Vehicle defect identification model training method and vehicle defect input method

Abstract

The invention relates to a vehicle defect recognition model training method and a vehicle defect input method, and the method comprises the steps: obtaining an initial sample set which comprises first defect description texts corresponding to a plurality of vehicle defects and first vehicle defect labels corresponding to each first defect description text; acquiring a data set comprising second defect description texts corresponding to a plurality of vehicle defects; according to the initial sample set, performing parameter adjustment on the initial vehicle defect tag generation model to obtain a target vehicle defect generation model, inputting a plurality of second defect description texts into the target vehicle defect generation model, and obtaining a second vehicle defect tag corresponding to each second defect description text; constructing a target sample set by utilizing the initial sample set, the plurality of second defect description texts and the plurality of second vehicle defect labels; and according to the target sample set, carrying out training processing on the initial vehicle defect identification model to obtain a target vehicle defect identification model. Therefore, a normative target vehicle defect label is obtained.

Description

Technical Field

[0001] The present disclosure relates to the technical field of vehicle evaluation, and particularly to a method for training a vehicle defect recognition model, a method for entering vehicle defects, a device, an electronic device, a medium, and a vehicle. Background Art

[0002] Currently, when conducting quality inspection on a vehicle, a professional quality inspector needs to inspect the vehicle at the inspection site. Specifically, based on their own understanding and experience, the quality inspector inspects each component of the vehicle by observing, touching, or with the aid of tools to determine whether there are defects in the current vehicle. When the quality inspector discovers a defect, they need to determine information such as the type and description corresponding to the defect and enter it into the system; when there is no corresponding defect type in the detection system, or the quality inspector cannot find the corresponding defect type in the detection system based on their own understanding and experience, the quality inspector needs to customize defect label information such as the defect type and description corresponding to the defect and manually enter the defect into the system.

[0003] However, using the above method, for the same defect, due to different quality inspectors selecting different defect label information, there is a problem of non-standard defect label information entry. At the same time, since the defect label information is entered manually, the efficiency of entering vehicle defect label information is reduced. Summary of the Invention

[0004] To solve the above technical problems, the present disclosure provides a method for training a vehicle defect recognition model, a method for entering vehicle defects, a device, an electronic device, a medium, and a vehicle.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for training a vehicle defect recognition model, including:

[0006] Obtaining an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each of the first defect description texts, and obtaining a data set including second defect description texts corresponding to the multiple vehicle defects, where the first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels;

[0007] Adjusting the parameters of an initial vehicle defect label generation model according to the initial sample set of the first vehicle defect labels to obtain a target vehicle defect generation model;

[0008] Inputting multiple of the second defect description texts into the target vehicle defect generation model to obtain second vehicle defect labels corresponding to each of the second defect description texts, where the second vehicle defect labels include: the system sub-labels, the assembly sub-labels, the part sub-labels, and the defect sub-labels;

[0009] Construct a target sample set by using the initial sample set of the first vehicle defect labels, multiple pieces of the second defect description texts, and multiple pieces of the second vehicle defect labels. The target sample set includes: multiple defect description texts, and a vehicle defect label corresponding to each defect description text. The vehicle defect label includes: the system sub-label, the assembly sub-label, the part sub-label, and the defect sub-label;

[0010] Train an initial vehicle defect recognition model according to the target sample set to obtain a target vehicle defect recognition model. The target vehicle defect recognition model is used to obtain a target vehicle defect label corresponding to a vehicle defect description text to be recognized. The target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label.

[0011] Optionally, the obtaining of the initial sample set including the first defect description texts corresponding to multiple vehicle defects and the first vehicle defect labels corresponding to each first defect description text, and the obtaining of the data set including the second defect description texts corresponding to the multiple vehicle defects includes:

[0012] For each of the vehicle defects, obtain multiple defect description texts obtained by performing text recognition on the defect description voice data of each vehicle defect;

[0013] Among the multiple defect description texts, determine multiple first defect description texts, and perform manual marking processing on the multiple first defect description texts to obtain a first vehicle defect label corresponding to each first defect description text;

[0014] Construct the initial sample set according to the multiple first defect description texts and the first vehicle defect label corresponding to each first defect description text;

[0015] Among the multiple defect description texts, determine multiple second defect description texts, and construct the data set, where the multiple second defect description texts are other defect description texts in the multiple defect description texts except the multiple first defect description texts.

[0016] Optionally, the defect description voice data includes: standard defect description voice data and non-standard defect description voice data;

[0017] The obtaining of multiple defect description texts by performing text recognition on the defect description voice data of each vehicle defect for different vehicle defects includes:

[0018] Perform text recognition on the standard defect description voice data and the non-standard defect description voice data to obtain a standard defect description text and a non-standard defect description text;

[0019] Use the standard defect description text and the non-standard defect description text to construct the defect description text.

[0020] Optionally, the artificial marking process for the multiple first defect description texts to obtain the first vehicle defect label corresponding to each first defect description text includes:

[0021] Perform word segmentation on the multiple first defect description texts, and use the multiple target word segments corresponding to each first defect description text to determine the vehicle component information to which the vehicle defect corresponding to the first defect description text belongs;

[0022] According to the vehicle component information, perform a marking process on the first defect description text to obtain the system sub-label, the assembly sub-label, the part sub-label, and the defect sub-label included in the first vehicle defect label.

[0023] In a second aspect, an embodiment of the present disclosure provides a method for entering vehicle defects, including:

[0024] Obtain a vehicle defect description text to be recognized input by a user, where the vehicle defect description text to be recognized is used to describe a target defect of a vehicle;

[0025] Input the vehicle defect description text into a target vehicle defect recognition model, and obtain a target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model;

[0026] Wherein, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, the target vehicle defect recognition model is trained according to a target sample set, the target sample set includes: multiple defect description texts, and a vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

[0027] Optionally, the method further includes:

[0028] Perform a matching process on the target defect sub-label among multiple vehicle defect levels to obtain the target vehicle defect level to which the target defect belongs.

