental protection key information to determine whether the corresponding environmental protection data is consistent. Based on the query result, a determination is made as to whether the corresponding environmentally friendly vehicle can pass environmental protection verification.

[0083] For example, if the identified first vehicle identification (VIN code) is "LSVNV133X12345678", the vehicle with the VIN code "LSVNV133X12345678" is queried from the vehicle information database as the target vehicle. For the target vehicle, the collected environmental protection information or environmental protection information requirements (such as the index requirements of some environmental protection parameters) of the target vehicle are compared with the first environmental protection key information uploaded by the user to check whether the two are consistent, or to check whether the first environmental protection key information meets the environmental protection information requirements of the corresponding vehicle. If they meet the consistency or the environmental protection information requirements, the environmentally friendly vehicle is determined to have passed the verification.

[0084] The on-board information processing method of an environmentally friendly vehicle in the present application identifies the first vehicle identification from the on-board list picture, calls at least one OCR model to identify the text information in the picture, extracts the first environmental key information from the text information, and performs environmental verification based on the first vehicle identification and the first environmental key information by comparing them with the known vehicle information, thereby improving the efficiency of environmental verification. In addition, by extracting the text information through at least one OCR model, the accuracy of the text information extracted from the low-resolution on-board list picture can be improved, thereby improving the subsequent extraction accuracy of the first environmental key information, and ultimately improving the accuracy of environmental verification.

[0085] In one embodiment, Figure 2 As shown, step 150 includes:

[0086] Step 210: Compare the first vehicle identifier with the second vehicle identifier in the known vehicle information to see if they are completely consistent.

[0087] In this embodiment, the first vehicle identification has characteristic format requirements, such as a specific number of digits, with the characters in some digits requiring specific numbers or letters. The recognized first vehicle identification can be checked to see if it meets the preset format requirements. If it does not, the vehicle identifications can be directly determined to be inconsistent. If the format requirements are met, the vehicle identification is searched in the vehicle information database. If a completely identical vehicle identification (i.e., the second vehicle identification) is found, the vehicle identifications are determined to be consistent. If no corresponding vehicle identification is found, the vehicle identifications are determined to be inconsistent.

[0088] Step 220: Verify the authenticity of the corresponding on-vehicle manifest based on the first environmental protection key information.

[0089] Environmental protection key information refers to core data directly related to the environmental performance of the vehicle. It is the basis for determining whether the vehicle meets environmental protection standards and whether it can legally be on the road or enter a specific area. For the first environmental protection key information identified, the vehicle identification contained in the first environmental protection key information can be verified with the first vehicle identification identified in the above-mentioned identification code to check whether the two are completely consistent. When the two are completely consistent, the first key field and the corresponding field content in the environmental protection key information are further identified to check whether the first key field is a field related to environmental protection verification and whether the corresponding field content meets the verification requirements, thereby determining whether the on-board manifest is a genuine manifest. For example, for the validity period field, it can be determined whether the validity period content is within the validity period required by the environmental protection requirements. If it is not within the validity period, it means that the on-board manifest is an expired on-board manifest and does not meet the authenticity requirements.

[0090] In one embodiment, Figure 3 As shown, step 220 includes:

[0091] Step 310: extract the corresponding first key field and the first field content corresponding to the first key field from the first environmental protection key information.

[0092] The first key field represents the name of the data field related to the vehicle's environmental protection verification, such as the emission standard and environmental protection number, identified from the vehicle inventory image. The first field content represents the specific data value corresponding to the first key field. For example, some of the first key fields and their first field contents extracted from the first environmental protection key information are as follows: Emission Standard: National VI; Environmental Protection Number: Beijing Environmental Standard 20230616; Effective Date: 2025-12-31; Emission Limit: 0.5mg / m³; Testing Agency: XX City Environmental Monitoring Center.

[0093] Step 320 , identifying the overlap rate between the extracted key fields and the standard key fields in the standard key field set required by the vehicle manifest.

[0094] In this embodiment, the electronic device predefines the standard key fields required to be included in the vehicle manifest. The electronic device then checks the first key field identified from the image to see if it complies with the standard key fields. Understandably, since the user's vehicle manifest may be wrinkled or worn, the first key field may be incomplete or some key fields may be incorrectly identified, such as misidentifying "emission standards" as "emission direction standards." Based on this, the electronic device can perform semantic recognition on each identified key field and correct any misidentified fields, thereby improving the accuracy of key field recognition.

