Vehicle loss assessment method and device based on artificial intelligence, equipment and medium

Abstract

The invention belongs to the field of artificial intelligence and financial science and technology, and relates to a vehicle loss assessment method and device based on artificial intelligence, computer equipment and a storage medium, and the method comprises the steps: receiving image data and text data which are input by a user and are related to a vehicle accident of a target vehicle; preprocessing the image data to obtain target image data; integrating the text data to obtain target text data; performing feature extraction on the target image data based on an image processing module in the survey loss assessment model to obtain image features; performing feature extraction on the target text data based on a text processing module in the survey and loss assessment model to obtain text features; fusing the image features and the text features to obtain target fusion features; and performing loss assessment processing on the target fusion feature based on a target full connection layer in the survey loss assessment model to obtain a corresponding vehicle loss assessment result. The processing efficiency of vehicle loss assessment is improved, and the accuracy of the vehicle loss assessment result is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence development technology and financial technology, and in particular to an artificial intelligence-based vehicle loss assessment method, device, computer equipment and storage medium. Background Art

[0002] In the insurance industry, the investigation and damage assessment process for vehicle collision accidents has long been a key link in ensuring the accuracy and efficiency of insurance claims. The traditional process usually starts with the insurance company receiving a report from the insured, and then responding quickly by sending surveyors with professional skills to the accident scene for on-site inspection.

[0003] However, with the continuous increase in the number of vehicles and the frequent occurrence of traffic accidents, the challenges faced by the traditional inspection and damage assessment process are becoming increasingly prominent. First of all, there are many nodes involved in the process, from on-site inspection to vehicle towing to the repair shop, to professional evaluation at the damage assessment center, and finally through back-end approval. Each link requires the participation of experienced professionals, which not only increases labor costs, but also prolongs the overall processing time.

[0004] Secondly, the cost of interpersonal communication in the process is high, and the information transmission between different nodes relies on traditional methods such as telephone and email, which is not only inefficient, but also prone to delays or errors due to poor information transmission or misunderstanding. Especially in the damage assessment stage, the judgment of professional damage assessors is often affected by multiple factors such as personal experience, skill level and on-site conditions, which leads to low accuracy of vehicle loss assessment results. Summary of the invention

[0005] The purpose of the embodiments of the present application is to propose a vehicle loss assessment method, device, computer equipment and storage medium based on artificial intelligence to solve the technical problems that the existing inspection and damage assessment process has low processing efficiency and low accuracy of vehicle loss assessment results.

[0006] In order to solve the above technical problems, the embodiment of the present application provides a vehicle loss assessment method based on artificial intelligence, which adopts the following technical solutions:

[0007] receiving image data and text data related to a vehicle accident of a target vehicle input by a user;

[0008] Performing image preprocessing on the image data to obtain corresponding target image data;

[0009] Performing data integration processing on the text data to obtain corresponding target text data;

[0010] Extracting features of the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features;

[0011] Perform feature extraction on the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features;

[0012] Fusing the image feature with the text feature to obtain a corresponding target fusion feature;

[0013] Based on the target fully connected layer in the inspection and damage assessment model, the target fusion features are processed for loss assessment to obtain a vehicle loss assessment result corresponding to the target vehicle.

[0014] Furthermore, the step of extracting features from the target image data based on the image processing module in the pre-built damage assessment model to obtain corresponding image features specifically includes:

[0015] Inputting the target image data into the image processing module in the survey and damage assessment model;

[0016] Extracting local features of the target image data through a convolutional layer in the image processing module to obtain a corresponding first feature map;

[0017] Performing dimensionality reduction processing on the first feature map based on the pooling layer in the image processing module to obtain a corresponding second feature map;

[0018] Based on the fully connected layer in the image processing module, feature conversion processing is performed on the second feature map to obtain image features corresponding to the target image data.

[0019] Furthermore, the step of extracting features from the target text data based on the text processing module in the damage assessment model to obtain corresponding text features specifically includes:

[0020] Inputting the target text data into the text processing module in the inspection and damage assessment model;

[0021] Performing text preprocessing on the target text data by the text processing module to obtain corresponding designated text data;

[0022] Perform text encoding processing on the specified text data to obtain corresponding word vector data;

[0023] The word vector data is processed based on a preset sequence processing strategy to obtain text features corresponding to the target text data.

[0024] Furthermore, the step of fusing the image feature with the text feature to obtain the corresponding target fusion feature specifically includes:

[0025] Get the preset fusion strategy;

[0026] Based on the fusion strategy, the image feature and the text feature are fused to obtain a fused specified feature;

[0027] The designated feature is used as the target fusion feature.

[0028] Furthermore, the step of performing image preprocessing on the image data to obtain corresponding target image data specifically includes:

[0029] Performing image enhancement processing on the image data to obtain corresponding first image data;

[0030] Performing denoising processing on the first image data to obtain corresponding second image data;

[0031] performing size normalization processing on the second image data to obtain corresponding third image data;

[0032] The third image data is used as the target image data.