[0029] Optionally, the method further includes:

[0030] Save the text describing the vehicle defect to be recognized, the target vehicle defect label corresponding to the vehicle defect description text, and the target vehicle defect level to which the target defect belongs to a preset database;

[0031] Periodically use the preset database to update the target sample set, and use the updated target sample set to train the target vehicle defect recognition model.

[0032] In a third aspect, an embodiment of the present disclosure provides a vehicle defect recognition model training device, including:

[0033] An acquisition module, configured to acquire an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each of the first defect description texts, and acquire a data set including second defect description texts corresponding to the multiple vehicle defects, where the first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels;

[0034] A target vehicle defect generation model obtaining module, configured to adjust the parameters of an initial vehicle defect label generation model according to the initial sample set of the first vehicle defect labels to obtain a target vehicle defect generation model;

[0035] A second vehicle defect label acquisition module, configured to input multiple second defect description texts into the target vehicle defect generation model to acquire second vehicle defect labels corresponding to each of the second defect description texts, where the second vehicle defect labels include: the system sub-labels, the assembly sub-labels, the part sub-labels, and the defect sub-labels;

[0036] A target sample set construction module, configured to construct a target sample set by using the initial sample set of the first vehicle defect labels, multiple second defect description texts, and multiple second vehicle defect labels, where the target sample set includes: multiple defect description texts, and vehicle defect labels corresponding to each of the defect description texts;

[0037] A target vehicle defect recognition model training module, configured to perform training processing on an initial vehicle defect recognition model according to the target sample set to acquire a target vehicle defect recognition model, where the target vehicle defect recognition model is used to acquire target vehicle defect labels corresponding to a text describing a vehicle defect to be recognized, and the target vehicle defect labels include: target system sub-labels, target assembly sub-labels, target part sub-labels, and target defect sub-labels.

[0038] In a fourth aspect, an embodiment of the present disclosure provides a vehicle defect entry device, including:

[0039] A module for obtaining a text description of a vehicle defect to be recognized, which is used to obtain the text description of the vehicle defect to be recognized input by the user, and the text description of the vehicle defect to be recognized is used to describe the target defect of the vehicle;

[0040] A target vehicle defect label acquisition module, which is used to input the text description of the vehicle defect into a target vehicle defect recognition model, and obtain the target vehicle defect label corresponding to the text description of the vehicle defect output by the target vehicle defect recognition model;

[0041] Among them, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label. The target vehicle defect recognition model is trained according to a target sample set, and the target sample set includes: multiple defect description texts, and the vehicle defect label corresponding to each defect description text. The vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

[0042] In a fifth aspect, an embodiment of the present disclosure provides an electronic device, including:

[0043] One or more processors;

[0044] A storage device for storing one or more programs,

[0045] When the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any one of the first aspects.

[0046] In a sixth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method described in any one of the first aspects is implemented.

[0047] In a seventh aspect, an embodiment of the present disclosure provides a vehicle, including: a processor, a memory, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the method described in any one of the first aspects is implemented.

[0048] The technical solution provided by the embodiment of the present disclosure has the following advantages compared with the prior art:

[0049] The vehicle defect recognition model training method provided by the embodiments of the present disclosure includes: obtaining an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, and obtaining a data set including second defect description texts corresponding to multiple vehicle defects, where the first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; adjusting the parameters of the initial vehicle defect label generation model according to the initial sample set of the first vehicle defect labels to obtain a target vehicle defect generation model; inputting multiple second defect description texts into the target vehicle defect generation model to obtain second vehicle defect labels corresponding to each second defect description text, where the second vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; constructing a target sample set by using the initial sample set of the first vehicle defect labels, multiple second defect description texts, and multiple second vehicle defect labels, where the target sample set includes: multiple defect description texts, and vehicle defect labels corresponding to each defect description text, and the vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; training the initial vehicle defect recognition model according to the target sample set to obtain a target vehicle defect recognition model, where the target vehicle defect recognition model is used to obtain target vehicle defect labels corresponding to the vehicle defect description text to be recognized, and the target vehicle defect labels include: target system sub-labels, target assembly sub-labels, target part sub-labels, and target defect sub-labels. In the above process, by using the initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, the parameters of the initial vehicle defect label generation model are adjusted to obtain a target vehicle defect generation model, so that the second vehicle defect labels corresponding to multiple second defect description texts can be obtained by using the target vehicle defect generation model, thereby improving the efficiency of obtaining vehicle defect labels. Since the vehicle defect recognition model is trained according to the target sample set including the initial sample set, multiple second defect description texts, and multiple second vehicle defect labels to obtain a target vehicle defect recognition model, the target vehicle defect recognition model is used to recognize the vehicle defect description text to be recognized, so as to obtain standardized target vehicle defect labels, avoiding the problem of non-standard defect label information in the prior art, and eliminating the need for manual input of vehicle defect labels, thus improving the efficiency of inputting vehicle defect labels. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings are incorporated herein and form a part of this specification, showing embodiments consistent with the present disclosure and, together with the specification, are used to explain the principles of the present disclosure.

[0051] To more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0052] Figure 1 It is a schematic flowchart of a method for training a vehicle defect recognition model provided by an embodiment of the present disclosure;

[0053] Figure 2 It is a schematic flowchart of a method for entering vehicle defects provided by an embodiment of the present disclosure;

[0054] Figure 3 It is a schematic structural diagram of a vehicle defect recognition model training device provided by an embodiment of the present disclosure;

[0055] Figure 4 It is a schematic structural diagram of a vehicle defect entry device provided by an embodiment of the present disclosure;

[0056] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. Specific embodiments

[0057] In order to more clearly understand the above objects, features, and advantages of the present disclosure, the following will further describe the solutions of the present disclosure. It should be noted that, without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0058] Many specific details are set forth in the following description to facilitate a thorough understanding of the present disclosure, but the present disclosure can also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only a part of the embodiments of the present disclosure, rather than all of the embodiments.

