Quality evaluation method, device and equipment for single-cycle vehicle-mounted CAN network data

Abstract

The invention relates to the technical field of image processing, and discloses a quality evaluation method, device and equipment for single-cycle vehicle-mounted CAN network data, and the method comprises the steps: obtaining a standard vehicle-mounted CAN network data set D1 and to-be-evaluated single-cycle vehicle-mounted CAN network data D2; identifying a period start CANID of the D1; performing standardization processing according to the period starting CANID, and extracting a plurality of single-period D1 data segments and single-period D2 data segments; according to a coloring classification system, the multiple single-cycle D1 data segments are converted into a color bar chart to obtain a reference color bar chart data set D1sts, and the single-cycle D2 data segments are converted into color bar charts to obtain a to-be-evaluated color bar chart D2st; calculating a feature reference library by using an unsupervised image feature extractor; and calculating the similarity between the to-be-evaluated feature vector and the feature mean vector, and judging the quality grade of D2 according to the similarity. According to the method, the quality evaluation accuracy and efficiency of the single-cycle vehicle-mounted CAN network data can be improved.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, and in particular relates to a method, apparatus and equipment for quality evaluation of single-cycle vehicle-mounted CAN network data. Background Technology

[0002] With the rapid development of intelligent connected vehicle technology, the Controller Area Network (CAN), as the neural hub of internal vehicle communication, has become a core resource for vehicle status monitoring, fault diagnosis, and autonomous driving algorithm training. In particular, the quality evaluation of data generated by the CAN network directly affects the reliability and security of downstream data applications. Currently, in the field of CAN network data quality evaluation, existing technologies typically involve three main stages: data processing, quality assessment, and deep learning model application. However, many shortcomings still exist in practical applications.

[0003] In the processing of vehicle CAN network data, existing technologies mostly rely on general-purpose tools such as SQL and Python for basic analysis, including data filtering, field extraction, and redundant data removal. While these methods can perform preliminary data organization and infer data periodicity by identifying the distribution characteristics of CAN_IDs, they lack standardized procedures for extracting "single-cycle" data. Specifically, existing technologies use relatively crude methods to identify the periodic distribution of CAN_IDs, making it difficult to accurately locate the "CycleStart CAN_ID," resulting in inaccurate segmentation of single-cycle data. Furthermore, the lack of unified standards for data redundancy removal and field filtering makes it easy to lose critical information or leave behind invalid data during processing, thus affecting the data foundation for subsequent quality evaluation.

[0004] In the specific implementation of data quality assessment, existing technologies mainly employ two types of methods: one is rule-based manual assessment, which involves manually verifying whether the data meets standards by pre-setting rules such as CAN_ID integrity and transmission stability; the other is a method based on simple statistical analysis, which judges data by calculating indicators such as missing rate and outlier percentage. However, these methods exhibit significant limitations when faced with large-scale generated data: manual assessment heavily relies on expert experience, is highly subjective, and extremely inefficient, failing to meet the needs of rapid screening; while simple statistical analysis can only cover surface-level quality indicators, failing to delve into the inherent reliability and consistency of the data, and its evaluation dimensions are too simplistic.

[0005] In recent years, some existing technologies have begun to introduce deep learning models to assist in data quality assessment, but most of them adopt supervised learning mode, that is, rely on a large number of data samples with labeled quality levels for training. This approach is poorly adapted to the vehicle CAN network data scenario: on the one hand, vehicle CAN network data lacks a unified quality labeling standard, and the cost of obtaining a large amount of high-quality labeled data is extremely high; on the other hand, general deep learning models are not optimized for the special characteristics of vehicle CAN network data (such as strict periodicity, temporal sequence, strong correlation between CAN_ID and the data length DLC field, etc.), and the feature extraction is not targeted enough, resulting in low accuracy and efficiency of quality assessment of single-cycle vehicle CAN network data. Summary of the Invention

[0006] This invention provides a method, apparatus, and device for quality evaluation of single-cycle vehicle CAN network data, which can improve the accuracy and efficiency of quality evaluation of single-cycle vehicle CAN network data.

[0007] To achieve the above objectives, the present invention provides a method for quality evaluation of single-cycle vehicle CAN network data, comprising: Obtain the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; Identify the cycle start CAN_ID of the standard automotive CAN network dataset D1; Based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted. Based on the pre-built CAN_ID color classification system, multiple single-period D1 data segments are converted into color bar maps to obtain the benchmark color bar map dataset D1sts, and single-period D2 data segments are converted into color bar maps to obtain the color bar map to be evaluated D2st. An unsupervised image feature extractor is used to extract features from the benchmark color bar image dataset D1sts to obtain benchmark feature vectors. The feature mean vector of the benchmark feature vectors is then calculated to obtain a feature benchmark library. An unsupervised image feature extractor is used to extract the feature vector to be evaluated from the color bar image D2st, and the similarity between the feature vector to be evaluated and the mean vector of features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity.