[0033] Furthermore, before the step of extracting features from the target image data by the image processing module in the pre-built survey and damage assessment model to obtain corresponding image features, the step further includes:

[0034] Obtain pre-collected vehicle inspection and damage assessment record data;

[0035] Classify and process the vehicle inspection and damage assessment record data based on a preset data dimension to obtain corresponding initial sample data;

[0036] Performing data annotation processing on the initial sample data to obtain corresponding sample data;

[0037] Determine a model structure of a preset multimodal deep learning model, and call the multimodal deep learning model;

[0038] Obtaining a target loss function corresponding to the multimodal deep learning model;

[0039] Based on the target loss function, the multimodal deep learning model is cross-validated and trained using the sample data to obtain a specified model that meets the preset model construction expectations;

[0040] The designated model is used as the inspection and damage assessment model, and the inspection and damage assessment model is deployed on the cloud.

[0041] Furthermore, after the step of performing loss assessment processing on the target fusion feature based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle, the method further includes:

[0042] Get the preset report generation strategy;

[0043] Integrate the vehicle loss assessment results based on the report generation strategy to obtain a corresponding vehicle loss report;

[0044] Get the target push method;

[0045] Based on the target push method, the vehicle loss report is pushed to the user.

[0046] In order to solve the above technical problems, the embodiment of the present application further provides a vehicle damage assessment device based on artificial intelligence, which adopts the following technical solution:

[0047] A receiving module, used for receiving image data and text data related to a vehicle accident of a target vehicle input by a user;

[0048] A first processing module, used for performing image preprocessing on the image data to obtain corresponding target image data;

[0049] A second processing module, used for performing data integration processing on the text data to obtain corresponding target text data;

[0050] A first extraction module is used to extract features from the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features;

[0051] A second extraction module is used to extract features from the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features;

[0052] A fusion module, used for fusing the image features with the text features to obtain corresponding target fusion features;

[0053] An assessment module is used to perform loss assessment processing on the target fusion features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle.

[0054] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0055] receiving image data and text data related to a vehicle accident of a target vehicle input by a user;

[0056] Performing image preprocessing on the image data to obtain corresponding target image data;

[0057] Performing data integration processing on the text data to obtain corresponding target text data;

[0058] Extracting features of the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features;

[0059] Perform feature extraction on the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features;

[0060] Fusing the image feature with the text feature to obtain a corresponding target fusion feature;

[0061] Based on the target fully connected layer in the inspection and damage assessment model, the target fusion features are processed for loss assessment to obtain a vehicle loss assessment result corresponding to the target vehicle.

[0062] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0063] receiving image data and text data related to a vehicle accident of a target vehicle input by a user;

[0064] Performing image preprocessing on the image data to obtain corresponding target image data;

[0065] Performing data integration processing on the text data to obtain corresponding target text data;

[0066] Extracting features of the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features;

[0067] Perform feature extraction on the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features;

[0068] Fusing the image feature with the text feature to obtain a corresponding target fusion feature;

[0069] Based on the target fully connected layer in the inspection and damage assessment model, the target fusion features are processed for loss assessment to obtain a vehicle loss assessment result corresponding to the target vehicle.

[0070] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0071] The present application first receives image data and text data related to a vehicle accident of a target vehicle input by a user; then performs image preprocessing on the image data to obtain corresponding target image data; and performs data integration processing on the text data to obtain corresponding target text data; then, based on the image processing module in the pre-built inspection and damage assessment model, feature extraction is performed on the target image data to obtain corresponding image features; subsequently, based on the text processing module in the inspection and damage assessment model, feature extraction is performed on the target text data to obtain corresponding text features; further, the image features and the text features are fused to obtain corresponding target fused features; finally, based on the target fully connected layer in the inspection and damage assessment model, loss assessment processing is performed on the target fused features to obtain a vehicle loss assessment result corresponding to the target vehicle. After acquiring image data and text data corresponding to a vehicle accident of a target vehicle, the present application performs feature extraction on the image data based on an image processing module in an inspection and damage assessment model to obtain image features, and performs feature extraction on the text data based on the image processing module in the inspection and damage assessment model to obtain text features, and then fuses the image features with the text features to obtain target fused features, and finally performs automatic and accurate loss assessment processing on the target fused features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle, thereby greatly reducing the consumption of personnel in the vehicle inspection and damage assessment process, and greatly reducing the influence of human factors on the inspection and damage assessment results, effectively improving the processing efficiency of vehicle loss assessment, and improving the accuracy of the generated vehicle loss assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0073] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0074] Figure 2 A flowchart of an embodiment of a vehicle damage assessment method based on artificial intelligence according to the present application;

[0075] Figure 3 is a structural schematic diagram of an embodiment of an artificial intelligence-based vehicle loss assessment device according to the present application;

[0076] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0077] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0078] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0079] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0080] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0081] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0082] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0083] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0084] It should be noted that the artificial intelligence-based vehicle loss assessment method provided in the embodiments of the present application is generally executed by a server / terminal device, and accordingly, the artificial intelligence-based vehicle loss assessment device is generally arranged in a server / terminal device.