[0059] Currently, for the quality inspection of vehicles, professional quality inspectors are required to conduct inspections on-site. Specifically, based on their understanding and experience, quality inspectors inspect various vehicle components by observing, touching, or using tools to determine whether there are defects in the current vehicle. When a quality inspector discovers a defect, they need to determine information such as the corresponding type and description of the defect and enter it into the system; when there is no corresponding defect type in the detection system, or when the quality inspector cannot find the corresponding defect type in the detection system according to their understanding and experience, the quality inspector needs to customize defect label information such as the defect type and description corresponding to the defect and manually enter the defect into the system. As a result, for the same defect, due to different defect label information defined by different quality inspectors, there is a problem of non-standardized defect label information entry. At the same time, since the defect label information is entered manually, the efficiency of entering vehicle defect label information is reduced.

[0060] Based on this, embodiments of the present disclosure provide a method for training a vehicle defect recognition model. By obtaining an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, a data set including second defect description texts corresponding to multiple vehicle defects is obtained. The first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels. According to the initial sample set of the first vehicle defect labels, the parameters of the initial vehicle defect label generation model are adjusted to obtain a target vehicle defect generation model. The multiple second defect description texts are input into the target vehicle defect generation model to obtain second vehicle defect labels corresponding to each second defect description text. The second vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels. Using the initial sample set of the first vehicle defect labels, the multiple second defect description texts, and the multiple second vehicle defect labels, a target sample set is constructed. The target sample set includes: multiple defect description texts, and vehicle defect labels corresponding to each defect description text. The vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels. According to the target sample set, the initial vehicle defect recognition model is trained to obtain a target vehicle defect recognition model. The target vehicle defect recognition model is used to obtain target vehicle defect labels corresponding to the vehicle defect description text to be recognized. The target vehicle defect labels include: target system sub-labels, target assembly sub-labels, target part sub-labels, and target defect sub-labels. In the above process, by using the initial sample set including the first defect description texts corresponding to multiple vehicle defects and the first vehicle defect labels corresponding to each first defect description text, the parameters of the initial vehicle defect label generation model are adjusted to obtain a target vehicle defect generation model, so that the target vehicle defect generation model can be used to obtain second vehicle defect labels corresponding to multiple second defect description texts, thereby improving the efficiency of obtaining vehicle defect labels. Since the vehicle defect recognition model is trained according to the target sample set including the initial sample set, the multiple second defect description texts, and the multiple second vehicle defect labels to obtain the target vehicle defect recognition model, the target vehicle defect recognition model is used to recognize the vehicle defect description text to be recognized, so as to obtain standardized target vehicle defect labels, avoiding the problem of non-standard defect label information in the prior art, and eliminating the need for manual entry of vehicle defect labels, thus improving the efficiency of entering vehicle defect labels.

[0061] Figure 1 is a schematic flowchart of a method for training a vehicle defect recognition model provided by an embodiment of the present disclosure. As Figure 1 shown, the method for training a vehicle defect recognition model provided by an embodiment of the present disclosure includes:

[0062] S11. Obtain an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect tags corresponding to each first defect description text, and obtain a data set including second defect description texts corresponding to multiple vehicle defects.

[0063] Among them, a vehicle defect may be an abnormality of a vehicle determined by detection, and this vehicle defect may be a defect in the vehicle production process. The multiple vehicle defects may be different vehicle defects. A defect description text refers to text information used to describe a vehicle defect. For example, "The gap between the rear door lamp and the tail lamp is uneven in the general assembly workshop". This defect description text may be obtained by recognizing defect description voice data input by a user, but is not limited thereto. The present disclosure does not specifically limit it, and those skilled in the art can specifically set it according to actual situations. For each vehicle defect, there may be a corresponding first defect description text and a second defect description text. The corresponding first defect description text and second defect description text can be understood as different descriptions of the same vehicle defect.

[0064] The first vehicle defect tag is obtained by manual marking. The first vehicle defect tag includes: a system sub-tag, an assembly sub-tag, a part sub-tag, and a defect sub-tag. The system sub-tag is used to represent the vehicle component system to which the current vehicle defect belongs, such as the interior system, the lighting system, etc. The part sub-tag is used to represent the vehicle component device to which the current vehicle defect belongs, such as the pillar sill interior trim panel device, the external lighting device. The assembly sub-tag is used to represent the vehicle assembly to which the current vehicle defect belongs, such as the left D-pillar upper trim panel assembly, the rear door lamp assembly. The defect sub-tag is used to represent the specific defect to which the current vehicle defect belongs, such as a large gap, an uneven gap, but is not limited thereto. The present disclosure does not specifically limit it, and those skilled in the art can specifically set it according to actual situations.

[0065] Specifically, for each of the multiple vehicle defects, obtain at least one first defect description text corresponding to the vehicle defect and the first vehicle defect tag corresponding to each first defect description text, and determine the initial sample set. The initial sample set includes multiple first defect description texts of multiple vehicle defects, and multiple first vehicle defect tags corresponding to the multiple first defect description texts. And, obtain a data set of second defect description texts corresponding to each of the multiple vehicle defects.

[0066] It should be noted that the first defect description texts included in the initial sample set and the second defect description texts included in the data set may be defect description texts respectively for the same multiple defects.

[0067] Optionally, based on the above embodiments, in some embodiments of the present disclosure, an implementation manner of obtaining an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect tags corresponding to each first defect description text, and obtaining a data set including second defect description texts corresponding to multiple vehicle defects may include:

[0068] Step a1, for each vehicle defect, obtain multiple defect description texts obtained by performing text recognition on the defect description voice data of each vehicle defect.

[0069] Among them, the defect description voice data includes standard defect description voice data and non-standard defect description voice data. The standard defect description voice data may refer to data recognized by experts with vehicle detection experience. The non-standard defect description voice data refers to voice data that is more colloquial when describing vehicle defects. Exemplarily, for the defect of the paint surface of the rear bumper of the same vehicle, the standard defect description voice data may be "the paint surface of the rear bumper assembly has peeled off", and the non-standard defect description voice data may be "there is a problem with the paint surface of the bumper peeling off", but it is not limited thereto. The present disclosure does not specifically limit it, and those skilled in the art can set it according to the actual situation.