[0008] To address the aforementioned problems, the present invention also provides a quality evaluation device for single-cycle vehicle CAN network data, the device comprising: The data acquisition module is used to acquire the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; The color bar conversion module is used to identify the cycle start CAN_ID of the standard vehicle CAN network dataset D1; based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted; according to the pre-constructed CAN_ID color classification system, the multiple single-cycle D1 data segments are converted into color bars to obtain the baseline color bar dataset D1sts, and the single-cycle D2 data segments are converted into color bars to obtain the color bar dataset D2st to be evaluated; The quality assessment module is used to extract features from the baseline color bar dataset D1sts using an unsupervised image feature extractor, obtain baseline feature vectors, and calculate the feature mean vector of the baseline feature vectors to obtain a feature benchmark library. It also uses an unsupervised image feature extractor to extract the feature vector to be evaluated from the color bar dataset D2st, calculates the similarity between the feature vector to be evaluated and the feature mean vector in the feature benchmark library, and determines the quality level of the single-cycle vehicle CAN network data D2 to be evaluated based on the similarity.

[0009] To address the above problems, the present invention also provides an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the quality evaluation method for single-cycle vehicle CAN network data described above.

[0010] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the aforementioned method for evaluating the quality of single-cycle vehicle CAN network data.

[0011] This invention obtains a standard automotive CAN network dataset D1 and constructs a unified quality evaluation standard. Furthermore, by identifying the cycle start CAN_ID of the standard automotive CAN network dataset D1, it is possible to periodically split the dataset and convert the data into a structured and visual representation through a CAN_ID classification and coloring system. Additionally, by standardizing the standard automotive CAN network dataset D1 and the single-cycle automotive CAN network data D2 to be evaluated based on the cycle start CAN_ID, information misalignment or loss caused by inconsistent data slicing can be eliminated. Moreover, by extracting multiple single-cycle D1 data segments from the standard automotive CAN network dataset D1 and the single-cycle D2 data segment from the single-cycle automotive CAN network data D2 to be evaluated based on the cycle start CAN_ID, single-cycle data can be accurately extracted. Based on the pre-constructed CAN_ID color classification system, multiple single-cycle D1 data segments are converted into color bars to obtain the baseline color bar dataset D1sts, and single-cycle D2 data segments are converted into color bars to obtain the color bar dataset D2st to be evaluated. This transforms time-series data into the image domain, better reflecting the deeper information of the time-series data. In addition, an unsupervised image feature extractor is used to extract features from the baseline color bar dataset D1sts to obtain baseline feature vectors. The mean vector of the baseline feature vectors is then calculated to obtain a feature benchmark library, which can solve the problems of high annotation costs and inconsistent annotation standards. Finally, the similarity between the feature vector to be evaluated and the mean vector of the features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity, which can improve the accuracy and efficiency of quality evaluation of single-cycle vehicle CAN network data. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a method for evaluating the quality of single-cycle vehicle CAN network data according to an embodiment of the present invention. Figure 2 This is a color bar diagram illustrating a method for evaluating the quality of single-cycle vehicle CAN network data according to an embodiment of the present invention. Figure 3 A functional block diagram of a single-cycle vehicle CAN network data quality evaluation device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device for implementing the quality evaluation method for single-cycle vehicle CAN network data, as provided in an embodiment of the present invention.

[0013] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0015] This application provides a method for quality evaluation of single-cycle vehicle CAN network data. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for quality evaluation of single-cycle vehicle CAN network data can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0016] Reference Figure 1 The diagram shown is a flowchart illustrating a method for evaluating the quality of single-cycle vehicle CAN network data according to an embodiment of the present invention. In this embodiment, the method for evaluating the quality of single-cycle vehicle CAN network data includes: S1. Obtain the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated.

[0017] Understandably, the Standard Vehicle CAN Network Dataset D1 refers to the benchmark reference data for quality evaluation. It comes from a vehicle CAN network that has undergone rigorous verification, long-term normal operation, and qualified quality. The Standard Vehicle CAN Network Dataset D1 contains complete single-cycle data characteristics, and the CAN_ID and DLC fields are free of anomalies. For example, the Standard Vehicle CAN Network Dataset D1 could be the car-hacking-dataset, which is an open-source dataset used for automotive network security research. This dataset can serve as a reliable benchmark dataset.

[0018] Understandably, the single-cycle vehicle CAN network data D2 to be evaluated refers to the data generated using a pre-built data generation model based on the standard vehicle CAN network dataset D1. The generated data must ensure that the data format is consistent with the standard vehicle CAN network dataset D1, including the CAN_ID field and the DLC field. The pre-built data generation model refers to a model that is pre-built based on the standard vehicle CAN network dataset D1 and can generate single-cycle CAN network data with the same data distribution as the standard vehicle CAN network dataset D1.

[0019] S2. Identify the periodic start CAN_ID of the standard vehicle CAN network dataset D1.

[0020] Understandably, the CycleStart CAN_ID refers to the key CAN message ID used to identify the start and end of a single data cycle. Statistical analysis of the standard automotive CAN network dataset D1 shows that the CycleStart CAN_ID must have a probability of appearing at the beginning of each cycle of at least 95%, and the time interval fluctuation must not exceed ±5%.