[0085] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0086] Continue to refer Figure 2 , showing a flow chart of an embodiment of an artificial intelligence-based vehicle loss assessment method according to the present application. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted. The artificial intelligence-based vehicle loss assessment method provided in the embodiment of the present application can be applied to any scenario where vehicle loss assessment is required, and the artificial intelligence-based vehicle loss assessment method can be applied to products in these scenarios, for example, vehicle loss assessment in the field of vehicle insurance. The artificial intelligence-based vehicle loss assessment method comprises the following steps:

[0087] Step S201 , receiving image data and text data related to a vehicle accident of a target vehicle input by a user.

[0088] In this embodiment, the vehicle damage assessment method based on artificial intelligence is run on an electronic device (e.g. Figure 1The server / terminal device shown in the figure) can obtain image data and text data through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future. The executor of the present application may specifically be a vehicle loss assessment system, or simply referred to as a system. Specifically, a user can upload image data and text data related to a vehicle accident of a target vehicle through the system interface of the vehicle loss assessment system. Among them, the above-mentioned image data may be pictures of damaged vehicles, which may be taken at the scene of the accident or close-ups of damaged parts of the vehicle. The above-mentioned text data is some basic information related to the target vehicle and vehicle accidents entered by the user, such as vehicle type, license plate number, accident type, accident description, etc. This information helps the system to more accurately assess the loss.

[0089] Step S202: performing image preprocessing on the image data to obtain corresponding target image data.

[0090] In this embodiment, the specific implementation process of performing image preprocessing on the image data to obtain the corresponding target image data will be described in further detail in subsequent specific embodiments of the present application and will not be elaborated on here.

[0091] Step S203, performing data integration processing on the text data to obtain corresponding target text data.

[0092] In this embodiment, the corresponding target text data may be obtained by integrating the vehicle type, license plate number, accident type, accident description and other data contained in the above text data.

[0093] Step S204, extracting features of the target image data based on the image processing module in the pre-constructed inspection and damage assessment model to obtain corresponding image features.

[0094] In this embodiment, the above-mentioned image processing module based on the pre-built inspection and damage assessment model performs feature extraction on the target image data to obtain the corresponding image features. The specific implementation process will be further described in detail in subsequent specific embodiments of the present application and will not be elaborated in detail here.

[0095] Step S205, extracting features from the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features.

[0096] In this embodiment, the above-mentioned text processing module based on the inspection and damage assessment model performs feature extraction on the target text data to obtain the corresponding text features. This application will provide further detailed descriptions of this in subsequent specific embodiments and will not be elaborated on here.

[0097] Step S206: fusing the image feature with the text feature to obtain a corresponding target fusion feature.

[0098] In this embodiment, the above-mentioned specific implementation process of fusing the image features with the text features to obtain the corresponding target fusion features will be further described in detail in subsequent specific embodiments of the present application and will not be elaborated on here.

[0099] Step S207, performing loss assessment processing on the target fusion features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle.

[0100] In this embodiment, the target fusion features can be input into the inspection and damage assessment model, so that the target fusion features can be processed for loss assessment through the target fully connected layer in the inspection and damage assessment model, and the vehicle loss assessment result corresponding to the target vehicle can be output. Among them, the inspection and damage assessment model can assess the damage information of the target vehicle, that is, the above-mentioned vehicle loss assessment result, through complex calculations and analysis. The vehicle loss assessment result may include the damaged part, the degree of damage, and the repair quotation. In addition, the construction process of the above-mentioned inspection and damage assessment model will be further described in detail in the subsequent specific embodiments of this application, and will not be elaborated on here.

[0101] The present application first receives image data and text data related to a vehicle accident of a target vehicle input by a user; then performs image preprocessing on the image data to obtain corresponding target image data; and performs data integration processing on the text data to obtain corresponding target text data; then, based on the image processing module in the pre-constructed inspection and damage assessment model, feature extraction is performed on the target image data to obtain corresponding image features; subsequently, based on the text processing module in the inspection and damage assessment model, feature extraction is performed on the target text data to obtain corresponding text features; further, the image features and the text features are fused to obtain corresponding target fused features; finally, based on the target fully connected layer in the inspection and damage assessment model, loss assessment processing is performed on the target fused features to obtain a vehicle loss assessment result corresponding to the target vehicle. After acquiring image data and text data corresponding to a vehicle accident of a target vehicle, the present application performs feature extraction on the image data based on an image processing module in an inspection and damage assessment model to obtain image features, and performs feature extraction on the text data based on the image processing module in the inspection and damage assessment model to obtain text features, and then fuses the image features with the text features to obtain target fused features, and finally performs automatic and accurate loss assessment processing on the target fused features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle, thereby greatly reducing the consumption of personnel in the vehicle inspection and damage assessment process, and greatly reducing the influence of human factors on the inspection and damage assessment results, effectively improving the processing efficiency of vehicle loss assessment, and improving the accuracy of the generated vehicle loss assessment results.

