Road surface information learning method, system and application

Abstract

The invention belongs to the technical field of visual image processing and computer vision, and provides a road surface information learning method and system and application, and the method comprises the steps: carrying out the preprocessing of low-resolution original road data, and obtaining a training data set and a test data set; constructing a core module which comprises a multi-source information interaction attention module and a multi-scale feature fusion module; respectively constructing a pavement humidity scale model and a pavement height scale model based on the core module; performing joint training on the models by using a training data set, and testing by using a test data set after the training is completed to obtain a trained pavement road parameter scale model; and obtaining to-be-processed high-resolution road coefficient information data, and inputting the to-be-processed high-resolution road coefficient information data into the pavement road parameter scale model for processing to obtain a high-resolution pavement parameter information product. By quickly and accurately identifying the road surface information, the driving system can be assisted to quickly adapt to the road surface, so that the vehicle passing comfort is improved.

Description

Technical Field

[0001] This application relates to the fields of visual image processing and computer vision technology, and in particular to a road information learning method, system and application. Background Technology

[0002] Currently, there is a significant delay in vehicles' perception and feedback of road information, preventing them from responding promptly to sudden road conditions (such as potholes and speed bumps). The core of this problem lies in the untimely updating of road information and the lack of an effective vehicle-road cooperative mechanism. When a vehicle enters a poor road section, due to the failure to obtain road surface information in advance, the chassis system (such as suspension and ride height control) cannot adjust to a corresponding state in advance, and the vehicle speed also fails to decrease in time. This passive approach forces the vehicle to withstand direct impacts from the road surface, severely affecting ride comfort and causing impact damage to tires, wheels, and chassis structure. In the long run, this will accelerate component wear and aging, directly leading to decreased vehicle passability and potential safety hazards.

[0003] Currently, there are various methods for road condition perception. For example, patent application CN120766231A discloses a road condition recognition method based on machine vision and force feedback, including the following steps: S1, continuously capturing road surface image streams using a vehicle front-end image acquisition device; S2, automatically segmenting abnormal road surface areas using an arbitrary target segmentation model; S3, arranging piezoelectric thin film vibration sensing units within the contact surface between the vehicle tires and the road surface to collect minute vibration signals in real time; S4, extracting time-domain and frequency-domain features of the vibration signals; S5, performing multi-scale encoding of visual and force feedback features; S6, dynamically adjusting the weights of the two-modal features to generate fused decision features; S7, outputting the road condition recognition result. This method is intended to improve the safety of autonomous driving. However, this method focuses on real-time perception and recognition, only able to perceive road conditions that the vehicle has already approached or is currently passing, lacking macroscopic perception capabilities for long-distance, large-scale road conditions ahead of the vehicle. Therefore, its application lacks foresight and predictive ability.

[0004] For example, patent application CN120726604A discloses a road surface recognition method, device, medium, and vehicle-mounted intelligent system based on an improved ShuffleNetV2 model. The method includes: acquiring road surface image data to be identified; inputting the road surface image data into a pre-trained improved ShuffleNetV2 model to obtain a prediction result of the road surface type. Specifically, an SENet module is introduced into the main branch of each basic unit of the improved ShuffleNetV2 model, and the ReLU activation function in the SENet module is replaced with an H-Swish activation function; the prediction result of the road surface type is fed back to the intelligent driving system, enabling the intelligent driving system to implement autonomous driving decision-making and vehicle driving mode control functions based on the road surface type prediction result. Thus, by introducing an SENet module into the main branch of the basic unit of the improved ShuffleNetV2 model, the model's road surface feature extraction capability is improved. Furthermore, by replacing the ReLU activation function in the SE module with an H-Swish activation function, the convergence speed is accelerated and the model's nonlinear expression capability is improved, thereby achieving efficient identification of complex road surface types. However, this method outputs a road surface type prediction as a classification label and its corresponding probability, which cannot provide continuous, quantitative physical parameters for precise vehicle control. Furthermore, the model architecture of this method is designed for general image classification, not specifically built for learning complex road surface information. It lacks an inherent mechanism for processing multi-source, heterogeneous data. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the present invention provides a road information learning method, system and application, which can assist the driving system to quickly adapt to the road surface by quickly and accurately identifying road information, thereby improving the comfort of vehicle passage and solving the problem of tire, wheel and chassis damage caused by poor road passability.