[0021] Specifically, the periodic start CAN_ID of the standard automotive CAN network dataset D1 is identified, including: The CAN_ID field of the standard vehicle CAN network dataset D1 is statistically analyzed using a preset data analysis script to obtain the total number of CAN_ID fields. The occurrence frequency and standard deviation of the time interval for each CAN_ID in the total number of CAN_ID fields are calculated in different time windows. Periodic CAN_ID field data is then constructed based on the occurrence frequency and standard deviation of the time interval. The CAN_ID corresponding to the start position of the period in the period CAN_ID field data is determined as the start CAN_ID of the period.

[0022] Understandably, the pre-defined data analysis scripts are built on Python data analysis libraries Pandas and NumPy, and can perform categorical statistical analysis on fields.

[0023] S3. Based on the cycle start CAN_ID, standardize the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated, and extract multiple single-cycle D1 data segments from the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments from the single-cycle vehicle CAN network data D2 to be evaluated.

[0024] Understandably, standardization refers to a series of preprocessing operations to eliminate data noise, such as deleting redundant information like timestamps.

[0025] Specifically, based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments from the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments from the single-cycle vehicle CAN network data D2 to be evaluated are extracted, including: Retain the CAN_ID and DLC fields in the standard vehicle CAN network dataset D1, and remove redundant fields in the standard vehicle CAN network dataset D1 other than the CAN_ID and DLC fields; The standard vehicle CAN network dataset D1 is divided into multiple independent single-cycle D1 data segments based on the cycle start CAN_ID. Retain the CAN_ID and DLC fields in the single-cycle vehicle CAN network data D2 to be evaluated, and delete redundant fields other than the CAN_ID and DLC fields in the single-cycle vehicle CAN network data D2 to be evaluated. Identify the cycle start CAN_ID in the single-cycle vehicle CAN network data D2 to be evaluated, and locate the start position of the single cycle based on the cycle start CAN_ID in the single-cycle vehicle CAN network data D2 to be evaluated. Remove the data before the first cycle start CAN_ID and after the second cycle start CAN_ID to obtain the complete single-cycle D2 data segment.

[0026] Understandably, the DLC (Data Length Code) field refers to the data length code field, which indicates the number of bytes in the data field of the CAN frame, that is, how many bytes of data the frame carries. The value range is usually 0-8, corresponding to the classic CAN, where each frame contains a maximum of 8 bytes of data.

[0027] Furthermore, before standardizing the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated based on the cycle start CAN_ID, the process further includes: performing a data check on the single-cycle vehicle CAN network data D2 to be evaluated to verify the field integrity and CAN_ID validity of the single-cycle vehicle CAN network data D2 to be evaluated.

[0028] Furthermore, a data check is performed on the single-cycle vehicle CAN network data D2 to be evaluated, verifying the field completeness and CAN_ID validity of the single-cycle vehicle CAN network data D2 to be evaluated, including: Level 1 inspection: Traverse the data records of the single-cycle vehicle CAN network data D2 to be evaluated, check whether each data record contains the CAN_ID field and the DLC field, count the number of data records with missing CAN_ID field or DLC field, if any data record has a missing CAN_ID field or DLC field, then D2 is determined to have failed the inspection and an evaluation conclusion of "incomplete data fields, unqualified quality" is generated. Second-level inspection: Based on the CAN_ID set in the standard vehicle CAN network dataset D1, verify whether all CAN_IDs in the single-cycle vehicle CAN network data D2 to be evaluated are within the valid range of the CAN_ID set. This verification includes comparing all CAN_IDs in the single-cycle vehicle CAN network data D2 to be evaluated with the CAN_ID set in the standard vehicle CAN network dataset D1 through set operations. If there are CAN_IDs in the single-cycle vehicle CAN network data D2 that do not belong to the standard vehicle CAN network dataset D1, then D2 is determined to have failed the inspection, and the evaluation conclusion "Invalid CAN_IDs exist, quality unqualified" is generated.

[0029] Understandably, set operations refer to comparisons between two sets.

[0030] S4. Based on the pre-built CAN_ID color classification system, convert multiple single-period D1 data segments into color bar charts to obtain the baseline color bar chart dataset D1sts, and convert single-period D2 data segments into color bar charts to obtain the color bar chart to be evaluated D2st.

[0031] Understandably, the pre-built CAN_ID color classification system refers to a rule that maps text data to image colors. It is constructed based on the transmission frequency of CAN_ID (such as 100Hz, 50Hz, etc.), same-frequency asynchronousity (same frequency but alternating occurrences), and same-frequency out-of-order (same frequency but fixed different positions). Its main function is to assign specific colors to CAN_IDs of different categories (such as the color corresponding to CAN_ID of category 1 being #FF2D55).