[0102] In some optional implementations, step S204 includes the following steps:

[0103] The target image data is input into the image processing module in the inspection and damage assessment model.

[0104] In this embodiment, the above-mentioned inspection and damage assessment model is a multimodal deep learning model based on a framework suitable for deep learning development, such as TensorFlow, PyTorch, etc. The inspection and damage assessment model includes at least an image processing module, a text processing module, a target fully connected layer, and may also include a fusion layer. Among them, the above-mentioned image processing module can specifically adopt a CNN model (convolutional neural network), and the CNN model includes a convolution layer, a pooling layer, and a fully connected layer. Among them, the number of convolution layers, the size and number of convolution kernels, and the activation function (such as ReLU) can be determined according to the actual CNN model construction requirements. The above-mentioned pooling layer is used to use maximum pooling or average pooling to reduce the dimension of the feature map and reduce the amount of calculation. The above-mentioned fully connected layer is used to convert the feature map output by the convolution layer into a feature vector of fixed length.

[0105] The local features of the target image data are extracted through the convolution layer in the image processing module to obtain a corresponding first feature map.

[0106] In this embodiment, the target image data is passed through a series of convolutional layers, each of which contains multiple convolutional kernels (filters) for extracting local features (such as edges, textures) in the target image data. As the number of layers increases, the image processing module can learn more complex feature representations and output corresponding feature maps.

[0107] Based on the pooling layer in the image processing module, the first feature map is subjected to dimensionality reduction processing to obtain a corresponding second feature map.

[0108] In this embodiment, the convolution layer is followed by a pooling layer (such as maximum pooling) to reduce the dimension of the feature map, reduce the amount of calculation, and increase the robustness of the feature.

[0109] Based on the fully connected layer in the image processing module, feature conversion processing is performed on the second feature map to obtain image features corresponding to the target image data.

[0110] In this embodiment, the feature maps output by the convolutional layer and the pooling layer are flattened and passed to the fully connected layer (Dense layer) to convert the feature maps into vectors of fixed length. The obtained vectors are the feature representations of the target image data, i.e., the above-mentioned image features.

[0111] The present application inputs the target image data into the image processing module in the inspection and damage assessment model; then performs local feature extraction on the target image data through the convolution layer in the image processing module to obtain a corresponding first feature map; then performs dimensionality reduction processing on the first feature map based on the pooling layer in the image processing module to obtain a corresponding second feature map; and subsequently performs feature conversion processing on the second feature map based on the fully connected layer in the image processing module to obtain image features corresponding to the target image data. The present application uses the convolution layer, pooling layer and fully connected layer in the image processing module to perform local feature extraction, dimensionality reduction processing and feature conversion processing on the target image data, respectively, so as to achieve fast and accurate feature extraction of the target image data and obtain corresponding image features, effectively improving the processing accuracy and processing intelligence of image feature extraction.

[0112] In some optional implementations of this embodiment, step S205 includes the following steps:

[0113] The target text data is input into the text processing module in the inspection and damage assessment model.

[0114] In this embodiment, the above-mentioned text processing module can specifically adopt an LSTM network (long short-term memory network). The number of LSTM layers and the number of hidden units corresponding to the LSTM network and whether to use bidirectional LSTM can be set according to actual business processing requirements, and the text processing flow corresponding to the LSTM network can be determined.

[0115] The target text data is preprocessed by the text processing module to obtain corresponding designated text data.

[0116] In this embodiment, the text preprocessing includes word segmentation, stop word removal, stem extraction or word form restoration of the target text data, so as to obtain the corresponding designated text data.

[0117] The specified text data is subjected to text encoding processing to obtain corresponding word vector data.

[0118] In this embodiment, the text encoding process refers to converting the specified text data into a numerical form. Specifically, each word corresponding to the specified text data can be converted into a vector of fixed dimension, i.e., the above-mentioned word vector data, by using word embedding technology. The obtained vector contains the semantic information of the word.

[0119] The word vector data is processed based on a preset sequence processing strategy to obtain text features corresponding to the target text data.

[0120] In this embodiment, for the LSTM network, the text is regarded as a sequence and is input into the network in sequence. In the LSTM network, a word vector is received at each time step, and the hidden state and cell state are updated. The strategy content of the above sequence processing strategy is to receive a word vector at each time step in the LSTM network, and update the hidden state and cell state accordingly.

[0121] The present application inputs the target text data into the text processing module in the inspection and damage assessment model; then performs text preprocessing on the target text data through the text processing module to obtain corresponding designated text data; then performs text encoding processing on the designated text data to obtain corresponding word vector data; and subsequently processes the word vector data based on a preset sequence processing strategy to obtain text features corresponding to the target text data. The present application performs text preprocessing, text encoding processing, and vector processing based on a sequence processing strategy on the target text data based on the use of a text processing module in the inspection and damage assessment model, thereby achieving rapid and accurate feature extraction of the target text data and obtaining corresponding text features, effectively improving the processing accuracy and processing intelligence of text feature extraction.