[0095] Standard key fields can be extracted from a pre-set standard template. The first key field is compared with the standard key field to determine the overlap ratio between the two. Because some fields in the vehicle manifest may be incomplete or missing, an appropriate overlap ratio threshold can be set. When the calculated overlap ratio exceeds the corresponding overlap ratio threshold, step 330 can be executed. If the overlap ratio is less than the corresponding overlap ratio threshold, the vehicle manifest is directly judged to be inauthentic, and the user is required to provide a new vehicle manifest. For example, the user can retake a higher-quality image of the vehicle manifest to re-perform environmental verification.

[0096] It is understood that the overlap rate threshold can be adaptively adjusted based on the quality of the uploaded vehicle inventory image, or based on business needs and actual recognition accuracy, thereby improving the applicability of vehicle environmental verification. For example, the overlap rate threshold can be set to 80%.

[0097] For example, the identified key fields include: {"emission standard", "environmental protection number", "effective date", "emission limit", "testing agency"}, a total of five. In addition to these five required standard key fields, the engine model is also included. Based on this, the corresponding overlap rate can be calculated as 5 / 6 ≈ 83.3%.

[0098] Step 330 , identifying whether the content of each first field complies with the field content specification of the corresponding overlapping standard key field, and calculating the corresponding content standardization degree based on the standardization result.

[0099] In this embodiment, different fields have corresponding content specification requirements. For example, for the environmental protection number, the content specification requirements are: provincial abbreviation + environmental label + 8 digits; the content specification requirements corresponding to the effective date field are: YYYY-MM-DD, and not expired; the content specification requirements corresponding to the emission limit field are 0.1-1.0mg / m 3 For each first field content, identify whether it complies with the corresponding field content specification, and calculate the corresponding content standardization degree based on the identification result.

[0100] Step 340 : When the overlap rate exceeds the corresponding overlap rate threshold and the content standardization exceeds the corresponding standardization threshold, the on-board manifest is determined to be a true manifest.

[0101] Step 350 : When the overlap rate does not exceed the corresponding overlap rate threshold, or the content standardization does not exceed the corresponding standardization threshold, it is determined that the on-board manifest is not a true manifest.

[0102] In this embodiment, by performing overlap rate and standardization verification, the accuracy of the authenticity determination of the list can be improved.

[0103] Step 230 : When the vehicle identifications are completely consistent and the accompanying vehicle list is a genuine list, the first environmental protection key information is compared with the second environmental protection key information in the known vehicle information to see if they match. If they do match, the verification is determined to be successful.

[0104] If the vehicle identification is completely consistent and the accompanying list is preliminarily identified as a real list, the first environmental protection key information can be compared with the second environmental protection key information corresponding to the target vehicle to see if they match. If they match, the verification is determined to be successful.

[0105] Specifically, the matching requirements for different key environmental information may not be the same. For example, some key environmental information requires a complete match, while others may only require a fuzzy match. Different key environmental information is judged based on the corresponding matching requirements. Verification is considered successful when all key environmental information meets the corresponding matching requirements. Alternatively, verification is considered successful when more than a preset number of first key environmental information meet the matching requirements.

[0106] By verifying key environmental information and vehicle identification, the accuracy of environmental vehicle verification can be improved.

[0107] In step 240 , if the vehicle identification is not completely consistent, or the accompanying vehicle list is not a genuine list, or the environmental protection key information does not match, the verification is determined to have failed.

[0108] In this embodiment, when any of the following situations occurs, such as the vehicle identification is not completely consistent, the accompanying list is not the real list, or the key environmental protection information does not match, the verification is determined to have failed, and the reason for the verification failure is prompted, so that the user can submit the real environmental protection information or re-perform the vehicle environmental protection verification to ensure the safety of the vehicle.

[0109] In one embodiment, the first enhanced image uses nearest neighbor interpolation to upsample the vehicle list image to the target resolution and then performs Gaussian filtering. The second enhanced image uses bicubic interpolation to upsample the image to the target resolution. The third enhanced image uses bilinear interpolation to upsample the image to the target resolution. Figure 4 As shown, calling at least one OCR model to extract text information from the vehicle list image from at least two enhanced images includes:

[0110] Step 410: Call at least one preset OCR model to perform text recognition on each enhanced image, and each OCR model outputs a character recognition result for each enhanced image.