[0070] Optionally, based on the above embodiments, in some embodiments of the present disclosure, an implementation manner of obtaining multiple defect description texts by performing text recognition on the defect description voice data of each vehicle defect for each vehicle defect may be:

[0071] Perform text recognition on the standard defect description voice data and the non-standard defect description voice data to obtain standard defect description texts and non-standard defect description texts.

[0072] Use the standard defect description texts and the non-standard defect description texts to construct the defect description texts.

[0073] Specifically, for multiple vehicle defects, obtain the standard defect description voice data and the non-standard defect description voice data of each vehicle defect. After obtaining the standard defect description voice data and the non-standard defect description voice data of each vehicle defect, perform text recognition on the standard defect description voice data and the non-standard defect description voice data of each vehicle defect respectively to obtain the standard defect description texts and the non-standard defect description texts corresponding to the standard defect description voice data and the non-standard defect description voice data respectively. Further, construct the defect description texts of each vehicle defect including the standard defect description texts and the non-standard defect description texts.

[0074] Optionally, based on the above embodiments, in some embodiments of the present disclosure, obtaining the standard defect description voice data and non-standard defect description voice data for each vehicle defect may be to rank the occurrences of multiple different vehicle defects in descending order, and according to the sorting result, for different vehicle defects, obtain different numbers of standard defect description voice data and non-standard defect description voice data. That is, it can be understood that for vehicle defects with a relatively high ranking, more standard defect description voice data and non-standard defect description voice data can be obtained.

[0075] Exemplarily, for vehicle defect 1, vehicle defect 2, vehicle defect 3, vehicle defect 4, vehicle defect 5, and vehicle defect 6, it is determined that the frequencies of vehicle defect 1, vehicle defect 2, vehicle defect 3, vehicle defect 4, vehicle defect 5, and vehicle defect 6 occurring in a preset duration of two months are 10 times, 7 times, 2 times, 4 times, 12 times, and 5 times respectively. Then, the sorting is as follows according to this frequency: vehicle defect 5, vehicle defect 1, vehicle defect 2, vehicle defect 6, vehicle defect 4, vehicle defect 3. Therefore, 30 standard defect description voice data and non-standard defect description voice data are obtained for vehicle defect 5 and vehicle defect 1 respectively, 20 standard defect description voice data and non-standard defect description voice data are obtained for vehicle defect 2 and vehicle defect 6 respectively, and 10 standard defect description voice data and non-standard defect description voice data are obtained for vehicle defect 4 and vehicle defect 3 respectively. However, this is not limited thereto, and the present disclosure does not specifically limit it. Those skilled in the art can set it specifically according to the actual situation.

[0076] In this way, the vehicle defect recognition model training method provided by the embodiments of the present disclosure can obtain the standard defect description text and non-standard defect description text for vehicle defects at the same time, thereby constructing the defect description text corresponding to the vehicle defect, enriching the description information of the vehicle defect, facilitating the subsequent training of the target vehicle defect recognition model, and improving the accuracy of obtaining the target vehicle defect label.

[0077] Step a2: Among the multiple defect description texts, determine multiple first defect description texts, and perform manual marking processing on the multiple first defect description texts to obtain the first vehicle defect label corresponding to each first defect description text.

[0078] Specifically, among the multiple defect description texts, randomly determine multiple first defect description texts, and perform manual marking processing on the multiple first defect description texts to obtain the first vehicle defect label corresponding to each first defect description text.

[0079] It should be noted that the number of multiple first defect description texts can be much smaller than the total number of multiple defect description texts. That is, it can be understood that a small number of defect description texts are randomly determined from multiple defect description texts as the first defect description texts.

[0080] Optionally, based on the above embodiments, in some embodiments of the present disclosure, in multiple defect description texts, a plurality of first defect description texts are randomly determined, and an implementation manner of manually marking the plurality of first defect description texts to obtain a first vehicle defect label corresponding to each first defect description text may be:

[0081] Perform word segmentation processing on the plurality of first defect description texts, and use the plurality of target word segments corresponding to each first defect description text to determine the vehicle component information to which the vehicle defect corresponding to the first defect description text belongs.

[0082] Wherein, the vehicle component information refers to the vehicle component to which the vehicle defect corresponding to the first defect description text belongs. Exemplarily, continuing with the above embodiments, for the first defect description text "the paint on the rear bumper assembly has peeled off", it belongs to the exterior system and the bumper, but is not limited thereto. The present disclosure does not specifically limit, and those skilled in the art can set it according to the actual situation.

[0083] The above word segmentation processing of the first defect description text can refer to the prior art and will not be elaborated here too much.

[0084] According to the vehicle component information, perform marking processing on the first defect description text to obtain the system sub-label, assembly sub-label, part sub-label, and defect sub-label included in the first vehicle defect label.

[0085] Specifically, perform word segmentation processing on the first defect description text corresponding to the vehicle defect to obtain a plurality of target word segments corresponding to the first defect description text. Use the plurality of target word segments corresponding to the first defect description text to determine the vehicle component information to which the vehicle defect corresponding to the first defect description text belongs. According to the vehicle component information, perform marking processing on the first defect description text to obtain the system sub-label, assembly sub-label, part sub-label, and defect sub-label included in the first vehicle defect label.

[0086] Step a3, construct an initial sample set according to the plurality of first defect description texts and the first vehicle defect label corresponding to each first defect description text.

[0087] Step a4, determine a plurality of second defect description texts in the plurality of defect description texts and construct a data set.

[0088] Wherein, the plurality of second defect description texts are other defect description texts in the plurality of defect description texts except the plurality of first defect description texts.