[0122] In some optional implementations, step S206 includes the following steps:

[0123] Get the preset fusion strategy.

[0124] In this embodiment, a fusion mechanism is pre-designed according to actual processing requirements to combine image features with text features to form a cross-modal joint representation. Specifically, fusion strategies include but are not limited to strategies such as concatenation, weighted sum, and attention mechanism. Among them. Concatenation refers to simply concatenating image features and text features into a longer vector. Weighted Sum refers to weighted summation of image features and text features according to certain weight distribution. Attention Mechanism refers to using the attention mechanism to dynamically adjust the importance of image features and text features in the fusion process.

[0125] The image feature and the text feature are fused based on the fusion strategy to obtain fused designated features.

[0126] In this embodiment, the image features and the text features can be fused according to the fusion processing steps included in the above fusion strategy to obtain the corresponding fused specified features. In addition, the image features and the text features can be fused at the fusion layer in the inspection and damage assessment model.

[0127] The designated feature is used as the target fusion feature.

[0128] The present application obtains a preset fusion strategy; then fuses the image features and the text features based on the fusion strategy to obtain the fused designated features; and subsequently uses the designated features as the target fusion features. The present application fuses the image features and the text features based on the use of a fusion strategy, so that the fusion processing of the image features and the text features can be completed quickly and accurately to obtain the target fusion features, thereby improving the processing efficiency of the feature fusion processing, ensuring the data accuracy of the obtained target fusion features, and facilitating the subsequent loss assessment processing of the target fusion features based on a pre-trained inspection and damage assessment model, so that the vehicle loss assessment results corresponding to the target vehicle can be accurately generated, thereby improving the processing accuracy and reliability of the vehicle loss assessment.

[0129] In some optional implementations, step S202 includes the following steps:

[0130] Perform image enhancement processing on the image data to obtain corresponding first image data.

[0131] In this embodiment, the image enhancement process refers to enhancing the brightness and contrast of an image by adjusting the pixel values ​​of the image, so as to achieve the effect of image enhancement. The pixel value adjustment process of the image data can be implemented by linear transformation.

[0132] The first image data is subjected to denoising processing to obtain corresponding second image data.

[0133] In this embodiment, the above-mentioned denoising process refers to a process of reducing noise in image data, and can be performed by using methods such as median filtering and Gaussian filtering.

[0134] The second image data is size normalized to obtain corresponding third image data.

[0135] In this embodiment, the size normalization process refers to adjusting the image data to a fixed size required by the model for batch processing.

[0136] The third image data is used as the target image data.

[0137] The present application performs image enhancement processing on the image data to obtain the corresponding first image data; then performs denoising processing on the first image data to obtain the corresponding second image data; then performs size normalization processing on the second image data to obtain the corresponding third image data; and subsequently uses the third image data as the target image data. The present application performs image enhancement processing, denoising processing, and size normalization processing on the image data, thereby achieving fast and accurate image preprocessing of the image data, effectively ensuring the accuracy and standardization of the obtained target image data, improving the instructions of the target image data, and facilitating the subsequent processing of the model.

[0138] In some optional implementations of this embodiment, before step S204, the electronic device may further perform the following steps:

[0139] Obtain pre-collected vehicle inspection and damage assessment record data.

[0140] In this embodiment, the vehicle inspection and damage assessment record data may be record data within a historical time period. There is no specific limitation on the time selection of the historical time period, for example, within the past year, or within the past three years, etc. In addition, the vehicle inspection and damage assessment record data includes but is not limited to accident pictures, process documents, processing results, accident reports, maintenance records, etc. In addition, in the process of collecting vehicle inspection and damage assessment record data, the comprehensiveness and diversity of the data can be further ensured, covering different vehicle types, different accident types and different degrees of damage.

[0141] The vehicle inspection and damage assessment record data is classified and processed based on a preset data dimension to obtain corresponding initial sample data.

[0142] In this embodiment, the above data dimensions may specifically include dimensions such as vehicle type, accident type, degree of damage, etc. The collected vehicle inspection and damage assessment record data may be classified according to the above data dimensions to obtain classified data, and then representative data samples may be screened out from the classified data to serve as the above initial sample data, and used for subsequent analysis and model training.

[0143] Perform data annotation processing on the initial sample data to obtain corresponding sample data.

[0144] In this embodiment, the generated initial sample data is annotated, which may include the degree of damage (such as slight, moderate, severe), vehicle type (such as sedan, truck, SUV, etc.), damage type (such as collision, scratch, flooding, etc.) and repair cost estimate, so as to obtain the corresponding sample data.

[0145] Determine a model structure of a preset multimodal deep learning model, and call the multimodal deep learning model.

[0146] In this embodiment, the above model structure can adopt a framework suitable for deep learning development, such as TensorFlow, PyTorch, etc. The multimodal deep learning model includes at least an image processing module, a text processing module, a target fully connected layer, and may also include a fusion layer. The multimodal deep learning model processes image data by combining the image processing module, processes text data by combining the text processing module, and effectively fuses image features and text features in the collection fusion layer, and performs loss assessment in combination with the target fully connected layer.