[0111] Step 420 : Determine the text information in the vehicle list image based on each character recognition result.

[0112] The multimodal resolution enhancement process may include at least two or three of the following enhancement methods: nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation. The first enhancement method involves upsampling the vehicle list image to the target resolution using nearest neighbor interpolation and then performing Gaussian filtering. The second enhancement method involves upsampling the image to the target resolution using bicubic interpolation. The third enhancement method involves upsampling the image to the target resolution using bilinear interpolation and then performing histogram equalization.

[0113] For the first enhanced image, during the upsampling process, each pixel of the pre-processed vehicle list image is mapped to the nearest position of the target resolution (for example, 300dpi) without generating new pixel values. The upsampling process retains the sharp edges of the original image (i.e., the pre-processed vehicle list image) but amplifies the jagged edges. Based on this, Gaussian filtering is used to process the edges, such as Gaussian low-pass filtering (for example, Gaussian filtering parameter standard deviation The value is 0, 0.5, 1.0, 1.5 or 2, and the convolution kernel size is 3×3 or other appropriate values), and the first enhanced image is obtained.

[0114] For the second enhanced image, during the upsampling process, the corresponding new pixel values ​​in the enhanced image are calculated based on the weighted average of the nearest neighbor pixels, resulting in the second enhanced image. The weights can be determined by a cubic function. The number of nearest neighbor pixels can be determined based on the resolution of the original image, so that the resolution of the enhanced image after upsampling reaches the target resolution. For example, a weighted average can be determined based on the nine nearest neighbor pixels (a 3×3 neighborhood).

[0115] For the third enhanced image, the upsampling process also calculates the corresponding new pixel value in the enhanced image based on the weighted average of the nearest neighbor pixels, but the weights are inversely proportional to the distance between the nearest neighbor pixels and the target pixel. Histogram equalization is applied to the resulting new pixel values ​​to produce the third enhanced image. Similarly, the number of nearest neighbor pixels selected can be determined based on the resolution of the original image.

[0116] Each OCR model performs OCR recognition on an enhanced image and obtains the character recognition result corresponding to the enhanced image. For example, if M OCR models and N enhanced images are set, M×N character recognition results can be obtained. For example, if M=2 (Tesseract-OCR model, Transformer-based OCR model) and N=3, 6 character recognition results can be output.

[0117] Taking the Tesseract-OCR model as an example, for the output enhanced image, the Tesseract-OCR model uses linear interpolation and unsharp masking to segment the text lines to form a single-line image, then calls the trained LSTM network for feature extraction, calls CTC decoding on the obtained feature vector, obtains the corresponding recognized character and the confidence of the character, and outputs the character recognition result.

[0118] For the Transformer-based OCR model, after image segmentation, the encoder extracts visual features, and the decoder decodes the extracted feature vector to obtain the corresponding recognized characters and the confidence of the characters, and outputs the character recognition results.

[0119] The character recognition results output by each OCR model include not only the characters but also the confidence level of each recognized character.

[0120] It is understandable that the characters in the multiple character recognition results output are not completely consistent. There must be some characters with incorrect recognition and some characters with correct recognition. Based on this, for the inconsistent characters in the recognition results, a weighted voting method can be used to select characters, and one character is selected from the inconsistent characters to finally form the text information in the vehicle checklist picture.

[0121] For example, for the character "准" in the "Emission Standard" field in the vehicle checklist, among the 6 character recognition results output, the corresponding recognized characters are: 准, 准, 淮, 堆, 准, 淮; the corresponding confidence levels are 83%, 81%, 82%, 78%, 79%, 77% respectively. Then through voting analysis, it is finally determined that the corresponding character is "准".

[0122] In this embodiment, by setting at least one OCR model and performing multiple image enhancement processes on the vehicle checklist picture with a low resolution, each OCR model outputs a character recognition result for each enhanced processing image, and based on the multiple character recognition results, the text information in the vehicle checklist picture is finally determined, which can improve the accuracy of character recognition for low-resolution images and ultimately improve the accuracy of environmental protection vehicle verification.

[0123] In one embodiment, each character recognition result includes the recognized character and the original confidence level of the corresponding character. As Figure 5 shown, step 430 includes:

[0124] Step 510, perform cross-model calibration on the original confidence levels output by each OCR model to obtain the standard confidence level.

