Data mining method, device, medium and system
Through the combination of text and graphics models and picture search models, automated processing of autonomous driving data mining has been solved, the problem of inefficiency in the existing technology has been solved, the quality and adaptability of data mining have been improved, and automation and iterative optimization have been realized.
Patent Information
- Application Number
- CN202411034252.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
The existing autonomous driving data mining methods are inefficient, difficult to adapt to new scenarios and complex environments, and lack adaptive and automatic optimization capabilities, making it difficult to meet the needs of rapid iteration and updates.
The target road test data and configuration files are filtered using the text and picture search model. The target map is determined by filtering the matching degree between the text and picture description text and the matching degree between the picture search model and the reference image. Relationships are stored in the database, and the model is iteratively updated to improve mining efficiency and quality.
It realizes automated data mining and verification, improves the efficiency and quality of data mining, solves the problem of inefficiency of existing methods, and enhances the scalability and adaptability of the system.
Smart Images

Figure CN118964444A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of autonomous driving, and more specifically, to a method, device, medium and system for data mining. Background Art
[0002] At present, scene data mining and retrieval in autonomous driving and related fields usually rely on manual mining, which involves manually reviewing past data pools and filtering out scenes with specific conditions. In addition, there are some automatic mining methods that are based on specific handwritten rules, which filter out scenes that meet the conditions by setting specific thresholds and conditions. For example, in an autonomous driving environment, it may be necessary to identify and classify road conditions, traffic signs, pedestrians, and other dynamic or static objects. These handwritten rules are often based on long-term accumulated experience and are determined through a large number of experiments and tests.
[0003] Disadvantages of existing solutions:
[0004] Low mining efficiency: Existing manual and handwritten rule-based methods are still relatively inefficient in the data mining process. Massive amounts of data need to be processed and verified one by one, making the entire process cumbersome and time-consuming. The design and adjustment of handwritten rules require professional knowledge and a large number of experimental verifications. Once new scenarios or new requirements emerge, the rules must be redesigned and optimized.
[0005] Poor data scalability: Fixed-rule mining methods have limited adaptability to data and poor processing scalability. When encountering new or complex scenarios, existing rules often cannot fully adapt and require human intervention to adjust or re-formulate new rules. The diversity of the environment and the complexity of driving scenarios require the system to be highly scalable, but the existing handwritten rules are not flexible enough to support large-scale and diverse data processing needs.
[0006] Small optimization space: Since the rules are determined based on experience accumulation and experiments, the optimization space is limited. It is difficult for human-written rules to cover all possible scenarios and situations, and the process of adjusting and optimizing these rules is complicated and inefficient. Autonomous driving systems need to constantly learn and adapt to new situations, while traditional rules and methods lack the ability to adapt and automatically optimize, and it is difficult to meet the needs of rapid iteration and updating. Summary of the invention
[0007] The main purpose of the present application is to provide a method, device, medium and system for data mining, so as to at least solve the problem that the efficiency of existing solutions in the data mining process using manual and handwritten rule-based methods is still relatively low.
[0008] In order to achieve the above objective, according to one aspect of the present application, a data mining method is provided, the method comprising:
[0009] Obtain target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring; use a text-image model to filter and process all the image description texts in the target drive test data and the configuration file to obtain a plurality of first target drive test images, and use an image-based search model to filter and process the first target drive test image and all the preset reference images in the configuration file to obtain a plurality of second target drive test images; determine a target mapping relationship as a mapping relationship between each of the second target drive test images and the corresponding image description text, and store the target mapping relationship in a database.
[0010] Optionally, a text-image model is used to filter and process all the image description texts in the target road test data and the configuration file to obtain multiple first target road test images, including: using a text-image model to process all the image description texts in the target road test data and the configuration file to obtain multiple first matching degrees; determining that the first target road test image is the road test image whose first matching degree is greater than or equal to a first matching degree threshold.
[0011] Optionally, the first target road test image and all the preset reference images in the configuration file are screened and processed using the image search model to obtain multiple second target road test images, including: using the image search model to process the first target road test image and all the preset reference images in the configuration file to obtain multiple second matching degrees; determining that the second target road test image is the first target road test image whose second matching degree is greater than or equal to a second matching degree threshold.
[0012] Optionally, the method further includes: determining the first matching degree threshold as a preset matching degree threshold corresponding to a range in which an average value of all matching degrees obtained during the last training of the text-image model lies.