[0147] Obtain a target loss function corresponding to the multimodal deep learning model.

[0148] In this embodiment, there is no specific limitation on the selection of the above-mentioned target loss function, which can be determined according to actual task requirements. For example, mean square error (MSE) or cross-entropy loss (Cross-Entropy Loss) can be used.

[0149] Based on the target loss function, the multimodal deep learning model is cross-validated and trained using the sample data to obtain a specified model that meets preset model building expectations.

[0150] In this embodiment, the above cross-validation training process specifically includes: 1. Model training. Sample data is input into the multimodal deep learning model to obtain a prediction result. Then the loss value is calculated based on the prediction result and the true label. The model parameters are then updated using the gradient descent method (such as the Adam optimizer) to minimize the above target loss function. Subsequently, the data set is divided into multiple folds for K-fold cross-validation. Each time training is performed, a portion of the data is used as a validation set and the rest as a training set. Through multiple trainings, the stability and generalization ability of the model on different data sets are evaluated. 2. Model evaluation. Select appropriate evaluation indicators such as accuracy, recall, F1 score, mean square error (MSE), etc. according to task requirements. Then use an independent test set to evaluate the model and calculate the values ​​of each evaluation indicator. If it is detected that the values ​​of each evaluation indicator meet the preset evaluation threshold, it is determined that the model meets the expectations of model construction.

[0151] The designated model is used as the inspection and damage assessment model, and the inspection and damage assessment model is deployed on the cloud.

[0152] In this embodiment, after obtaining the inspection and damage assessment model, the inspection and damage assessment model can be further deployed in the cloud to utilize distributed computing resources to improve processing efficiency and response speed.

[0153] In addition, the above-mentioned inspection and damage assessment model can be trained not only for vehicle insurance, but also for home property insurance and agricultural insurance through corresponding data. As long as enough and high-quality data is provided, the model can be trained through the above process, thereby realizing the continuous expansion of the model's usage scenarios.

[0154] In addition, the survey and damage assessment model is also equipped with a feedback mechanism and a timed automatic update mechanism. If there are problems with the returned results, the model can be updated based on manual feedback to continuously improve the model. The model can also be continuously updated automatically to expand the data set to obtain more updated data to keep pace with the times.

[0155] This application obtains the vehicle inspection and damage assessment record data collected in advance; and classifies and processes the vehicle inspection and damage assessment record data based on the preset data dimension to obtain the corresponding initial sample data; then performs data annotation processing on the initial sample data to obtain the corresponding sample data; then determines the model structure of the preset multimodal deep learning model, and calls the multimodal deep learning model; subsequently obtains the target loss function corresponding to the multimodal deep learning model; further based on the target loss function, uses the sample data to cross-validate and train the multimodal deep learning model to obtain a designated model that meets the preset model construction expectations; finally, uses the designated model as the inspection and damage assessment model, and deploys the inspection and damage assessment model to the cloud. This application obtains sample data by classifying and processing the vehicle inspection and damage assessment record data collected in advance and performing data annotation processing, and then based on the target loss function, uses the sample data to cross-validate and train the multimodal deep learning model to obtain a designated model that meets the preset model construction expectations and is used as the final inspection and damage assessment model, which effectively improves the construction efficiency of the inspection and damage assessment model and ensures the model effect of the obtained inspection and damage assessment model.

[0156] In some optional implementations of this embodiment, after step S207, the electronic device may further perform the following steps:

[0157] Gets the preset report generation strategy.

[0158] In this embodiment, the strategy content of the above-mentioned report generation strategy is to generate a vehicle loss report having multiple detailed parts including text description, image annotation, repair plan diagram and quotation.

[0159] The vehicle loss assessment results are integrated and processed based on the report generation strategy to obtain a corresponding vehicle loss report.

[0160] In this embodiment, a pre-built vehicle loss report template can be obtained, and then according to the policy content of the above report generation strategy, all vehicle loss information contained in the vehicle loss assessment result can be integrated into the corresponding positions in the vehicle loss report template, thereby generating a corresponding vehicle loss report. Among them, the above vehicle loss report template at least includes a plurality of areas to be filled, such as text description, image annotation, repair plan diagram and quotation. The above vehicle loss information may at least include information such as the damaged parts of the target vehicle, the degree of damage, the repair quotation, and the repair suggestions.

[0161] Get the target push method.

[0162] In this embodiment, the above-mentioned target push method can be any one of SMS sending, email sending, interface display, etc.

[0163] Based on the target push method, the vehicle loss report is pushed to the user.

[0164] In this embodiment, the vehicle loss report can be pushed to the user by adopting the above-mentioned target push method to improve the push intelligence of the vehicle loss report.