[0125] Step 520, select characters according to the recognized characters and the corresponding standard confidence levels in each character recognition result.

[0126] Step 530, form the text information in the vehicle checklist picture based on the selected characters.

[0127] In this embodiment, the original confidence level represents the confidence level calculated by the corresponding OCR model according to its own confidence evaluation logic. It is understandable that the confidence evaluation logics of different OCR models are not necessarily the same. For example, the confidence level of a certain model is generally high, while the confidence level given by a certain model is generally conservative, but their recognition capabilities are basically the same. Based on this, for the confidence levels of each character obtained by each model, they can be cross-model calibrated to obtain a normalized standard confidence level, so that the confidence levels of the characters output between different models can be compared.

[0128] Specifically, a calibration dataset, independent of the training and test datasets, can be further configured. This calibration dataset contains calibration images and their corresponding real-world character sequences. Each OCR model runs on this calibration dataset, recording the original confidence c0 of each output character and whether the character prediction was accurate (e.g., outputting "1" for accuracy and "0" for error). A logistic regression model is trained on the original confidence by configuring a corresponding calibrator (e.g., using Platt Scaling). This original confidence is then mapped to the actual accuracy rate, and the calibrator is used to convert the original confidence into a calibrated confidence.

[0129] The obtained calibration confidence levels are normalized to form a standard confidence level that can be compared between various OCR models.

[0130] Specifically, the original confidence c0 can be mapped to the corresponding standard confidence c1 according to the following piecewise linear function:

[0131] ,

[0132] Among them, k1, k2, b1, b2, and c2 are preset confidence calibration parameters, c2 represents the segmentation threshold of the original confidence c0, k1 and b1 are the first piecewise linear function parameters taken when the original confidence c0 is less than the segmentation threshold c2, and k2 and b2 are the second piecewise linear function parameters taken when the original confidence c0 is greater than or equal to the segmentation threshold c2. The values ​​of the above calibration parameters can be calculated according to the values ​​of the above calibrator during the calibration process, and the calibration parameters of different OCR models are not necessarily the same. For example, the obtained piecewise linear function of the standard confidence c1 of the Tesseract-OCR model is:

[0133] .

[0134] For each calculated standard confidence, the character with the highest standard confidence can be selected for each character as the character in the text information in the vehicle list image. Alternatively, a weighted vote can be performed based on the confidence of each standard to select the character.

[0135] Furthermore, based on the various character recognition results output by the various OCR models, one character can be determined from the multiple character recognition results based on the standard confidence levels of the corresponding characters to serve as the primary recognition result. If the standard confidence level of a character in the primary recognition result exceeds a preset confidence threshold, the character is used as the character in the text information in the accompanying vehicle list image. For characters that do not exceed the preset confidence threshold, the characters recognized in other character recognition results and their standard confidence levels are further referenced to determine the final character as the character in the text information in the accompanying vehicle list image, thereby further improving the accuracy of character recognition.

[0136] Furthermore, a weighted voting calculation can be performed on the standard confidence of the characters to be selected in the various character recognition results output by each OCR model to calculate the weighted voting value of the standard confidence of each character to be selected, and then based on the weighted voting value of the standard confidence of each character to be selected, the character to be selected with the largest weighted voting value is selected as the character in the text information in the vehicle list picture, thereby further improving the accuracy of character recognition.

[0137] In one embodiment, the weighted voting value The calculation formula can be:

[0138] ;

[0139] .

[0140] in, It is the set of all characters to be selected. is the set of all characters to be selected A sample value of a character to be selected in ; Indicates the sample value equal to the character to be selected The i-th character to be selected; Indicates all satisfaction The index set of characters to be selected. Indicates all satisfaction The number of characters to be selected. Indicates the sample value equal to the character to be selected The i-th character to be selected The standard confidence level of w best Indicates the weighted voting value of the selection The maximum sample value of the character to be selected The character selected as the final vote.

[0141] In this embodiment, the weighted voting value The weighted voting calculation prevents over-selection of small sample sizes of candidate characters due to their higher average confidence levels. Furthermore, for multiple candidate character sample values ​​with the same sample size, the sample values ​​of the candidate characters with higher average confidence levels are favored. When the sample sizes of each candidate character vary, this weighted voting mechanism strikes a balance between average and sample size, significantly improving the accuracy of the final recognized characters.