[0089] Specifically, perform manual marking on multiple first defect description texts to obtain the first vehicle defect labels corresponding to each first defect description text, combine the multiple first defect description texts and the first vehicle defect labels corresponding to each first defect description text to obtain an initial sample set including the multiple first defect description texts and the first vehicle defect labels corresponding to each first defect description text, and among the multiple defect description texts, determine the other defect description texts other than the multiple first defect description texts, determine the other defect description texts as multiple second defect description texts, and construct a data set.

[0090] S12. According to the initial sample set of the first vehicle defect labels, adjust the parameters of the initial vehicle defect label generation model to obtain a target vehicle defect generation model.

[0091] Among them, the vehicle defect label generation model refers to a model used to generate labels corresponding to defect description texts. This model can refer to existing models such as FastTEXT and TextCNN models, but is not limited to this. The present disclosure does not specifically limit it, and those skilled in the art can specifically set it according to the actual situation.

[0092] S13. Input the multiple second defect description texts into the target vehicle defect generation model to obtain the second vehicle defect labels corresponding to each second defect description text.

[0093] Among them, the second vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels.

[0094] Specifically, according to the initial sample set including the first defect description texts corresponding to different vehicle defects and the first vehicle defect labels corresponding to each of the first defect description texts, adjust the parameters of the initial vehicle defect label generation model to obtain a target vehicle defect generation model. Input the multiple second defect description texts into the target vehicle defect generation model, and use the target vehicle defect generation model to output the second vehicle defect labels corresponding to each second defect description text. Among them, the second vehicle defect labels include system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels.

[0095] S14. Use the initial sample set of the first vehicle defect labels, the multiple second defect description texts, and the multiple second vehicle defect labels to construct a target sample set.

[0096] Among them, the target sample set includes: multiple defect description texts, and the vehicle defect labels corresponding to each defect description text. The vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels.

[0097] Specifically, an initial sample set including first defect description texts corresponding to different vehicle defects and first vehicle defect labels corresponding to each of the first defect description texts, a plurality of second defect description texts, and a plurality of second vehicle defect labels are combined to construct a target sample set, where the target sample set includes: a plurality of defect description texts and vehicle defect labels corresponding to each defect description text, and the vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels.

[0098] S15. According to the target sample set, perform training processing on the initial vehicle defect recognition model to obtain a target vehicle defect recognition model.

[0099] Among them, the target vehicle defect recognition model is used to obtain target vehicle defect labels corresponding to the vehicle defect description text to be recognized, and the target vehicle defect labels include: target system sub-labels, target assembly sub-labels, target part sub-labels, and target defect sub-labels.

[0100] Specifically, input the target sample set into the initial vehicle defect recognition model, perform training processing on the initial vehicle defect recognition model, extract feature parameters corresponding to a plurality of defect description texts, and use the vehicle defect labels corresponding to each defect description text sample as supervised data for supervised training, continuously adjust the parameter information in the initial vehicle defect recognition model until the training function corresponding to the initial vehicle defect recognition model converges, thereby obtaining a trained target vehicle defect recognition model.

[0101] Thus, the vehicle defect recognition model training method provided by the embodiments of the present disclosure obtains an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, and obtains a data set including second defect description texts corresponding to multiple vehicle defects. Among them, the first vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label; according to the initial sample set of the first vehicle defect label, the parameters of the initial vehicle defect label generation model are adjusted to obtain a target vehicle defect generation model; the multiple second defect description texts are input into the target vehicle defect generation model to obtain second vehicle defect labels corresponding to each second defect description text. The second vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label; using the initial sample set of the first vehicle defect label, the multiple second defect description texts, and the multiple second vehicle defect labels, a target sample set is constructed. The target sample set includes: multiple defect description texts, and vehicle defect labels corresponding to each defect description text. The vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label; according to the target sample set, the initial vehicle defect recognition model is trained to obtain a target vehicle defect recognition model. The target vehicle defect recognition model is used to obtain a target vehicle defect label corresponding to the vehicle defect description text to be recognized. The target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label. In the above process, by using an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, the parameters of the initial vehicle defect label generation model are adjusted to obtain a target vehicle defect generation model, so as to be able to use the target vehicle defect generation model to obtain second vehicle defect labels corresponding to multiple second defect description texts, thereby improving the efficiency of obtaining vehicle defect labels. Since the vehicle defect recognition model is trained according to the target sample set including the initial sample set, the multiple second defect description texts, and the multiple second vehicle defect labels to obtain the target vehicle defect recognition model, therefore, the target vehicle defect recognition model is used to recognize the vehicle defect description text to be recognized, so as to obtain a standardized target vehicle defect label, avoiding the problem of non-standard defect label information in the prior art, and eliminating the need for manual entry of vehicle defect labels, improving the efficiency of entering vehicle defect labels.

[0102] Figure 2 is a schematic flowchart of a vehicle defect entry method provided by the embodiments of the present disclosure. As Figure 2 shown, a vehicle defect entry method provided by the embodiments of the present disclosure includes:

[0103] S21. Obtain the vehicle defect description text to be recognized input by the user.

[0104] Among them, the vehicle defect description text to be recognized is used to describe the target defect of the vehicle. The user can be a vehicle quality inspector.

[0105] S22. Input the vehicle defect description text into the target vehicle defect recognition model, and obtain the target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model.

[0106] Among them, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label. The target vehicle defect recognition model is trained according to a target sample set. The target sample set includes: a plurality of defect description texts, and the vehicle defect label corresponding to each defect description text. The vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

[0107] Specifically, obtain the vehicle defect description text to be recognized input by the user for describing the target defect of the vehicle, input the vehicle defect description text to be recognized into the trained target vehicle defect recognition model, and obtain the target vehicle defect label corresponding to the vehicle defect description text to be recognized. The target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label.