[0006] To achieve the above and related objectives, the present invention adopts the following technical solution:

[0007] The first aspect of this invention provides a method for learning road surface information, comprising the following steps:

[0008] Step S100: Preprocess the low-resolution raw road data to obtain the training dataset and the test dataset.

[0009] Step S200: Construct the core modules, including a multi-source information interaction attention module and a multi-scale feature fusion module; construct the road surface humidity scale model and the road surface height scale model based on the core modules respectively;

[0010] Step S300: Use the training dataset to jointly train the road surface moisture scale model and the road surface height scale model, and use the test dataset to test after training to obtain the trained road surface parameter scale model. The joint training includes supervised training and degradation training.

[0011] Step S400: Obtain the high-resolution road coefficient information data to be processed, input it into the road surface parameter scale model for processing, and obtain high-resolution road surface parameter information products.

[0012] Furthermore, the preprocessing in step S100 includes: performing a quality assessment on the original road data; if it is determined that there are missing data, then performing missing data imputation, resampling, and data augmentation operations to generate training and testing datasets.

[0013] Furthermore, in step S200, the multi-source information interaction attention module includes a single-to-multi-factor interaction attention sub-module; the multi-scale feature fusion module includes a residual space channel attention sub-module.

[0014] Further, in step S300, low-resolution road coefficient information data, low-resolution road surface moisture information data, and low-resolution road surface elevation difference information data are obtained from the training dataset to generate a supervised training dataset.

[0015] Furthermore, in step S300, supervised training includes: using a supervised training dataset and its corresponding high-resolution ground truth data for supervised learning.

[0016] Furthermore, in step S300, the degradation training includes: comparing and learning low-resolution data with corresponding high-resolution ground truth data, wherein the low-resolution data is artificially generated from high-resolution road parameter information data through degradation.

[0017] Furthermore, in step S300, the joint training is optimized using an adaptive dual-domain joint loss function until the road surface parameter scaling model converges.

[0018] A second aspect of the present invention provides a road surface information learning system, comprising:

[0019] The data processing module is used to preprocess the low-resolution raw road data to obtain training and test datasets.

[0020] The model building module is used to construct the core modules, including a multi-source information interaction attention module and a multi-scale feature fusion module; based on the core modules, a road surface humidity scale model and a road surface height scale model are constructed respectively.

[0021] The model training module is used to jointly train the road surface moisture scale model and the road surface height scale model using the training dataset, and to test the model using the test dataset after training to obtain the trained road surface parameter scale model. The joint training includes supervised training and degradation training.

[0022] The information generation module is used to acquire high-resolution road coefficient information data to be processed, input it into the road surface parameter scale model for processing, and obtain high-resolution road surface parameter information products.

[0023] A third aspect of the present invention provides a computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by a computer's processor, cause the computer to perform the aforementioned road information learning method.

[0024] A fourth aspect of the present invention provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the road information learning method described above.

[0025] The beneficial technical effects of this invention are as follows:

[0026] This invention achieves deep interaction and enhancement of multiple elements such as road coefficient, humidity, and elevation difference through a multi-source information interaction attention module, thereby improving information fusion and feature extraction capabilities. Through a multi-scale feature fusion module, key features are accurately extracted and calibrated from low-resolution data to generate high-resolution information, thereby achieving high-precision road surface information inversion.

[0027] This invention employs a joint training strategy of supervised training and degenerate training, enabling the model to not only learn ideal mappings but also cope with complex scenarios with poor data quality. This enhances the anti-interference ability and generalization performance of road surface parameter scale models in practical applications.

[0028] This invention generates high-resolution road surface parameter information products, which can assist intelligent driving systems in perceiving road conditions in advance, thereby proactively adjusting parameters such as vehicle speed, suspension stiffness, and chassis height to effectively reduce damage to tires, rims, and chassis, improve vehicle passability and ride comfort, and enhance adaptability to different road surfaces.