[0142] For example, for a certain character to be selected, the three character recognition results involved in the selection and their corresponding standardized confidence levels are shown in Table 1 below.

[0143] Table 1

[0144]

[0145] In Table 1, the various character recognition results output by each OCR model have two sample values, namely "National VI Emission" and "National V Emission". Through the above weighted voting calculation, the weighted voting values ​​of "National VI Emission" and "National V Emission" are 116 and 90 respectively, and the final character w is obtained. best It is "National VI Emission".

[0146] In one embodiment, step 520 includes: for each character recognition result, when the characters recognized at the same position are inconsistent, performing a weighted voting calculation on each character at the same position based on a standard confidence level, and selecting a candidate character from a plurality of characters based on the weighted voting calculation result.

[0147] In this embodiment, the weighted bidding calculation method can be calculated according to the above formula, and the character with the highest score is selected as the candidate character.

[0148] In one embodiment, Figure 6 As shown, step 530 includes:

[0149] Step 610: forming initial text information based on the selected candidate characters.

[0150] Step 620 , performing a vehicle information compliance check on the initial text information, and identifying candidate characters in the initial text information that do not comply with the vehicle information compliance as characters to be corrected.

[0151] Step 630: Correct the character to be corrected based on each character recognition result to obtain a corrected character corresponding to the character to be corrected.

[0152] Step 640: forming text information in the vehicle list image based on the initial text information and the corrected characters.

[0153] In this embodiment, for the initially formed text information, it can be detected whether it conforms to the vehicle information standardization check. For example, for the text "Emission Standard" in the initial text information, it can be recognized that it does not conform to the vehicle information standardization, and then the candidate character "Emission Standard" is taken as the character to be corrected. For the character to be corrected, it can be corrected based on semantic recognition by combining the corresponding content in each character recognition result. For example, the candidate character "Emission Standard" can be finally corrected to "Emission Standard".

[0154] By further performing standardization checks on the initially formed text information and making corrections according to the standardization check results, the accuracy of the text information is further improved.

[0155] In one embodiment, as Figure 7 shown, a vehicle information processing device for an environmental protection vehicle is provided. The device includes:

[0156] An image acquisition module 710, configured to acquire a vehicle list image to be recognized;

[0157] An identification extraction module 720, configured to identify a corresponding first vehicle identification from the identification code in the vehicle list image;

[0158] An OCR recognition module 730, configured to perform multi-modal resolution enhancement processing on the vehicle list image to generate at least two enhanced images, call at least one OCR model to extract the text information in the vehicle list image from at least two enhanced images; and identify the corresponding first environmental protection key information according to the extracted text information;

[0159] A verification module 740, configured to compare the first vehicle identification and the first environmental protection key information with known vehicle information to determine whether the corresponding environmental protection vehicle passes the verification.

[0160] In one embodiment, the verification module 740 is further configured to compare whether the first vehicle identification is exactly the same as the second vehicle identification in the known vehicle information; verify the authenticity of the corresponding vehicle list based on the first environmental protection key information; when the vehicle identifications are exactly the same and the vehicle list is a genuine list, compare whether the first environmental protection key information matches the second environmental protection key information in the known vehicle information. If they match, it is determined that the verification passes; if the vehicle identifications are not exactly the same, or the vehicle list is not a genuine list, or the environmental protection key information does not match, it is determined that the verification fails.

[0161] In one embodiment, the verification module 740 is also used to extract the corresponding first key field and the first field content corresponding to the first key field from the first environmental protection key information; identify the overlap rate between the extracted key field and the standard key field in the standard key field set required by the vehicle manifest; identify whether each first field content meets the field content specification of the corresponding overlapping standard key field, and calculate the corresponding content standardization according to the specification result; when the overlap rate exceeds the corresponding overlap rate threshold, and the content standardization exceeds the corresponding standardization threshold, the vehicle manifest is determined to be a real list; when the overlap rate does not exceed the corresponding overlap rate threshold, or the content standardization does not exceed the corresponding standardization threshold, the vehicle manifest is determined not to be a real list.