[0108] Exemplarily, for the vehicle defect description text to be recognized such as "The gap between the rear door lamp and the tail lamp is uneven in the general assembly workshop", the "The gap between the rear door lamp and the tail lamp is uneven in the general assembly workshop" is recognized through the target vehicle defect recognition model. The obtained target system sub-label corresponding to "The gap between the rear door lamp and the tail lamp is uneven in the general assembly workshop" is: Lighting system; the target assembly sub-label is: External lighting device; the target part sub-label is: Rear door lamp assembly; and the target defect sub-label is: Uneven gap. However, it is not limited thereto. The present disclosure does not specifically limit, and those skilled in the art can specifically set according to the actual situation.

[0109] In this way, the vehicle defect entry method provided by the embodiments of the present disclosure obtains the text description of the vehicle defect to be recognized input by the user, and the text description of the vehicle defect to be recognized is used to describe the target defect of the vehicle; inputs the text description of the vehicle defect into the target vehicle defect recognition model, and obtains the target vehicle defect label corresponding to the text description of the vehicle defect output by the target vehicle defect recognition model; wherein, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, and the target vehicle defect recognition model is trained according to a target sample set, and the target sample set includes: a plurality of defect description texts, and the vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label. In the above process, the target vehicle defect recognition model trained through the target sample set is used to recognize the text description of the vehicle defect to be recognized, and a standardized target vehicle defect label is obtained, avoiding the problem of non-standard defect label information in the prior art, and there is no need to manually enter the vehicle defect label, improving the efficiency of entering the vehicle defect label.

[0110] Optionally, based on the above embodiments, in some embodiments of the present disclosure, the method further includes:

[0111] Performing a matching process on the target defect sub-label among multiple vehicle defect levels to obtain the target vehicle defect level to which the target defect belongs.

[0112] Wherein, the vehicle defect level is used to indicate the severity of the target defect of the vehicle. It should be noted that corresponding vehicle defect levels are preset for different defect sub-labels.

[0113] Specifically, after obtaining the target vehicle defect label corresponding to the text description of the vehicle defect to be recognized, according to the target defect sub-label included in the target vehicle defect label, performing a matching process among multiple vehicle defect levels to obtain the target vehicle defect level corresponding to the target defect sub-label, so as to obtain the target vehicle defect level to which the target defect belongs.

[0114] In this way, the vehicle defect entry method provided by the embodiments of the present disclosure can determine the target vehicle defect level to which the target defect belongs according to the target defect sub-label, facilitating the user to obtain the severity of the target defect and being able to repair the target defect in a timely manner.

[0115] Optionally, based on the above embodiments, in some embodiments of the present disclosure, the method further includes:

[0116] Step b1, saving the text description of the vehicle defect to be recognized, the target vehicle defect label corresponding to the text description of the vehicle defect, and the target vehicle defect level to which the target defect belongs to a preset database.

[0117] Among them, the preset database can be the local storage database of the vehicle terminal, or the database corresponding to the server, but is not limited thereto. The present disclosure does not specifically limit it, and those skilled in the art can set it specifically according to the actual situation.

[0118] Step b2, periodically update the target sample set by using the preset database, and train the target vehicle defect recognition model by using the updated target sample set.

[0119] Specifically, save the vehicle defect description text to be recognized corresponding to the target defect of the vehicle, the target vehicle defect label corresponding to the vehicle defect description text, and the target vehicle defect level to which the target defect belongs to the preset database, periodically update the target sample set by using the preset database, and update the target vehicle defect recognition model by using the updated target sample set.

[0120] In this way, the vehicle defect entry method provided by the embodiments of the present disclosure can periodically update the target sample set and the target vehicle defect recognition model, thereby improving the robustness of the target vehicle defect recognition model and thus improving the accuracy of obtaining the target vehicle defect label.

[0121] Figure 3 It is a schematic structural diagram of a vehicle defect recognition model training device provided by the embodiments of the present disclosure, as Figure 4 shown, the vehicle defect recognition model training device includes: an acquisition module 11, a target vehicle defect generation model obtaining module 12, a second vehicle defect label obtaining module 13, a target sample set construction module 14, and a target vehicle defect recognition model training module 15.

[0122] Among them, the acquisition module 11 is configured to acquire an initial sample set including first defect description texts corresponding to a plurality of vehicle defects and first vehicle defect labels corresponding to each of the first defect description texts, and acquire a data set including second defect description texts corresponding to the plurality of vehicle defects, where the first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels;

[0123] The target vehicle defect generation model obtaining module 12 is configured to adjust the parameters of the initial vehicle defect label generation model according to the initial sample set of the first vehicle defect labels to obtain a target vehicle defect generation model;

[0124] The second vehicle defect label obtaining module 13 is configured to input a plurality of the second defect description texts into the target vehicle defect generation model to obtain second vehicle defect labels corresponding to each of the second defect description texts, where the second vehicle defect labels include: the system sub-labels, the assembly sub-labels, the part sub-labels, and the defect sub-labels;

[0125] The target sample set construction module 14 is configured to construct a target sample set by using the initial sample set of the first vehicle defect labels, the multiple second defect description texts, and the multiple second vehicle defect labels. The target sample set includes: multiple defect description texts, and vehicle defect labels corresponding to each of the defect description texts. The vehicle defect labels include: the system sub-label, the assembly sub-label, the part sub-label, and the defect sub-label;

[0126] The target vehicle defect recognition model training module 15 is configured to train an initial vehicle defect recognition model according to the target sample set to obtain a target vehicle defect recognition model. The target vehicle defect recognition model is used to obtain a target vehicle defect label corresponding to a vehicle defect description text to be recognized. The target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label.