[0013] Optionally, the method further includes: determining a preset matching degree threshold corresponding to a range where an average value of all first matching degrees obtained by processing the target drive test data and all the image description texts in the configuration file by the text-image model is within a second matching degree threshold.
[0014] Optionally, after storing the target mapping relationship in a database, the method further includes: using the target mapping relationship to iteratively update the text-image model and the image-based search model respectively to obtain an optimized text-image model and an optimized image-based search model.
[0015] Optionally, obtaining target drive test data includes: obtaining initial drive test data, where the initial drive test data is a drive test video; and performing data segmentation processing on the drive test video to obtain a plurality of target drive test data.
[0016] According to another aspect of the present application, a data mining device is provided, the device comprising:
[0017] An acquisition unit, configured to acquire target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring;
[0018] a first processing unit, configured to use a text-image model to filter the target drive test data and all the image description texts in the configuration file to obtain a plurality of first target drive test images, and use an image-based search model to filter the first target drive test images and all the preset reference images in the configuration file to obtain a plurality of second target drive test images;
[0019] The second processing unit is used to determine that the target mapping relationship is a mapping relationship between each of the second target drive test images and the corresponding image description text, and store the target mapping relationship in a database.
[0020] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the data mining methods.
[0021] According to another aspect of the present application, a vehicle system is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing any one of the data mining methods described.
[0022] By applying the technical solution of the present application, the present application provides a method, device, medium and vehicle system for data mining. The method adopts a text-image model to filter and process the target road test data and all the image description texts in the configuration file to obtain multiple first target road test images, and adopts an image-based search model to filter and process the first target road test images and all the preset reference images in the configuration file to obtain multiple second target road test images, thereby achieving the purpose of double screening, thereby improving the quality and efficiency of data mining; realizing automated data mining and verification, and solving the problem that the efficiency of existing solutions using manual and handwritten rule-based methods in the data mining process is still relatively low. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0024] Figure 1 A schematic diagram of a data mining method according to an embodiment of the present application is shown;
[0025] Figure 2 A schematic flow chart of another data mining method provided according to an embodiment of the present application is shown;
[0026] Figure 3 A structural block diagram of a data mining device provided according to an embodiment of the present application is shown. DETAILED DESCRIPTION
[0027] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0028] 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 drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0030] As introduced in the background technology, the existing manual and handwritten rule-based methods are still relatively inefficient in the data mining process. Massive amounts of data need to be processed and verified one by one, making the entire process cumbersome and time-consuming. The design and adjustment of handwritten rules require professional knowledge and a large number of experimental verifications. Once new scenarios or new requirements appear, the rules must be redesigned and optimized. The mining method of fixed rules has limited adaptability to data and poor processing scalability. When encountering new or complex scenarios, the existing rules often cannot fully adapt, and human intervention is required to adjust or re-formulate new rules. The diversity of the environment and the complexity of the driving scenarios require the system to be highly scalable, but the existing handwritten rules are not flexible enough to support large-scale and diverse data processing needs. Since the rules are determined based on experience accumulation and experiments, the optimization space is limited. It is difficult for manually written rules to cover all possible scenarios and situations, and the process of adjusting and optimizing these rules is complex and inefficient. Autonomous driving systems need to constantly learn and adapt to new situations, but traditional rules and methods lack the ability to adapt and automatically optimize, and are unable to meet the needs of rapid iteration and updating. In order to solve the problem that the existing solutions use manual and handwritten rule-based methods in the data mining process, the efficiency is still relatively low, the embodiments of the present application provide a data mining method, device, medium and system.
[0031] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0032] In this embodiment, a data mining method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0033] Figure 1 is a flow chart of a data mining method according to an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:
[0034] Step S101, obtaining target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring;
[0035] Specifically, for example, in a "driving scene in the rain", in a 30-second autonomous driving scene, some front camera images are sampled in each frame.
[0036] The step S101 of acquiring target drive test data includes: acquiring initial drive test data, where the initial drive test data is a drive test video; and performing data segmentation processing on the drive test video to obtain a plurality of target drive test data.
[0037] Specifically, the segmentation process can achieve the purpose of processing the drive test video in segments, thereby improving the processing efficiency.