[0029] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0030] The accompanying drawings, incorporated in and forming part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without inventive effort. In the drawings:

[0031] Figure 1 This is a flowchart of the road surface information learning method in this application;

[0032] Figure 2 This is a flowchart of another exemplary road surface information learning method of this application;

[0033] Figure 3 This is a framework diagram of the road surface information learning system of this application;

[0034] Figure 4 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. Detailed Implementation

[0035] Unless otherwise defined, all technical and / or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should be understood that certain features of the invention (described in the context of separate embodiments for clarity) may also be provided in a single embodiment. Conversely, multiple features of the invention (described in the context of a single embodiment for brevity) may also be provided separately or in any suitable combination or, where appropriate, in any other described embodiment of the invention. Certain features described in the context of various embodiments will not be considered essential features of those embodiments unless the embodiment is inoperable without those elements. The invention is further illustrated below by specific examples; however, it should be noted that the specific process conditions and results described in the embodiments of the invention are merely illustrative and should not be construed as limiting the scope of protection of the invention. All equivalent changes or modifications made in accordance with the spirit and essence of the invention should be covered within the scope of protection of the invention.

[0036] Please see Figure 1 The flowchart of the road surface information learning method of this application is detailed below:

[0037] Step S100: Preprocess the low-resolution raw road data to obtain the training dataset and the test dataset.

[0038] Specifically, in combination Figure 2The preprocessing in this application includes: quality assessment of the original road data; if data missing is determined, missing data imputation, resampling, and data augmentation operations are performed to generate training and testing datasets. More specifically, the original road data in this application is a multi-source heterogeneous dataset, including data from various channels such as vehicle sensors, roadside equipment, high-precision maps, and crowdsourced data, with relatively low resolution or accuracy, but controllable cost and easy large-scale acquisition. This application first determines whether the acquired original road data has quality issues related to missing data. If missing data exists, the following operations are performed in sequence: 1) Missing data imputation: interpolating and filling in missing or invalid values ​​in the original road data to ensure data integrity; 2) Resampling: including spatial resampling and temporal resampling to unify data from different sources and of different specifications onto the same spatial or temporal reference; 3) Data augmentation: expanding the existing data through algorithms such as geometric transformation and mixed sampling to increase the scale and diversity of the dataset.

[0039] Step S200: Construct core modules, including a multi-source information interaction attention module and a multi-scale feature fusion module; construct road surface humidity scale model and road surface height scale model based on the core modules respectively.

[0040] Specifically, the multi-source information interaction attention module of this application includes a single-to-multi-factor interaction attention submodule. This submodule is used to process multi-source auxiliary data from different sources, such as road coefficients, humidity, and height data. The single-to-multi-factor interaction attention submodule of this application traverses the multi-source auxiliary data, acquiring the numerical distribution characteristics of single elements and the spatial texture information of multiple elements, and performs feature interaction enhancement to generate multiple sets of single-to-multi-factor interaction features. Furthermore, this application fuses the multiple sets of single-to-multi-factor interaction features through cascade operators to generate multi-source information interaction features, thereby constructing the multi-source information interaction attention module. More specifically, this application, through the single-to-multi-factor interaction attention submodule, can simultaneously focus on the numerical distribution of a single parameter and the spatial texture correlation of multiple parameters, achieving interactive perception. Through this interactive perception, the multi-source information interaction attention module of this application can uncover deeper correlation features than the original road data, deeply exploring the intrinsic correlation between road surface information parameters and generating fused features after interactive enhancement.

[0041] Specifically, the multi-scale feature fusion module of this application includes a residual spatial channel attention submodule, which is used to refine and integrate the fused features to adapt to road surface features of different sizes. This application uses the residual spatial channel attention submodule to extract key road surface spatial features and road surface spectral features, and performs feature weight calibration to obtain a feature map that is doubly enhanced by both key road surface spatial features and road surface spectral features; then, feature-level fusion is performed based on the feature map to obtain feature map images at different scales, thereby constructing the multi-scale feature fusion module. More specifically, this application mines the mapping relationship between road surface parameters by exploring the information interaction between them, achieving efficient fusion of multi-source data and realizing effective extraction and accurate calibration of multi-scale features.