[0127] The vehicle defect identification model training device provided by an embodiment of the present disclosure includes an acquisition module 11, configured to acquire an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text, and acquire a data set including second defect description texts corresponding to multiple vehicle defects, wherein the first vehicle defect labels include: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; a target vehicle defect generation model obtaining module 12, configured to adjust the parameters of the initial vehicle defect label generation model according to the initial sample set of the first vehicle defect labels to obtain a target vehicle defect generation model; a second vehicle defect label acquisition module 13, configured to input multiple second defect description texts into the target vehicle defect generation model to acquire second vehicle defect labels corresponding to each second defect description text, the second vehicle defect labels including: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; a target sample set construction module 14, configured to construct a target sample set by using the initial sample set of the first vehicle defect labels, multiple second defect description texts, and multiple second vehicle defect labels, the target sample set including: multiple defect description texts, and vehicle defect labels corresponding to each defect description text, the vehicle defect labels including: system sub-labels, assembly sub-labels, part sub-labels, and defect sub-labels; a target vehicle defect identification model training module 15, configured to perform training processing on the initial vehicle defect identification model according to the target sample set to acquire a target vehicle defect identification model, the target vehicle defect identification model being used to acquire target vehicle defect labels corresponding to a vehicle defect description text to be identified, the target vehicle defect labels including: target system sub-labels, target assembly sub-labels, target part sub-labels, and target defect sub-labels. In the above device, by using an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect labels corresponding to each first defect description text to adjust the parameters of the initial vehicle defect label generation model to obtain a target vehicle defect generation model, the second vehicle defect labels corresponding to multiple second defect description texts can be acquired by using the target vehicle defect generation model, thereby improving the efficiency of acquiring vehicle defect labels. Since the vehicle defect identification model is trained according to the target sample set including the initial sample set, multiple second defect description texts, and multiple second vehicle defect labels to obtain a target vehicle defect identification model, the target vehicle defect identification model is used to identify the vehicle defect description text to be identified, so as to acquire standardized target vehicle defect labels, avoiding the problem of non-standard defect label information in the prior art, and eliminating the need for manual entry of vehicle defect labels, thus improving the efficiency of entering vehicle defect labels.

[0128] Figure 4 is a structural schematic diagram of a vehicle defect entry device provided by an embodiment of the present disclosure, as Figure 4As shown in the figure, the vehicle defect entry device includes: a to-be-recognized vehicle defect description text acquisition module 21 and a target vehicle defect label acquisition module 22.

[0129] Among them, the to-be-recognized vehicle defect description text acquisition module 21 is used to acquire the to-be-recognized vehicle defect description text input by the user, and the to-be-recognized vehicle defect description text is used to describe the target defect of the vehicle.

[0130] The target vehicle defect label acquisition module 22 inputs the vehicle defect description text into the target vehicle defect recognition model, and acquires the target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model; among them, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, and the target vehicle defect recognition model is trained according to a target sample set, and the target sample set includes: a plurality of defect description texts, and the vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

[0131] In the vehicle defect entry device provided by the embodiment of the present disclosure, the to-be-recognized vehicle defect description text acquisition module 21 is used to acquire the to-be-recognized vehicle defect description text input by the user, and the to-be-recognized vehicle defect description text is used to describe the target defect of the vehicle; the target vehicle defect label acquisition module 22 is used to input the vehicle defect description text into the target vehicle defect recognition model, and acquire the target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model; among them, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, and the target vehicle defect recognition model is trained according to a target sample set, and the target sample set includes: a plurality of defect description texts, and the vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label. In the above process, the to-be-recognized vehicle defect description text is recognized by using the target vehicle defect recognition model trained through the target sample set, and the standardized target vehicle defect label is acquired, avoiding the problem of non-standard defect label information in the prior art, and there is no need to manually enter the vehicle defect label, improving the efficiency of entering the vehicle defect label.

[0132] The device provided by the embodiment of the present invention can execute the method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0133] It should be noted that in the embodiments of the above device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0134] Figure 5 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. As Figure 5 shown, the electronic device includes a processor 510, a memory 520, an input device 530, and an output device 540; the number of processors 510 in the computer device can be one or more, Figure 5 and one processor 510 is taken as an example herein; the processor 510, the memory 520, the input device 530, and the output device 540 in the electronic device can be connected through a bus or other means, Figure 5 and taking the connection through a bus as an example herein.

[0135] The memory 520, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of the present invention. The processor 510 executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory 520, that is, implements the methods provided by the embodiments of the present invention.

[0136] The memory 520 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some instances, the memory 520 may further include a memory remotely set relative to the processor 510, and these remote memories may be connected to the computer device through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0137] The input device 530 can be used to receive input digital or character information, and generate key signal inputs related to the user settings and function controls of the electronic device, and may include a keyboard, a mouse, etc. The output device 540 may include a display device such as a display screen.

[0138] The embodiments of the present disclosure also provide a storage medium containing computer-executable instructions, and the computer-executable instructions are used to implement the methods provided by the embodiments of the present invention when executed by a computer processor.

[0139] Of course, for the storage medium containing computer-executable instructions provided by the embodiments of the present invention, the computer-executable instructions are not limited to the method operations described above, and can also execute the related operations in the methods provided by any embodiments of the present invention.

[0140] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and the necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FLASH), a hard disk, or an optical disc of a computer, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0141] It should be noted that in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0142] The above are only specific embodiments of the present disclosure, enabling those skilled in the art to understand or implement the present disclosure. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for training a vehicle defect recognition model, characterized in that, Including: Obtain an initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect tags corresponding to each of the first defect description texts, and obtain a data set including second defect description texts corresponding to the multiple vehicle defects, wherein the first vehicle defect tags include: system sub-tags, assembly sub-tags, part sub-tags, and defect sub-tags; According to the initial sample set of the first vehicle defect tags, adjust the parameters of the initial vehicle defect tag generation model to obtain a target vehicle defect generation model; Input multiple second defect description texts into the target vehicle defect generation model, and obtain second vehicle defect tags corresponding to each of the second defect description texts, where the second vehicle defect tags include: the system sub-tags, the assembly sub-tags, the part sub-tags, and the defect sub-tags; Utilize the initial sample set of the first vehicle defect tags, multiple second defect description texts, and multiple second vehicle defect tags to construct a target sample set, where the target sample set includes: multiple defect description texts, and vehicle defect tags corresponding to each of the defect description texts, and the vehicle defect tags include: the system sub-tags, the assembly sub-tags, the part sub-tags, and the defect sub-tags; According to the target sample set, perform training processing on the initial vehicle defect recognition model to obtain a target vehicle defect recognition model, where the target vehicle defect recognition model is used to obtain target vehicle defect tags corresponding to a vehicle defect description text to be recognized, and the target vehicle defect tags include: target system sub-tags, target assembly sub-tags, target part sub-tags, and target defect sub-tags.