[0032] Understandably, the CAN_ID coloring classification system can be represented as shown in the table below:

[0033] Understandably, the different types and numbers of electronic control units on the CAN bus of different car manufacturers will lead to differences in the period and transmission frequency of CAN_ID. However, since the transmission rules of all CAN buses are the same, the appearance of CAN_ID always shows a periodicity. Therefore, even if different car manufacturers have different hardware that leads to data differences, a three-in-one approach of CAN_ID transmission frequency, same frequency but different sequence, and same frequency but different asynchrony can still be used for color classification, and the color classification system can be set freely.

[0034] Understandably, the color bar image is a square image generated by visualizing single-cycle CAN text data according to a pre-built CAN_ID color classification system.

[0035] Specifically, the step of converting multiple single-period D1 data segments into color bar charts to obtain a baseline color bar chart dataset D1sts based on a pre-constructed CAN_ID color classification system, and converting single-period D2 data segments into color bar charts to obtain the color bar chart to be evaluated D2st, includes: Query the CAN_ID category, CAN_ID range, occurrence period and transmission frequency in the CAN_ID coloring classification system in the single-cycle D1 data segment, assign the preset color of the color bar to the CAN_ID in the single-cycle D1 data segment, and assign the color bar length according to the length of the DLC field in the single-cycle D1 data segment; Query the CAN_ID category, CAN_ID range, occurrence period and transmission frequency in the CAN_ID coloring classification system in the single-cycle D2 data segment, assign the preset color of the color bar to the CAN_ID in the single-cycle D2 data segment, and assign the color bar length according to the length of the DLC field in the single-cycle D2 data segment; The color bar chart consists of several vertical color bars with corresponding heights and horizontal color bar lengths. The vertical color bars correspond to the maximum data volume of each single-cycle data segment, and the height of each color bar is fixed. The horizontal color bar length is set according to the DLC field length of each single-cycle data segment. If the DLC field length is less than 8, it is supplemented with black. The color of the color bars is determined according to the pre-built CAN_ID coloring classification system. After the baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st are constructed, the baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st are uniformly scaled to obtain the standard-sized baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st.

[0036] For example, the color bar diagram of this invention can be shaped as follows: the color bar diagram is 450×450 square pixels in size, with 45 color bars in the vertical direction, each corresponding to the maximum data volume of each cycle, and the height of each color bar is 10 pixels. The color bar diagram always has 45 bars. If there is insufficient data, the color bars without data are replaced with white. The horizontal direction is divided into 8 equal blocks. The length of the color bars is set according to the length of the DLC field (0-8) of each data, which reflects the data size transmitted corresponding to the CAN_ID of the vehicle CAN network data. If the length of the DLC field is less than 8, it is supplemented with black.

[0037] S5. Use an unsupervised image feature extractor to extract features from the baseline color bar image dataset D1sts, obtain the baseline feature vector, and calculate the feature mean vector of the baseline feature vector to obtain the feature baseline library.

[0038] It is understandable that an unsupervised image feature extractor refers to a deep learning model that does not require pre-labeling "qualified" or "unqualified" for training, but directly uses pre-trained weights to extract image features. The embodiments of this invention are constructed based on a pre-trained MobileNetV2 backbone network.

[0039] Understandably, the feature benchmark library refers to the set of feature vectors extracted from the color bar graphs generated from the benchmark color bar graph dataset D1sts.

[0040] Specifically, an unsupervised image feature extractor is used to extract features from the benchmark color bar image dataset D1sts. The construction of the unsupervised image feature extractor includes: Step 1: Load the pre-trained MobileNetV2 backbone network and remove the top-level classification head module of the MobileNetV2 backbone network; Step 2: Add a global average pooling layer after the output of the MobileNetV2 backbone network, and freeze the weight parameters of the MobileNetV2 backbone network; Step 3: Use a fully connected layer with a linear activation function as the feature dimensionality reduction module; An unsupervised image feature extractor is obtained through model processing in steps 1-3 above.

[0041] For example, the construction of an unsupervised image feature extractor can be achieved by the following implementation steps 1-5: Step 1: Load the original MobileNetV2 pre-trained backbone network: The Google-released MobileNetV2 pre-trained model (trained on the ImageNet dataset) is loaded according to the following rules: ① The input size is fixed at 224×224×3, which is the standard input size for RGB three-channel images; ② ImageNet pre-trained weights are loaded to utilize the general visual features learned from massive datasets and improve the feature representation capability in small dataset scenarios. Step 2: Modification of the two core architectures of MobileNetV2: ① Remove the top-level classification head module of MobileNetV2 (implemented by include_top=False): Delete the fully connected layer, softmax activation layer, and classification output layer of the original model. This module is dedicated to supervised classification tasks and has no effect in unsupervised feature extraction scenarios. Removing it avoids task-oriented distortion of features and preserves the original visual features of the image; ② Enable and solidify the global average pooling layer (GAP) (implemented by pooling='avg'): Add a global average pooling layer after the feature map output of the MobileNetV2 backbone network. Compress the 7×7×1280 high-dimensional feature map output by the backbone into a 1×1×1280 feature tensor, and finally flatten it into a one-dimensional 1280-dimensional feature vector, realizing "any input image → fixed-dimensional feature output", providing a foundation for subsequent feature dimensionality reduction and similarity calculation; Step 3: Backbone Network Weight Freeze Processing: Perform a weight freeze operation (trainable=False) on the modified MobileNetV2 basic backbone network, freezing all weight parameters of all convolutional layers, inverted residual blocks, and linear bottleneck layers, so that they will no longer participate in training; the reasons for freezing are: ① This task is unsupervised feature extraction, so there is no need to retrain the model, and the pre-trained weights already have the ability to learn general visual features; ② Avoid overfitting of the model in small dataset scenarios and ensure the stability and robustness of feature extraction; ③ Freezing can significantly improve the feature extraction inference speed and reduce memory usage.