[0162] In one embodiment, the OCR recognition module 730 is also used to perform multimodal resolution enhancement processing on the vehicle list image to generate at least two enhanced images, wherein the first enhanced image uses nearest neighbor interpolation to upsample the vehicle list image to the target resolution and then performs Gaussian filtering, the second enhanced image uses bicubic interpolation to upsample the image to the target resolution; the third enhanced image uses bilinear interpolation to upsample the image to the target resolution; call at least one preset OCR model to perform text recognition on each enhanced image, each OCR model outputs a character recognition result for each enhanced image, and the OCR model includes a Tesseract-OCR model; and determine the text information in the vehicle list image based on each character recognition result.

[0163] In one embodiment, each character recognition result includes the recognized character and the original confidence of the corresponding character; the OCR recognition module 730 is also used to perform cross-model calibration on the original confidence output by each OCR model to obtain a standard confidence; character selection is performed based on the recognized characters in each character recognition result and the corresponding standard confidence; and text information in the vehicle list picture is formed based on the selected characters.

[0164] In one embodiment, the OCR recognition module 730 is also used to perform weighted voting calculation on each character in the same position based on the standard confidence level for each character recognition result when the characters recognized in the same position are inconsistent, and select a candidate character from multiple characters based on the weighted voting calculation result.

[0165] In one embodiment, the OCR recognition module 730 is also used to form initial text information based on the selected candidate characters; perform vehicle information standardization verification on the initial text information, and identify candidate characters in the initial text information that do not meet the vehicle information standardization as characters to be corrected; correct the characters to be corrected based on each character recognition result to obtain the corrected characters corresponding to the characters to be corrected; and form the text information in the vehicle list picture based on the initial text information and the corrected characters.

[0166] In one embodiment, the OCR recognition module 730 is further configured to map the original confidence c0 to the corresponding standard confidence c1 according to the following piecewise linear function:

[0167] ,

[0168] Among them, k1, k2, b1, b2, and c2 are preset confidence calibration parameters.

[0169] In one embodiment, a computer-readable storage medium is provided, on which executable instructions are stored. When the instructions are executed by a processor, the processor executes the steps in the above-mentioned method embodiments.

[0170] In one embodiment, an electronic device is also provided, comprising one or more processors; a memory, wherein one or more programs are stored in the memory, wherein when the one or more programs are executed by one or more processors, the one or more processors execute the steps in the above-mentioned method embodiments.

[0171] In one embodiment, Figure 8 , which shows a schematic diagram of the structure of an electronic device for implementing an embodiment of the present application. Electronic device 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 802 or programs loaded from a storage unit 808 into a random access memory (RAM) 803. RAM 803 also stores various programs and data required for the operation of electronic device 800. CPU 801, ROM 802, and RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to bus 804.

[0172] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, mouse, and the like; an output section 807 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 808 including a hard disk; and a communication section 809 including a network interface card such as a LAN card or a modem. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. Removable media 811, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 810 as needed, so that computer programs read from the removable media can be installed in the storage section 808 as needed.

[0173] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer-readable medium carrying instructions. In such embodiments, the instructions can be downloaded and installed from a network via the communication portion 809 and / or installed from removable media 811. When the instructions are executed by the central processing unit (CPU) 801, the various method steps described in this application are performed.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

[0175] Furthermore, those skilled in the art will appreciate that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this application and to form different embodiments. For example, all of the above embodiments may be used in any combination. The information disclosed in this background section is intended solely to enhance understanding of the overall background of this application and should not be construed as an admission or any form of implication that such information constitutes prior art known to those skilled in the art.

Claims

1. A method for processing on-board information of an environmentally friendly vehicle, characterized in that: The method comprises: Get the vehicle list image to be identified; Identify the corresponding first vehicle identifier from the identification code in the vehicle list image; performing multimodal resolution enhancement processing on the vehicle list image to generate at least two enhanced images, wherein the first enhanced image is obtained by upsampling the vehicle list image to a target resolution using nearest neighbor interpolation and then performing Gaussian filtering, the second enhanced image is obtained by upsampling the vehicle list image to the target resolution using bicubic interpolation, and the third enhanced image is obtained by upsampling the vehicle list image to the target resolution using bilinear interpolation; Calling at least one preset OCR model to perform text recognition on each enhanced image, each OCR model outputting a character recognition result for each enhanced image, the OCR model including a Tesseract-OCR model, and determining text information in the vehicle list image based on each character recognition result; Identifying the corresponding first environmental protection key information according to the extracted text information; Based on the comparison between the first vehicle identification, the first environmental protection key information and known vehicle information, it is determined whether the corresponding environmental protection vehicle has passed the verification.