[0038] Step S102, using a text-image model to filter the target drive test data and all the image description texts in the configuration file to obtain a plurality of first target drive test images, and using an image-based search model to filter the first target drive test images and all the preset reference images in the configuration file to obtain a plurality of second target drive test images;
[0039] Specifically, the text-image model is used for autonomous driving data mining. The corresponding autonomous driving scenarios are mined according to the given text description, for example: "scene of driving in rainy days". In a 30-second autonomous driving scenario, some front camera pictures are sampled in each frame, and the text-image model is matched for each picture and the text description. If it meets the text description scene, the picture is considered to meet the mining conditions. The same operation is taken for all sampled pictures, and the results of all pictures are fused to determine whether the autonomous driving scene meets the text description. The image search model is used for mining data verification. The given example pictures and the sampled pictures that meet the mining results of the text-image model are input into the image search model for feature extraction and similarity matching.
[0040] In a specific implementation process, the step S102 uses a text-image model to filter and process all the image description texts in the target drive test data and the configuration file to obtain a plurality of first target drive test images, including: using a text-image model to process all the image description texts in the target drive test data and the configuration file to obtain a plurality of first matching degrees; determining that the first target drive test image is the drive test image whose first matching degree is greater than or equal to a first matching degree threshold;
[0041] Specifically, if the autonomous driving scene obtained by fusing all the image results is a scene that conforms to the text description, it is included in the mining result, and the text image matching score is used as its mining score. By setting the second matching degree threshold, the mining quality of the first target road test image is improved.
[0042] Determine the first matching degree threshold as a preset matching degree threshold corresponding to the range of the average value of all matching degrees obtained in the last training of the text-image model;
[0043] By setting a corresponding mapping relationship table, the method can set the first matching degree threshold more in line with the actual scenario.
[0044] The step S102 of using the image search model to screen the first target drive test image and all the preset reference images in the configuration file to obtain a plurality of second target drive test images includes: using the image search model to process the first target drive test image and all the preset reference images in the configuration file to obtain a plurality of second matching degrees; determining that the second target drive test image is the first target drive test image whose second matching degree is greater than or equal to a second matching degree threshold;
[0045] Specifically, after feature extraction and similarity matching, if the matching degree reaches a certain threshold, the autonomous driving scene is counted as a successful verification result, and the matching degree is used as its verification score. By setting a second matching degree threshold, the mining quality of the second target road test image can be improved.
[0046] A second matching degree threshold is determined to be a preset matching degree threshold corresponding to a range within which an average value of all first matching degrees obtained by processing the target drive test data and all the image description texts in the configuration file by the text-image model falls.
[0047] Specifically, by setting a corresponding mapping relationship table, the setting of the second matching degree threshold will be more in line with the actual scenario.
[0048] Step S103: determining a target mapping relationship as a mapping relationship between each of the second target drive test images and the corresponding image description text, and storing the target mapping relationship in a database.
[0049] Specifically, the successfully verified scenes are used as training data to train the text-image model and the image search model to improve their mining capabilities. At the same time, the data in the database can also be filtered out according to its text description to provide it to other autonomous driving models.
[0050] In one embodiment of the present application, after storing the above-mentioned target mapping relationship in a database, the above-mentioned method further includes: using the above-mentioned target mapping relationship to iteratively update the above-mentioned text-image model and the above-mentioned image search model respectively to obtain an optimized text-image model and an optimized image search model.
[0051] Specifically, the method enables the subsequent use of the optimized text-image model and the optimized image-based search model for data mining, thereby improving the quality and efficiency of subsequent data mining.
[0052] In the above steps, the target drive test data and all the image description texts in the configuration file are screened by using a text-image model to obtain a plurality of first target drive test images, and the first target drive test images and all the preset reference images in the configuration file are screened by using an image-based search model to obtain a plurality of second target drive test images, thereby achieving the purpose of double screening, thereby improving the quality and efficiency of data mining; realizing automated data mining and verification, and solving the problem that the existing solutions using manual and handwritten rule-based methods in the data mining process are still relatively inefficient.
[0053] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the data mining method of the present application will be described in detail below in conjunction with specific embodiments.