[0042] Step 4: Lightweight Feature Dimensionality Reduction Module Design and Integration: A fully connected layer (Dense layer) is selected as the core module for feature dimensionality reduction. This module has a simple structure, high computational efficiency, and strong adaptability, making it the optimal choice for CNN feature dimensionality reduction. It is integrated into the feature output end of the MobileNetV2 backbone network to achieve linear mapping of feature dimensions. The core parameters of the dimensionality reduction layer are configured as follows: ① Input dimension: 1280 dimensions (output of Step 2); ② Output dimension: 128 dimensions (target feature dimension); ③ Activation function: linear activation (activation='linear'), disabling non-linear activation functions such as ReLU and Sigmoid. Step 5: Key Post-Processing Optimization of the Feature Extractor: Perform a standardization post-processing operation on the 128-dimensional feature vector output by the feature extractor. This step is necessary for all image feature vectors and is a prerequisite for subsequent feature library construction and similarity calculation. Otherwise, the quality of subsequent similarity calculations will be poor. The 128-dimensional eigenvectors of the color bar chart are normalized using the L2 norm, as shown in the following formula: ; in, For the first The 128-dimensional normalized eigenvector of the color bar chart. The output of the unsupervised image feature extractor is the first The original 128-dimensional feature vector of the color bar chart, For the first The L2 norm of the eigenvectors of the color bar graph.

[0043] Understandably, normalization has the following effects: ① It eliminates the differences in the numerical magnitude of feature vectors, retains only the directional information of the features, and improves the accuracy of cosine similarity calculation by 100%; ② After normalization, the magnitude of all feature vectors is always equal to 1, the feature space distribution is more regular, and the aggregation of similar image features is stronger; ③ It adapts to scenarios with uneven distribution of small datasets and reduces the interference of noisy features on the judgment results.

[0044] Understandably, linear activation can completely preserve the original numerical distribution and spatial relationships of features, while nonlinear activation leads to the loss of feature information and a decrease in discriminative ability, making it the only viable activation method for unsupervised feature dimensionality reduction.

[0045] For example, the feature benchmark library can be constructed using the following implementation steps 1-3: Step 1: Initialize the feature benchmark library: Initialize an empty feature set matrix (feature benchmark library implementation method) to store the 128-dimensional feature vectors of all original target class images in the benchmark color bar dataset D1sts. The storage format is a two-dimensional matrix, and the matrix dimension is defined as follows: ,in, The total number of original target class images is 128, and the feature vector dimension is 128. Each row of the matrix represents a 128-dimensional normalized feature vector of an image, and each column represents a dimension of the feature. Step 2: Feature Baseline Data Population: For each image in the normalized image dataset D1sts_norm (the baseline color bar image dataset used for feature extraction), after normalization, a normalized 128-dimensional feature vector is extracted using the constructed unsupervised image feature extractor. This vector is then stored row-wise in the initialized feature baseline dataset, resulting in a complete feature set matrix. ; Step 3: Calculate the feature mean vector: For the complete feature set matrix The mean vector of features is obtained by taking the mean of all 128-dimensional feature vectors in the dataset and calculating the mean of each dimension. Dimension The value is 1 × 128, and the calculation formula is as follows: ; in, ∈[1,128] represents the dimension index of the feature. This represents the number of images in the normalized image dataset D1sts_norm, which is also the number of rows in the feature set matrix. For the complete feature set matrix No. Color bar chart 1-dimensional eigenvalues.

[0046] S6. Use an unsupervised image feature extractor to extract the feature vector to be evaluated from the color bar image D2st, and calculate the similarity between the feature vector to be evaluated and the mean vector of features in the feature benchmark library. Determine the quality level of the single-cycle vehicle CAN network data D2 to be evaluated based on the similarity.