2. The method according to claim 1, characterized in that The comparing the first vehicle identifier, the first environmental protection key information, and known vehicle information to determine whether the corresponding environmentally friendly vehicle has passed the verification includes: comparing the first vehicle identifier and the second vehicle identifier in the known vehicle information to determine whether they are completely consistent; Verifying the authenticity of the corresponding on-board manifest based on the first environmental protection key information; When the vehicle identification is completely consistent and the accompanying vehicle list is a genuine list, the first environmental protection key information is compared with the second environmental protection key information in the known vehicle information to see if they match. If so, the verification is determined to be successful. If the vehicle identification is not completely consistent, or the accompanying list is not a real list, or the key environmental protection information does not match, the verification will be deemed to have failed.

3. The method according to claim 2, characterized in that The verifying the authenticity of the corresponding on-board manifest based on the first environmental protection key information includes: Extracting a corresponding first key field and a first field content corresponding to the first key field from the first environmental protection key information; Identify the overlap rate between the extracted first key field and the standard key fields in the standard key field set required by the vehicle manifest; Identify whether the content of each first field complies with the field content specification of the corresponding overlapping standard key field, and calculate the corresponding content standardization degree based on the standardization result; When the coincidence rate exceeds the corresponding coincidence rate threshold, and the content standardization exceeds the corresponding standardization threshold, the on-board manifest is determined to be a true manifest; When the overlap rate does not exceed the corresponding overlap rate threshold, or the content standardization does not exceed the corresponding standardization threshold, it is determined that the on-board manifest is not a true manifest.

4. The method according to claim 1, wherein Each character recognition result includes the recognized character and the original confidence of the corresponding character; The determining of the text information in the vehicle list image according to each character recognition result includes: The original confidence output by each OCR model is calibrated across models to obtain the standard confidence; Character selection is performed based on the recognized characters in each character recognition result and the corresponding standard confidence level; The text information in the vehicle list picture is formed based on the selected characters.

5. The method according to claim 4, characterized in that The character selection according to the recognized characters and the corresponding standard confidence in each character recognition result includes: for each character recognition result, when the recognized characters belonging to the same position are inconsistent, performing a weighted voting calculation on each character in the same position based on the standard confidence, and selecting a candidate character from a plurality of characters based on the weighted voting calculation result; The forming of text information in the vehicle list picture based on the selected characters includes: forming initial text information based on the selected candidate characters; Performing a vehicle information standardization check on the initial text information, and identifying candidate characters in the initial text information that do not conform to the vehicle information standardization as characters to be corrected; Correcting the character to be corrected based on each character recognition result to obtain a corrected character corresponding to the character to be corrected; The text information in the vehicle list picture is formed according to the initial text information and the corrected characters.

6. The method according to claim 4, characterized in that The cross-model calibration of the original confidence output by each OCR model to obtain the standard confidence includes: mapping the original confidence c0 to the corresponding standard confidence c1 according to the following piecewise linear function: ， Among them, k1, k2, b1, b2, and c2 are preset confidence calibration parameters.

7. An onboard information processing device for an environmentally friendly vehicle, characterized in that: The device comprises: Image acquisition module, used to obtain the vehicle list image to be identified; an identification extraction module, configured to identify a corresponding first vehicle identification from the identification code in the vehicle list image; An OCR recognition module is configured to perform multimodal resolution enhancement processing on the on-board list image to generate at least two enhanced images, wherein the first enhanced image uses nearest neighbor interpolation to upsample the on-board list image to a target resolution and then performs Gaussian filtering, the second enhanced image uses bicubic interpolation to upsample the on-board list image to a target resolution, and the third enhanced image uses bilinear interpolation to upsample the on-board list image to a target resolution; calling at least one preset OCR model to perform text recognition on each enhanced image, each OCR model outputting a character recognition result for each enhanced image, the OCR model including a Tesseract-OCR model, determining text information in the on-board list image based on each character recognition result; and identifying corresponding first environmental protection key information based on the extracted text information; The verification module is used to compare the first vehicle identifier, the first environmental protection key information and known vehicle information to determine whether the corresponding environmental protection vehicle has passed the verification.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores executable instructions, which, when executed by a processor, enable the processor to perform the method according to any one of claims 1 to 6.

9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs, which, when executed by the one or more processors, causes the one or more processors to perform the method according to any one of claims 1 to 6.

Images