2. The method according to claim 1, wherein The obtaining of the initial sample set including first defect description texts corresponding to multiple vehicle defects and first vehicle defect tags corresponding to each of the first defect description texts, and the obtaining of the data set including second defect description texts corresponding to the multiple vehicle defects, includes: For each of the vehicle defects, obtain multiple defect description texts obtained by performing text recognition on defect description voice data of each of the vehicle defects; Among the multiple defect description texts, determine multiple first defect description texts, and perform manual marking processing on the multiple first defect description texts to obtain first vehicle defect tags corresponding to each of the first defect description texts; Construct the initial sample set according to the multiple first defect description texts and the first vehicle defect tags corresponding to each of the first defect description texts; Among the multiple defect description texts, determine multiple second defect description texts, and construct the data set, where the multiple second defect description texts are other defect description texts in the multiple defect description texts except the multiple first defect description texts.

3. The method according to claim 2, characterized in that, The defect description voice data includes: standard defect description voice data and non-standard defect description voice data; The obtaining, for each of the vehicle defects, of multiple defect description texts obtained by performing text recognition on defect description voice data of each of the vehicle defects, includes: Perform text recognition on the standard defect description voice data and the non-standard defect description voice data to obtain a standard defect description text and a non-standard defect description text; Utilize the standard defect description text and the non-standard defect description text to construct the defect description text.

4. The method according to claim 2, wherein The artificial marking process for multiple first defect description texts to obtain a first vehicle defect label corresponding to each first defect description text includes: Perform word segmentation on multiple first defect description texts, and use multiple target word segments corresponding to each first defect description text to determine the vehicle component information to which the vehicle defect corresponding to the first defect description text belongs; According to the vehicle component information, perform a marking process on the first defect description text to obtain the system sub-label, the assembly sub-label, the part sub-label, and the defect sub-label included in the first vehicle defect label.

5. A method for entering vehicle defects, characterized in that, Including: Obtain a vehicle defect description text to be recognized input by a user, where the vehicle defect description text to be recognized is used to describe a target defect of a vehicle; Input the vehicle defect description text into a target vehicle defect recognition model, and obtain a target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model; Wherein, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, the target vehicle defect recognition model is trained according to a target sample set, the target sample set includes: multiple defect description texts, and a vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

6. The method according to claim 5, characterized in that, The method further includes: Perform a matching process on the target defect sub-label among multiple vehicle defect levels to obtain a target vehicle defect level to which the target defect belongs.

7. The method according to claim 6, wherein The method further includes: Save the vehicle defect description text to be recognized, the target vehicle defect label corresponding to the vehicle defect description text, and the target vehicle defect level to which the target defect belongs to a preset database; Periodically utilize the preset database to update the target sample set, and train the target vehicle defect recognition model by using the updated target sample set.

8. A vehicle defect recognition model training device, characterized in that Including: An acquisition module, configured to acquire an initial sample set including multiple first defect description texts corresponding to vehicle defects and a first vehicle defect label corresponding to each first defect description text, and acquire a data set including multiple second defect description texts corresponding to the vehicle defects, wherein the first vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label; A target vehicle defect generation model obtaining module, configured to adjust parameters of an initial vehicle defect label generation model according to the initial sample set of the first vehicle defect label to obtain a target vehicle defect generation model; A second vehicle defect label acquisition module, configured to input multiple pieces of the second defect description text into the target vehicle defect generation model, and acquire a second vehicle defect label corresponding to each piece of the second defect description text, where the second vehicle defect label includes: the system sub-label, the assembly sub-label, the part sub-label, and the defect sub-label; A target sample set construction module, configured to construct a target sample set by using the initial sample set of the first vehicle defect label, multiple pieces of the second defect description text, and multiple pieces of the second vehicle defect label, where the target sample set includes: multiple defect description texts, and a vehicle defect label corresponding to each defect description text; A target vehicle defect recognition model training module, configured to perform training processing on an initial vehicle defect recognition model according to the target sample set, and acquire a target vehicle defect recognition model, where the target vehicle defect recognition model is configured to acquire a target vehicle defect label corresponding to a vehicle defect description text to be recognized, and the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label.

9. A vehicle defect entry device, characterized in that, including: A vehicle defect description text to be recognized acquisition module, configured to acquire a vehicle defect description text to be recognized input by a user, where the vehicle defect description text to be recognized is used to describe a target defect of a vehicle; A target vehicle defect label acquisition module, configured to input the vehicle defect description text into the target vehicle defect recognition model, and acquire a target vehicle defect label corresponding to the vehicle defect description text output by the target vehicle defect recognition model; wherein, the target vehicle defect label includes: a target system sub-label, a target assembly sub-label, a target part sub-label, and a target defect sub-label, the target vehicle defect recognition model is trained according to a target sample set, the target sample set includes: multiple defect description texts, and a vehicle defect label corresponding to each defect description text, and the vehicle defect label includes: a system sub-label, an assembly sub-label, a part sub-label, and a defect sub-label.

10. An electronic device, characterized in that, including: One or more processors; A storage device, configured to store one or more programs, when the one or more programs are executed by the one or more processors, enabling the one or more processors to implement the vehicle defect recognition model training method according to any one of claims 1 to 4, or the vehicle defect entry method according to any one of claims 5 to 7.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the vehicle defect recognition model training method according to any one of claims 1 to 4, or the vehicle defect entry method according to any one of claims 5 to 7.

12. A vehicle, characterized in that, including: A processor, a memory, and a computer program stored on the memory and executable on the processor, where when the computer program is executed by the processor, it implements the vehicle defect recognition model training method according to any one of claims 1 to 4, or the vehicle defect entry method according to any one of claims 5 to 7.

Images