[0042] Specifically, the method for constructing a road surface moisture scale model in this application includes: using a single-to-multi-factor interactive attention submodule to perform information interaction between single and multi-factor data in low-resolution road coefficient information data, generating multi-source information interaction features; based on the multi-source information interaction features, using a residual spatial channel attention submodule to calibrate and extract road surface spectral features, generating feature mapping maps at different scales, thereby constructing a road surface moisture scale model. The feature mapping maps under this model are used to accurately reflect the dryness and wetness of the road surface and changes in material properties.

[0043] Specifically, the method for constructing the pavement height scale model in this application includes: using a single-to-multi-factor interactive attention submodule to perform information interaction between single and multi-factor data in low-resolution road coefficient information data, generating multi-source information interaction features; based on the multi-source information interaction features, using a residual space channel attention submodule to calibrate and extract pavement height features, generating feature maps at different scales, thereby constructing the pavement height scale model. The feature maps under this model are used to accurately reflect pavement undulations, pothole depths, etc.

[0044] Step S300: Jointly train the road surface moisture scale model and the road surface height scale model using the training dataset, and test them using the test dataset after training to obtain the trained road surface parameter scale model. The joint training includes supervised training and degradation training.

[0045] Specifically, this application obtains low-resolution road coefficient information data, low-resolution pavement moisture information data, and low-resolution pavement height information data from the training dataset to generate a supervised training dataset. This application also obtains high-quality, high-resolution ground truth data that spatially corresponds precisely to the aforementioned low-resolution data from the training dataset. More specifically, the supervised training of this application includes: supervised learning using the supervised training dataset and its corresponding high-resolution ground truth data. This application inputs the supervised training dataset into the pavement moisture scale model and the pavement height scale model respectively, outputting their corresponding high-resolution prediction results. These prediction results are compared with the actual high-resolution ground truth data. An adaptive dual-domain joint loss function is used to calculate the difference between the prediction and the ground truth. The backpropagation algorithm is then used to continuously adjust the model's internal parameters, making the model's prediction results infinitely close to the ground truth, thus completing optimization and ensuring that the model learns accurate and physically reliable pavement parameter mapping patterns.

[0046] Specifically, the degradation training method of this application includes: comparing low-resolution data with corresponding high-resolution ground truth data, wherein the low-resolution data is artificially generated by degradation of high-resolution road parameter information data. More specifically, this application inputs the artificially degraded low-resolution data into the road surface moisture scale model and the road surface height scale model respectively, and compares their outputs with the original, undegraded high-resolution ground truth data. An adaptive dual-domain joint loss function is used to calculate the difference between the prediction and the ground truth, and the backpropagation algorithm is used to continuously adjust the internal parameters of the model, making the model's prediction result infinitely close to the ground truth, thus completing the optimization. During the degradation training process, the model learns to inversely recover the high-frequency details and real information lost during the degradation process, thereby endowing the model with information recovery and anti-interference capabilities, enabling it to handle real-world data of varying quality.

[0047] Specifically, after the training cycle, this application evaluates the model performance using an independent test dataset, which also contains pairs of low-resolution inputs and high-resolution ground truth values. Through testing, this application assesses the model's generalization ability, prevents overfitting, and selects the best-performing model as the trained road surface parameter scaling model, also known as a road surface parameter network.

[0048] Step S400: Obtain the high-resolution road coefficient information data to be processed, input it into the road surface parameter scale model for processing, and obtain high-resolution road surface parameter information products.

[0049] Specifically, the high-resolution road surface parameter information products generated by this application, such as the precise three-dimensional shape of potholes and the distribution of road surface slippage, can be input into the intelligent driving domain controller to achieve advance and precise passability control, assisting the intelligent driving system to quickly adjust its own chassis status, adjust vehicle speed, and adjust the vehicle suspension system to smoothly pass through potholes and other road surfaces.

[0050] Please see Figure 3 Here is a framework diagram of the road surface information learning system 300 of this application, including:

[0051] The data processing module 310 is used to preprocess the low-resolution raw road data to obtain training and test datasets.

[0052] Model building module 320 is used to build core modules, including a multi-source information interaction attention module and a multi-scale feature fusion module; based on the core modules, road surface humidity scale model and road surface height scale model are built respectively.

[0053] The model training module 330 is used to jointly train the road surface moisture scale model and the road surface height scale model using the training dataset, and to test the model using the test dataset after training to obtain the trained road surface parameter scale model. The joint training includes supervised training and degradation training.