[0047] For example, determining the quality level of the single-cycle vehicle CAN network data D2 to be evaluated based on similarity can be achieved using the following implementation steps 1-3: Step 1: Extraction of the feature vector to be evaluated: The D2st color bar image is scaled and pixel normalized according to the standard in Step 6, and then input into the feature extractor module to extract the normalized 128-dimensional feature vector to be evaluated. =[ 1, 2,…, 128]; Step 2, Similarity Calculation: The cosine similarity algorithm is used to calculate the feature vector to be evaluated. Compared with the feature mean vector in the feature benchmark library similarity The calculation formula is: ; in, For vector dot product, The feature vector to be evaluated L2 norm, The L2 norm of the feature mean vector in the feature benchmark library; Step 3, Quality Level Determination: Preset Similarity Threshold , usually the optimal threshold Set to 0.85: ① If ≥ If the quality of the single-cycle vehicle CAN network data D2 to be evaluated is deemed acceptable, the evaluation conclusion "Similarity Qualified" will be output; ② If < If the quality of the single-cycle vehicle CAN network data D2 to be evaluated is deemed unqualified, the evaluation conclusion "similarity unqualified" will be output.

[0048] This invention obtains a standard automotive CAN network dataset D1 and constructs a unified quality evaluation standard. Furthermore, by identifying the cycle start CAN_ID of the standard automotive CAN network dataset D1, it is possible to periodically split the dataset and convert the data into a structured and visual representation through a CAN_ID classification and coloring system. Additionally, by standardizing the standard automotive CAN network dataset D1 and the single-cycle automotive CAN network data D2 to be evaluated based on the cycle start CAN_ID, information misalignment or loss caused by inconsistent data slicing can be eliminated. Moreover, by extracting multiple single-cycle D1 data segments from the standard automotive CAN network dataset D1 and the single-cycle D2 data segment from the single-cycle automotive CAN network data D2 to be evaluated based on the cycle start CAN_ID, single-cycle data can be accurately extracted. Based on the pre-constructed CAN_ID color classification system, multiple single-cycle D1 data segments are converted into color bars to obtain the baseline color bar dataset D1sts, and single-cycle D2 data segments are converted into color bars to obtain the color bar dataset D2st to be evaluated. This transforms time-series data into the image domain, better reflecting the deeper information of the time-series data. In addition, an unsupervised image feature extractor is used to extract features from the baseline color bar dataset D1sts to obtain baseline feature vectors. The mean vector of the baseline feature vectors is then calculated to obtain a feature benchmark library, which can solve the problems of high annotation costs and inconsistent annotation standards. Finally, the similarity between the feature vector to be evaluated and the mean vector of the features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity, which can improve the accuracy and efficiency of quality evaluation of single-cycle vehicle CAN network data.

[0049] Reference Figure 2 The diagram shown is a color bar diagram illustrating a method for evaluating the quality of single-cycle vehicle CAN network data according to an embodiment of the present invention.

[0050] like Figure 3 The diagram shown is a functional block diagram of a single-cycle vehicle CAN network data quality evaluation device provided in an embodiment of the present invention.

[0051] The quality evaluation device 100 for single-cycle vehicle CAN network data described in this invention can be installed in an electronic device. Depending on the functions implemented, the quality evaluation device 100 for single-cycle vehicle CAN network data may include a data acquisition module 101, a color bar graph conversion module 102, and a quality assessment module 103.

[0052] The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0053] In this embodiment, the functions of each module / unit are as follows: The data acquisition module 101 is used to acquire the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated.

[0054] The color bar conversion module 102 is used to identify the cycle start CAN_ID of the standard vehicle CAN network dataset D1; based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted; based on the pre-constructed CAN_ID color classification system, the multiple single-cycle D1 data segments are converted into color bars to obtain the baseline color bar dataset D1sts, and the single-cycle D2 data segments are converted into color bars to obtain the color bar dataset D2st to be evaluated.

[0055] The quality assessment module 103 is used to extract features from the baseline color bar dataset D1sts using an unsupervised image feature extractor to obtain baseline feature vectors, and calculate the feature mean vector of the baseline feature vectors to obtain a feature benchmark library; it also uses an unsupervised image feature extractor to extract the feature vector to be evaluated from the color bar dataset D2st, and calculates the similarity between the feature vector to be evaluated and the feature mean vector in the feature benchmark library, and determines the quality level of the single-cycle vehicle CAN network data D2 to be evaluated based on the similarity.

[0056] like Figure 4 The diagram shown is a structural schematic of an electronic device for implementing a single-cycle vehicle CAN network data quality evaluation method according to an embodiment of the present invention.

[0057] The electronic device may include a processor 10, a memory 11, a communication bus 12, and a communication interface 13. It may also include a computer program stored in the memory 11 and capable of running on the processor 10, such as a method program for evaluating the quality of single-cycle vehicle CAN network data.

[0058] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a method for evaluating the quality of single-cycle vehicle CAN network data), and calls data stored in the memory 11 to perform various functions of the electronic device and process data.

[0059] The memory 11 includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of an electronic device, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device, such as a plug-in portable hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc. Furthermore, the memory 11 can include both internal and external storage units of the electronic device. The memory 11 can be used not only to store application software and various types of data installed in the electronic device, such as the code of a single-cycle vehicle CAN network data quality evaluation method program, but also to temporarily store data that has been output or will be output.

[0060] The communication bus 12 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0061] The communication interface 13 is used for communication between the aforementioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, Bluetooth interface, etc.), typically used to establish communication connections between the electronic device and other electronic devices. The user interface may be a display, an input unit (such as a keyboard), or optionally, a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device and to display a visual user interface.