[0165] This application obtains a preset report generation strategy; then integrates and processes the vehicle loss assessment results based on the report generation strategy to obtain a corresponding vehicle loss report; then obtains a target push method; and subsequently pushes the vehicle loss report to the user based on the target push method. This application performs loss assessment processing on the target fusion features based on the target fully connected layer in the investigation and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle. It will also automatically and intelligently integrate and process the vehicle loss assessment results based on the use of the report generation strategy to obtain a corresponding vehicle loss report, thereby effectively improving the generation efficiency and intelligence of the vehicle loss report. In addition, the vehicle loss report will also be pushed to the user based on the obtained target push method, so that the user can promptly and quickly understand the loss situation of the target vehicle by consulting the vehicle loss report, and then can execute corresponding processing measures, which is conducive to improving the user experience.

[0166] In some optional implementations, the user information obtained is subject to the user's consent and complies with relevant laws and policies.

[0167] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0168] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned vehicle loss assessment results, the above-mentioned vehicle loss assessment results can also be stored in a node of a blockchain.

[0169] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0170] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0171] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0172] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0173] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0174] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a vehicle damage assessment device based on artificial intelligence, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0175] like Figure 3As shown, the vehicle loss assessment device 300 based on artificial intelligence in this embodiment includes: a receiving module 301, a first processing module 302, a second processing module 303, a first extraction module 304, a second extraction module 305, a fusion module 306 and an assessment module 307. Among them:

[0176] The receiving module 301 is used to receive image data and text data related to the vehicle accident of the target vehicle input by the user;

[0177] A first processing module 302, configured to perform image preprocessing on the image data to obtain corresponding target image data;

[0178] The second processing module 303 is used to perform data integration processing on the text data to obtain corresponding target text data;

[0179] A first extraction module 304 is used to extract features from the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features;

[0180] The second extraction module 305 is used to extract features from the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features;

[0181] A fusion module 306 is used to fuse the image feature with the text feature to obtain a corresponding target fusion feature;

[0182] The evaluation module 307 is used to perform loss evaluation processing on the target fusion features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss evaluation result corresponding to the target vehicle.

[0183] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0184] In some optional implementations of this embodiment, the first extraction module 304 includes:

[0185] A first input submodule, used for inputting the target image data into the image processing module in the inspection and damage assessment model;

[0186] An extraction submodule, configured to extract local features of the target image data through a convolutional layer in the image processing module to obtain a corresponding first feature map;

[0187] A dimensionality reduction submodule, used for performing dimensionality reduction processing on the first feature map based on the pooling layer in the image processing module to obtain a corresponding second feature map;

[0188] A conversion submodule is used to perform feature conversion processing on the second feature map based on the fully connected layer in the image processing module to obtain image features corresponding to the target image data.

[0189] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0190] In some optional implementations of this embodiment, the second extraction module 305 includes:

[0191] A second input submodule, used for inputting the target text data into the text processing module in the inspection and damage assessment model;

[0192] A preprocessing submodule, used for performing text preprocessing on the target text data through the text processing module to obtain corresponding designated text data;

[0193] The encoding submodule is used to perform text encoding processing on the specified text data to obtain corresponding word vector data;

[0194] The first processing submodule is used to process the word vector data based on a preset sequence processing strategy to obtain text features corresponding to the target text data.

[0195] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0196] In some optional implementations of this embodiment, the fusion module 306 includes:

[0197] The acquisition submodule is used to obtain the preset fusion strategy;

[0198] A fusion submodule, used for fusing the image feature and the text feature based on the fusion strategy to obtain a fused specified feature;

[0199] The first determination submodule is used to use the specified feature as the target fusion feature.

[0200] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0201] In some optional implementations of this embodiment, the first processing module 302 includes:

[0202] A second processing submodule is used to perform image enhancement processing on the image data to obtain corresponding first image data;

[0203] A third processing submodule, configured to perform denoising on the first image data to obtain corresponding second image data;

[0204] a fourth processing submodule, configured to perform size normalization processing on the second image data to obtain corresponding third image data;

[0205] The second determining submodule is configured to use the third image data as the target image data.

[0206] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0207] In some optional implementations of this embodiment, the vehicle damage assessment device based on artificial intelligence further includes:

[0208] The first acquisition module is used to acquire the vehicle inspection and damage assessment record data collected in advance;

[0209] A classification module, used to classify the vehicle inspection and damage assessment record data based on a preset data dimension to obtain corresponding initial sample data;

[0210] A labeling module, used to perform data labeling processing on the initial sample data to obtain corresponding sample data;

[0211] A calling module, used to determine the model structure of a preset multimodal deep learning model and call the multimodal deep learning model;

[0212] A determination module, used to obtain a target loss function corresponding to the multimodal deep learning model;

[0213] A training module, used to perform cross-validation training on the multimodal deep learning model using the sample data based on the target loss function to obtain a specified model that meets the preset model construction expectations;

[0214] A deployment module is used to use the designated model as the inspection and damage assessment model, and deploy the inspection and damage assessment model to the cloud.