[0054] This embodiment relates to a specific data mining method, such as Figure 2 As shown, the following steps are included:
[0055] Obtain target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring;
[0056] Using a text-image model to filter and process all image description texts in the target drive test data and the configuration file to obtain a plurality of first target drive test images;
[0057] The image-based search model is used to screen the first target drive test image and all preset reference images in the configuration file to obtain a plurality of second target drive test images;
[0058] Determine a target mapping relationship as a mapping relationship between each second target drive test image and a corresponding image description text, and store the target mapping relationship in a database;
[0059] The target mapping relationship is used to iteratively update the text-image model and the image-searching model, respectively, to obtain an optimized text-image model and an optimized image-searching model.
[0060] The present application embodiment also provides a kind of data mining device, it should be noted that the data mining device of the present application embodiment can be used to execute the method for data mining provided by the present application embodiment. The device is used to implement the above-mentioned embodiment and preferred implementation mode, and the description has been made no further. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware is also possible and conceived.
[0061] The following is an introduction to the data mining device provided in the embodiments of the present application.
[0062] Figure 3 is a structural block diagram of a data mining device provided according to an embodiment of the present application. Figure 3 As shown, the device comprises:
[0063] An acquisition unit 31 is used to acquire target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring;
[0064] A first processing unit 32 is configured to filter the target drive test data and all the image description texts in the configuration file using a text-image model to obtain a plurality of first target drive test images, and filter the first target drive test images and all the preset reference images in the configuration file using an image-based search model to obtain a plurality of second target drive test images;
[0065] The second processing unit 33 is used to determine that the target mapping relationship is a mapping relationship between each of the second target drive test images and the corresponding image description text, and store the target mapping relationship in a database.
[0066] In the above-mentioned device, the present application provides a method, device, medium and vehicle system for data mining. The method adopts a text-image model to filter and process the above-mentioned target road test data and all the above-mentioned image description texts in the above-mentioned configuration file to obtain multiple first target road test images, and adopts an image-based search model to filter and process the above-mentioned first target road test images and all the above-mentioned preset reference images in the above-mentioned configuration file to obtain multiple second target road test images, thereby achieving the purpose of double screening, thereby improving the quality and efficiency of data mining; and realizing automated data mining and verification.
[0067] In one embodiment of the present application, the first processing unit includes a first processing module and a first determination module. The first processing module is used to process the target drive test data and all the image description texts in the configuration file using a text-image model to obtain multiple first matching degrees; the first determination module is used to determine that the first target drive test image is the drive test image whose first matching degree is greater than or equal to a first matching degree threshold.
[0068] In one embodiment of the present application, the first processing unit further includes a second processing module and a second determination module. The second processing module is used to process the first target drive test image and all the preset reference images in the configuration file using the image search model to obtain multiple second matching degrees; the second determination module is used to determine that the second target drive test image is the first target drive test image whose second matching degree is greater than or equal to a second matching degree threshold.
[0069] In one embodiment of the present application, the data mining device further comprises a first determination unit configured to determine the first matching threshold as a preset matching threshold corresponding to the range of the average values of all matching degrees obtained during the last training of the text-image model.
[0070] In one embodiment of the present application, the data mining device further comprises a second determination unit for determining the second matching degree threshold as a preset matching degree threshold corresponding to the range of the average value of all matching degrees obtained during the last training of the text-image model.
[0071] In one embodiment of the present application, the data mining device further includes a third processing unit. The third processing unit is used to iteratively update the text-image model and the image-based search model using the target mapping relationship to obtain an optimized text-image model and an optimized image-based search model.
[0072] In one embodiment of the present application, the first acquisition unit includes an acquisition module and a third processing module. The acquisition module is used to acquire initial drive test data, the initial drive test data is a drive test video; the third processing module is used to perform data segmentation processing on the drive test video to obtain a plurality of target drive test data.
[0073] The data mining device includes a processor and a memory. The first acquisition unit, the first processing unit, the second processing unit, the third processing unit, the first determination unit, and the second determination unit are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions. The modules are all located in the same processor; or, the modules are located in different processors in any combination.
[0074] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters can be adjusted to solve the problem that the existing solutions use manual and handwritten rule-based methods in the data mining process, and the efficiency is still relatively low.
[0075] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0076] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the data mining method, device, medium and vehicle system method.
[0077] Specifically, data mining methods include:
[0078] Step S101, obtaining target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring;
[0079] Specifically, for example, in a "driving scene in the rain", in a 30-second autonomous driving scene, some front camera images are sampled in each frame.