[0062] Figure 4 Only electronic devices with components are shown; those skilled in the art will understand that... Figure 4 The structure shown does not constitute a limitation on the electronic device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0063] For example, although not shown, the electronic device may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0064] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0065] The memory 11 in the electronic device stores a quality evaluation method program for single-cycle vehicle CAN network data, which is a combination of multiple instructions. When run in the processor 10, it can achieve the following: Obtain the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; Identify the cycle start CAN_ID of the standard automotive CAN network dataset D1; Based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted. Based on the pre-built CAN_ID color classification system, multiple single-period D1 data segments are converted into color bar maps to obtain the benchmark color bar map dataset D1sts, and single-period D2 data segments are converted into color bar maps to obtain the color bar map to be evaluated D2st. An unsupervised image feature extractor is used to extract features from the benchmark color bar image dataset D1sts to obtain benchmark feature vectors. The feature mean vector of the benchmark feature vectors is then calculated to obtain a feature benchmark library. An unsupervised image feature extractor is used to extract the feature vector to be evaluated from the color bar image D2st, and the similarity between the feature vector to be evaluated and the mean vector of features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity.

[0066] Specifically, the specific implementation method of the processor 10 for the above instructions can be referred to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, and will not be repeated here.

[0067] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0068] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following: Obtain the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; Identify the cycle start CAN_ID of the standard automotive CAN network dataset D1; Based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted. Based on the pre-built CAN_ID color classification system, multiple single-period D1 data segments are converted into color bar maps to obtain the benchmark color bar map dataset D1sts, and single-period D2 data segments are converted into color bar maps to obtain the color bar map to be evaluated D2st. An unsupervised image feature extractor is used to extract features from the benchmark color bar image dataset D1sts to obtain benchmark feature vectors. The feature mean vector of the benchmark feature vectors is then calculated to obtain a feature benchmark library. An unsupervised image feature extractor is used to extract the feature vector to be evaluated from the color bar image D2st, and the similarity between the feature vector to be evaluated and the mean vector of features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity.

[0069] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0070] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0071] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0072] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0073] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0074] The blockchain referred to in this invention is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0075] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0076] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for quality evaluation of single-cycle vehicle-mounted CAN network data, characterized in that, The method includes: Obtain the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; Identify the cycle start CAN_ID of the standard automotive CAN network dataset D1; Based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted. Based on the pre-built CAN_ID color classification system, multiple single-period D1 data segments are converted into color bar maps to obtain the benchmark color bar map dataset D1sts, and single-period D2 data segments are converted into color bar maps to obtain the color bar map to be evaluated D2st. An unsupervised image feature extractor is used to extract features from the benchmark color bar image dataset D1sts to obtain benchmark feature vectors. The feature mean vector of the benchmark feature vectors is then calculated to obtain a feature benchmark library. An unsupervised image feature extractor is used to extract the feature vector to be evaluated from the color bar image D2st, and the similarity between the feature vector to be evaluated and the mean vector of features in the feature benchmark library is calculated. The quality level of the single-cycle vehicle CAN network data D2 to be evaluated is determined based on the similarity.

2. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, The periodic start CAN_ID of the identification standard vehicle CAN network dataset D1 includes: The CAN_ID field of the standard vehicle CAN network dataset D1 is statistically analyzed using a preset data analysis script to obtain the total number of CAN_ID fields. The occurrence frequency and standard deviation of the time interval for each CAN_ID in the total number of CAN_ID fields are calculated in different time windows. Periodic CAN_ID field data is then constructed based on the occurrence frequency and standard deviation of the time interval. The CAN_ID corresponding to the start position of the period in the period CAN_ID field data is determined as the start CAN_ID of the period.

3. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, The standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized according to the cycle start CAN_ID, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and the single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted, including: Retain the CAN_ID and DLC fields in the standard vehicle CAN network dataset D1, and remove redundant fields in the standard vehicle CAN network dataset D1 other than the CAN_ID and DLC fields; The standard vehicle CAN network dataset D1 is divided into multiple independent single-cycle D1 data segments based on the cycle start CAN_ID. Retain the CAN_ID and DLC fields in the single-cycle vehicle CAN network data D2 to be evaluated, and delete redundant fields other than the CAN_ID and DLC fields in the single-cycle vehicle CAN network data D2 to be evaluated. Identify the cycle start CAN_ID in the single-cycle vehicle CAN network data D2 to be evaluated, and locate the start position of the single cycle based on the cycle start CAN_ID in the single-cycle vehicle CAN network data D2 to be evaluated. Remove the data before the first cycle start CAN_ID and after the second cycle start CAN_ID to obtain the complete single-cycle D2 data segment.

4. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, Before standardizing the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated based on the cycle start CAN_ID, the process further includes: performing a data check on the single-cycle vehicle CAN network data D2 to be evaluated to verify the field integrity and CAN_ID validity of the single-cycle vehicle CAN network data D2 to be evaluated.

5. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, The single-cycle vehicle CAN network data D2 to be evaluated is subjected to data inspection to verify the field integrity and CAN_ID validity of the single-cycle vehicle CAN network data D2 to be evaluated, including: Level 1 inspection: Traverse the data records of the single-cycle vehicle CAN network data D2 to be evaluated, check whether each data record contains the CAN_ID field and the DLC field, count the number of data records with missing CAN_ID field or DLC field, if any data record has a missing CAN_ID field or DLC field, then D2 is determined to have failed the inspection and an evaluation conclusion of "incomplete data fields, unqualified quality" is generated. Second-level inspection: Based on the CAN_ID set in the standard vehicle CAN network dataset D1, verify whether all CAN_IDs in the single-cycle vehicle CAN network data D2 to be evaluated are within the valid range of the CAN_ID set. This verification includes comparing all CAN_IDs in the single-cycle vehicle CAN network data D2 to be evaluated with the CAN_ID set in the standard vehicle CAN network dataset D1 through set operations. If there are CAN_IDs in the single-cycle vehicle CAN network data D2 that do not belong to the standard vehicle CAN network dataset D1, then D2 is deemed to have failed the inspection, and the evaluation conclusion "Invalid CAN_IDs exist, quality unqualified" is generated.

6. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, The process of converting multiple single-period D1 data segments into color bars to obtain a baseline color bar dataset D1sts, and converting single-period D2 data segments into color bars to obtain the color bar dataset D2st to be evaluated, based on the pre-constructed CAN_ID color classification system, includes: Query the CAN_ID category, CAN_ID range, occurrence period and transmission frequency in the CAN_ID coloring classification system in the single-cycle D1 data segment, assign the preset color of the color bar to the CAN_ID in the single-cycle D1 data segment, and assign the color bar length according to the length of the DLC field in the single-cycle D1 data segment; Query the CAN_ID category, CAN_ID range, occurrence period and transmission frequency in the CAN_ID coloring classification system in the single-cycle D2 data segment, assign the preset color of the color bar to the CAN_ID in the single-cycle D2 data segment, and assign the color bar length according to the length of the DLC field in the single-cycle D2 data segment; The color bar chart consists of several vertical color bars with corresponding heights and horizontal color bar lengths. The vertical color bars correspond to the maximum data volume of each single-cycle data segment, and the height of each color bar is fixed. The horizontal color bar length is set according to the DLC field length of each single-cycle data segment. If the DLC field length is less than 8, it is supplemented with black. The color of the color bars is determined according to the pre-built CAN_ID coloring classification system. After the baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st are constructed, the baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st are uniformly scaled to obtain the standard-sized baseline color bar chart dataset D1sts and the color bar chart to be evaluated D2st.

7. The method for quality evaluation of single-cycle vehicle CAN network data as described in claim 1, characterized in that, The method of using an unsupervised image feature extractor to extract features from the benchmark color bar chart dataset D1sts includes the following: Step 1: Load the pre-trained MobileNetV2 backbone network and remove the top-level classification head module of the MobileNetV2 backbone network; Step 2: Add a global average pooling layer after the output of the MobileNetV2 backbone network, and freeze the weight parameters of the MobileNetV2 backbone network; Step 3: Use a fully connected layer with a linear activation function as the feature dimensionality reduction module; An unsupervised image feature extractor is obtained through model processing in steps 1-3 above.

8. A quality evaluation device for single-cycle vehicle-mounted CAN network data, characterized in that, The apparatus is used to implement the quality evaluation method for single-cycle vehicle CAN network data as described in any one of claims 1 to 7, and the apparatus includes: The data acquisition module is used to acquire the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated; The color bar conversion module is used to identify the cycle start CAN_ID of the standard vehicle CAN network dataset D1; based on the cycle start CAN_ID, the standard vehicle CAN network dataset D1 and the single-cycle vehicle CAN network data D2 to be evaluated are standardized, and multiple single-cycle D1 data segments of the standard vehicle CAN network dataset D1 and single-cycle D2 data segments of the single-cycle vehicle CAN network data D2 to be evaluated are extracted; according to the pre-constructed CAN_ID color classification system, the multiple single-cycle D1 data segments are converted into color bars to obtain the baseline color bar dataset D1sts, and the single-cycle D2 data segments are converted into color bars to obtain the color bar dataset D2st to be evaluated; The quality assessment module is used to extract features from the baseline color bar dataset D1sts using an unsupervised image feature extractor, obtain baseline feature vectors, and calculate the feature mean vector of the baseline feature vectors to obtain a feature benchmark library. It also uses an unsupervised image feature extractor to extract the feature vector to be evaluated from the color bar dataset D2st, calculates the similarity between the feature vector to be evaluated and the feature mean vector in the feature benchmark library, and determines the quality level of the single-cycle vehicle CAN network data D2 to be evaluated based on the similarity.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the quality evaluation method for single-cycle vehicle CAN network data as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the quality evaluation method for single-cycle vehicle CAN network data as described in any one of claims 1 to 7.

Images