[0215] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0216] In some optional implementations of this embodiment, the vehicle damage assessment device based on artificial intelligence further includes:

[0217] A second acquisition module is used to acquire a preset report generation strategy;

[0218] An integration module, used to integrate the vehicle loss assessment results based on the report generation strategy to obtain a corresponding vehicle loss report;

[0219] The third acquisition module is used to obtain the target push mode;

[0220] A push module is used to push the vehicle loss report to the user based on the target push method.

[0221] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the vehicle loss assessment method based on artificial intelligence in the aforementioned embodiment, and will not be repeated here.

[0222] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0223] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0224] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0225] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the vehicle loss assessment method based on artificial intelligence, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0226] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for running the artificial intelligence-based vehicle loss assessment method.

[0227] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0228] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the above-mentioned artificial intelligence-based vehicle loss assessment method.

[0229] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0230] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A vehicle loss assessment method based on artificial intelligence, characterized in that: The steps include: receiving image data and text data related to a vehicle accident of a target vehicle input by a user; Performing image preprocessing on the image data to obtain corresponding target image data; Performing data integration processing on the text data to obtain corresponding target text data; Extracting features of the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features; Perform feature extraction on the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features; Fusing the image feature with the text feature to obtain a corresponding target fusion feature; Based on the target fully connected layer in the inspection and damage assessment model, the target fusion features are processed for loss assessment to obtain a vehicle loss assessment result corresponding to the target vehicle.

2. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: The step of extracting features from the target image data based on the image processing module in the pre-built survey and damage assessment model to obtain corresponding image features specifically includes: Inputting the target image data into the image processing module in the survey and damage assessment model; Extracting local features of the target image data through a convolutional layer in the image processing module to obtain a corresponding first feature map; Performing dimensionality reduction processing on the first feature map based on the pooling layer in the image processing module to obtain a corresponding second feature map; Based on the fully connected layer in the image processing module, feature conversion processing is performed on the second feature map to obtain image features corresponding to the target image data.

3. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: The step of extracting features from the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features specifically includes: Inputting the target text data into the text processing module in the inspection and damage assessment model; Performing text preprocessing on the target text data by the text processing module to obtain corresponding designated text data; Perform text encoding processing on the specified text data to obtain corresponding word vector data; The word vector data is processed based on a preset sequence processing strategy to obtain text features corresponding to the target text data.

4. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: The step of fusing the image feature with the text feature to obtain the corresponding target fusion feature specifically includes: Get the preset fusion strategy; Based on the fusion strategy, the image feature and the text feature are fused to obtain a fused specified feature; The designated feature is used as the target fusion feature.

5. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: The step of performing image preprocessing on the image data to obtain corresponding target image data specifically includes: Performing image enhancement processing on the image data to obtain corresponding first image data; Performing denoising processing on the first image data to obtain corresponding second image data; performing size normalization processing on the second image data to obtain corresponding third image data; The third image data is used as the target image data.

6. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: Before the step of extracting features from the target image data by the image processing module in the pre-built survey and damage assessment model to obtain corresponding image features, the method further includes: Obtain pre-collected vehicle inspection and damage assessment record data; Classify and process the vehicle inspection and damage assessment record data based on a preset data dimension to obtain corresponding initial sample data; Performing data annotation processing on the initial sample data to obtain corresponding sample data; Determine a model structure of a preset multimodal deep learning model, and call the multimodal deep learning model; Obtaining a target loss function corresponding to the multimodal deep learning model; Based on the target loss function, the multimodal deep learning model is cross-validated and trained using the sample data to obtain a specified model that meets the preset model construction expectations; The designated model is used as the inspection and damage assessment model, and the inspection and damage assessment model is deployed on the cloud.

7. The vehicle loss assessment method based on artificial intelligence according to claim 1 is characterized in that: After the step of performing loss assessment processing on the target fusion feature based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle, the method further includes: Get the preset report generation strategy; Integrate the vehicle loss assessment results based on the report generation strategy to obtain a corresponding vehicle loss report; Get the target push method; Based on the target push method, the vehicle loss report is pushed to the user.

8. A vehicle loss assessment device based on artificial intelligence, characterized in that: include: A receiving module, used for receiving image data and text data related to a vehicle accident of a target vehicle input by a user; A first processing module, used for performing image preprocessing on the image data to obtain corresponding target image data; A second processing module, used for performing data integration processing on the text data to obtain corresponding target text data; A first extraction module is used to extract features from the target image data based on an image processing module in a pre-built survey and damage assessment model to obtain corresponding image features; A second extraction module is used to extract features from the target text data based on the text processing module in the inspection and damage assessment model to obtain corresponding text features; A fusion module, used for fusing the image features with the text features to obtain corresponding target fusion features; An assessment module is used to perform loss assessment processing on the target fusion features based on the target fully connected layer in the inspection and damage assessment model to obtain a vehicle loss assessment result corresponding to the target vehicle.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the vehicle loss assessment method based on artificial intelligence as described in any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the vehicle damage assessment method based on artificial intelligence as described in any one of claims 1 to 7.

Images