[0054] The information generation module 340 is used to acquire high-resolution road coefficient information data to be processed, input it into the road surface parameter scale model for processing, and obtain high-resolution road surface parameter information products.

[0055] It should be noted that the road information learning system and the road information learning method provided in the above embodiments belong to the same concept. The specific methods of operation of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the road information learning system provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0056] Embodiments of this application also provide a computer device, including: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the computer device to implement the road information learning method provided in the above embodiments.

[0057] Figure 4 A schematic diagram of the structure of a computer system suitable for an embodiment of this application is shown. It should be noted that... Figure 4 The computer system 400 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0058] like Figure 4 As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage section 408 into a random access memory (RAM) 403, such as performing the methods described in the above embodiments. Various programs and data required for system operation are also stored in the RAM 403. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404. The following components are connected to the I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (local area network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A driver 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as disks, optical discs, magneto-optical discs, semiconductor memories, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0059] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer tool programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs various functions defined in the system of this application.

[0060] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, flash memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. Computer programs contained on computer-readable media can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0061] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0062] The units described in the embodiments of this application can be implemented by tools or by hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the unit itself.

[0063] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the road information learning method as described above. This computer-readable storage medium may be included in the computer device described in the above embodiments, or it may exist independently and not incorporated into the computer device.

[0064] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the road information learning method provided in the various embodiments described above.

[0065] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A road surface information learning method characterized by comprising: The method comprises the following steps: Step S100, preprocessing low-resolution original road data to obtain a training data set and a test data set; Step S200, constructing a core module, including a multi-source information interaction attention module and a multi-scale feature fusion module; Based on the core module, a road surface humidity scale model and a road surface height scale model are constructed respectively; Step S300, using the training data set to jointly train the road surface humidity scale model and the road surface height scale model, and after training is completed, using the test data set to test, obtaining a trained road surface road parameter scale model, wherein the joint training includes supervised training and degenerative training; Step S400, obtaining high-resolution road coefficient information data to be processed, inputting it into the road surface road parameter scale model for processing, and obtaining high-resolution road surface parameter information products.

2. The method of claim 1, wherein, The preprocessing in step S100 includes: quality judgment on the original road data, if it is judged that there is data missing, performing missing filling, resampling and data augmentation operations to generate the training data set and the test data set.

3. The method of claim 1, wherein, In step S200, the multi-source information interaction attention module includes a single-multielement interaction attention submodule; the multi-scale feature fusion module includes a residual spatial channel attention submodule.

4. The method of claim 3, wherein, In step S300, low-resolution road coefficient information data, low-resolution road surface humidity information data and low-resolution road surface fall information data are obtained from the training data set to generate a supervised training data set.

5. The method of claim 4, wherein, In step S300, the supervised training includes: using the supervised training data set and its corresponding high-resolution true value data for supervised learning.

6. The method of claim 4, wherein, In step S300, the degenerative training includes: comparing low-resolution data with corresponding high-resolution true value data for contrast learning, wherein the low-resolution data is artificially degraded from high-resolution road parameter information data.

7. The method of claim 6, wherein, In step S300, the joint training is optimized using an adaptive dual-domain joint loss function until the road surface road parameter scale model converges.

8. A road surface information learning system characterized by comprising: It comprises: A data processing module for preprocessing low-resolution original road data to obtain a training data set and a test data set; A model construction module for constructing a core module, including a multi-source information interaction attention module and a multi-scale feature fusion module; Based on the core module, a road surface humidity scale model and a road surface height scale model are constructed respectively; A model training module for using the training data set to jointly train the road surface humidity scale model and the road surface height scale model, and after training is completed, using the test data set to test, obtaining a trained road surface road parameter scale model, wherein the joint training includes supervised training and degenerative training; An information generation module for obtaining high-resolution road coefficient information data to be processed, inputting it into the road surface road parameter scale model for processing, and obtaining high-resolution road surface parameter information products.

9. A computer-readable storage medium, characterized in that, A computer-readable medium storing computer readable instructions which, when executed by a processor of a computer, cause the computer to perform the road surface information learning method according to any one of claims 1 to 7.

10. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the road surface information learning method according to any one of claims 1 to 7.

Images