[0080] Step S102, using a text-image model to filter the target drive test data and all the image description texts in the configuration file to obtain a plurality of first target drive test images, and using an image-based search model to filter the first target drive test images and all the preset reference images in the configuration file to obtain a plurality of second target drive test images;
[0081] Specifically, the text-image model is used for autonomous driving data mining. The corresponding autonomous driving scenarios are mined according to the given text description, for example: "scene of driving in rainy days". In a 30-second autonomous driving scenario, some front camera pictures are sampled in each frame, and the text-image model is matched for each picture and the text description. If it meets the text description scene, the picture is considered to meet the mining conditions. The same operation is taken for all sampled pictures, and the results of all pictures are fused to determine whether the autonomous driving scene meets the text description. The image search model is used for mining data verification. The given example pictures and the sampled pictures that meet the mining results of the text-image model are input into the image search model for feature extraction and similarity matching.
[0082] Step S103, determining a target mapping relationship as a mapping relationship between each of the second target drive test images and the corresponding image description text, and storing the target mapping relationship in a database;
[0083] Specifically, the successfully verified scenes are used as training data to train the text-image model and the image search model to improve their mining capabilities. At the same time, the data in the database can also be filtered out according to its text description to provide it to other autonomous driving models.
[0084] Optionally, a text-image model is used to filter and process all the image description texts in the target drive test data and the configuration file to obtain multiple first target road test images, including: using a text-image model to process all the image description texts in the target drive test data and the configuration file to obtain multiple first matching degrees; determining that the first target road test image is the road test image whose first matching degree is greater than or equal to a first matching degree threshold.
[0085] Optionally, the first target road test image and all the preset reference images in the configuration file are screened and processed using the image search model to obtain a plurality of second target road test images, including: using the image search model to process the first target road test image and all the preset reference images in the configuration file to obtain a plurality of second matching degrees; and determining that the second target road test image is the first target road test image whose second matching degree is greater than or equal to a second matching degree threshold.
[0086] Optionally, the method further includes: determining the first matching degree threshold as a preset matching degree threshold corresponding to a range in which an average value of all matching degrees obtained during the last training of the text-image model lies.
[0087] Optionally, the method further includes: determining a preset matching degree threshold corresponding to a range where the second matching degree threshold is an average value of all first matching degrees obtained by processing the target drive test data and all the image description texts in the configuration file by the text-image model.
[0088] Optionally, after storing the above-mentioned target mapping relationship in the database, the above-mentioned method further includes: using the above-mentioned target mapping relationship to iteratively update the above-mentioned text-image model and the above-mentioned image search model respectively to obtain an optimized text-image model and an optimized image search model.
[0089] Optionally, obtaining the target drive test data includes: obtaining initial drive test data, where the initial drive test data is a drive test video; and performing data segmentation processing on the drive test video to obtain a plurality of the target drive test data.
[0090] An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein the data mining method is executed when the program is run.
[0091] An embodiment of the present invention provides a device, the device includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, at least the following steps are implemented: obtaining target road test data and a configuration file, the target road test data includes multiple road test images, the configuration file includes multiple image description texts and multiple preset reference images, and the road test images are images for road monitoring; using a text-image model to filter and process all the image description texts in the target road test data and the configuration file to obtain multiple first target road test images, and using a search-by-image model to filter and process all the preset reference images in the first target road test images and the configuration file to obtain multiple second target road test images; determining a target mapping relationship as a mapping relationship between each of the second target road test images and the corresponding image description text, and storing the target mapping relationship in a database. The device in this article may be a server, a PC, a PAD, a mobile phone, etc.
[0092] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with at least the following method steps: obtaining target drive test data and a configuration file, the target drive test data including multiple drive test images, the configuration file including multiple image description texts and multiple preset reference images, the drive test images being images for road monitoring; using a text-image model to filter and process all of the image description texts in the target drive test data and the configuration file to obtain multiple first target drive test images, and using an image-based search model to filter and process the first target drive test images and all of the preset reference images in the configuration file to obtain multiple second target drive test images; determining a target mapping relationship as a mapping relationship between each of the second target drive test images and the corresponding image description texts, and storing the target mapping relationship in a database.
[0093] The present application also provides a vehicle system, which includes: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing any of the above-mentioned data mining.
[0094] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.
[0095] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0096] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0097] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0099] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0100] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0101] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0102] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0103] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:
[0104] 1) The data mining method of the present application provides a data mining method, device, medium and vehicle system. The method uses a text-image model to filter and process the above-mentioned target road test data and all the above-mentioned image description texts in the above-mentioned configuration file to obtain multiple first target road test images, and uses an image-based search model to filter and process the above-mentioned first target road test images and all the above-mentioned preset reference images in the above-mentioned configuration file to obtain multiple second target road test images, thereby achieving the purpose of double screening, thereby improving the quality and efficiency of data mining; and realizing automated data mining and verification.
[0105] 2) The data mining device of the present application provides a data mining method, device, medium and vehicle system. The method uses a text-image model to filter and process the above-mentioned target road test data and all the above-mentioned image description texts in the above-mentioned configuration file to obtain multiple first target road test images, and uses an image-based search model to filter and process the above-mentioned first target road test images and all the above-mentioned preset reference images in the above-mentioned configuration file to obtain multiple second target road test images, thereby achieving the purpose of double screening, thereby improving the quality and efficiency of data mining; and realizing automated data mining and verification.
[0106] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A data mining method, characterized in that: include: Acquire target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring; Using a text-image model to filter all the image description texts in the target drive test data and the configuration file to obtain a plurality of first target drive test images, and using an image-based search model to filter all the preset reference images in the first target drive test image and the configuration file to obtain a plurality of second target drive test images; Determine a target mapping relationship as a mapping relationship between each of the second target drive test images and the corresponding image description text, and store the target mapping relationship in a database.
2. The method according to claim 1, characterized in that The target drive test data and all the image description texts in the configuration file are filtered and processed using a text-image model to obtain a plurality of first target drive test images, including: Using a text-image model to process the target drive test data and all the image description texts in the configuration file to obtain a plurality of first matching degrees; The first target drive test image is determined to be the drive test image whose first matching degree is greater than or equal to a first matching degree threshold.
3. The method according to claim 1, characterized in that The first target drive test image and all the preset reference images in the configuration file are screened and processed by using an image search model to obtain a plurality of second target drive test images, including: Using an image-based search model to process the first target drive test image and all the preset reference images in the configuration file to obtain a plurality of second matching degrees; The second target drive test image is determined to be the first target drive test image whose second matching degree is greater than or equal to a second matching degree threshold.
4. The method according to claim 1, characterized in that: The method further comprises: The first matching degree threshold is determined to be a preset matching degree threshold corresponding to a range within which an average value of all matching degrees obtained during the last training of the text-image model lies.
5. The method according to claim 1, characterized in that The method further comprises: A second matching degree threshold is determined to be a preset matching degree threshold corresponding to a range within which an average value of all first matching degrees obtained by processing the target drive test data and all the image description texts in the configuration file by the text-image model lies.
6. The method according to claim 1, characterized in that After storing the target mapping relationship in the database, the method further includes: The target mapping relationship is used to iteratively update the text-image model and the image-searching model respectively to obtain an optimized text-image model and an optimized image-searching model.
7. The method according to any one of claims 1 to 6, characterized in that Obtain target drive test data, including: Acquiring initial drive test data, where the initial drive test data is a drive test video; The drive test video is subjected to data segmentation processing to obtain a plurality of target drive test data.
8. A data mining device, characterized in that: include: An acquisition unit, configured to acquire target drive test data and a configuration file, wherein the target drive test data includes a plurality of drive test images, the configuration file includes a plurality of image description texts and a plurality of preset reference images, and the drive test image is an image for road monitoring; a first processing unit, configured to use a text-image model to filter the target drive test data and all the image description texts in the configuration file to obtain a plurality of first target drive test images, and use an image-based search model to filter the first target drive test images and all the preset reference images in the configuration file to obtain a plurality of second target drive test images; The second processing unit is used to determine that the target mapping relationship is a mapping relationship between each of the second target drive test images and the corresponding image description text, and store the target mapping relationship in a database.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the data mining method according to any one of claims 1 to 7.
10. A vehicle system, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a method for executing the data mining described in any one of claims 1 to 7.
Citation Information
Patent Citations
Retrieval optimization method and system for complex retrieval formula
CN103699574A
Information retrieval method and device based on picture matching and computer device
CN110532413A
Pedestrian re-identification method, device and equipment, and storage medium
CN113780172A
Image-based information search method and device, equipment and storage medium
CN114528424A
Data mining system, method and device based on graphics and text information combination
CN116595064A
