Artificial intelligence-based data processing method and device, computer device and medium

Abstract

The application belongs to the technical field of artificial intelligence, and relates to a data processing method based on artificial intelligence, which comprises the following steps: generating collection task queue data based on case reporting data entered by a user on a device; when a collection page is opened, collecting a preview video stream based on a camera of the device; frame extraction is performed on the preview video stream to obtain a target frame image; image analysis is performed on the target frame image based on a lightweight model to obtain an image analysis result; information generation is performed on the analysis result based on a rule mapping engine to obtain guide information; the guide information is fed back based on a feedback form, and picture data entered by the user after picture collection operation based on the guide information is received; fine detection is performed on the picture data based on a detection algorithm; and if the detection result passes the fine detection, the picture data is uploaded. The application can be applied to a vehicle claim business scene in the field of financial technology, can improve the compliance degree of uploaded image data, and can improve the processing efficiency of image uploading.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology and can be applied to the financial technology field, particularly to data processing methods, devices, computer equipment and storage media based on artificial intelligence. Background Technology

[0002] In the vehicle claims business within the financial insurance sector, with the continuous iteration and popularization of mobile internet technology, online self-service claims processing has gradually replaced the traditional offline operation mode and become the mainstream development trend in the industry due to its convenience and efficiency. In a typical online self-service claims process, when an applicant experiences a vehicle accident, they initiate a claim application through an application or mini-program provided by the insurance institution and submit on-site image data according to the system's instructions. Subsequently, the remote platform receives this image data and performs verification and vehicle damage assessment. However, in actual deployment and application, this conventional operation method relying on the "collection-transmission-cloud verification-return" model has exposed significant technical shortcomings and business bottlenecks. Specifically, accident investigation images have extremely high requirements for professional standards, such as the need to include panoramic license plate information and damage details from specific perspectives. However, ordinary car owners are not professional damage assessors and generally lack the ability to judge key elements such as shooting distance, lighting environment, and composition logic.

[0003] Current technologies mostly rely on displaying static images as references on the user interface. This simplistic approach is insufficient to effectively guide car owners in capturing images that meet professional standards. Consequently, the quality of uploaded images is extremely low, frequently resulting in issues such as incomplete captures, low clarity, and incorrect perspectives. Once an image fails to meet standards, the car owner must retake it, which not only reduces the overall efficiency of image uploading but may also directly impact the accuracy of subsequent vehicle damage assessments due to missing crucial evidence. Furthermore, it weakens the effectiveness of anti-fraud risk control.

[0004] Therefore, existing technologies suffer from low compliance and inefficiency in image uploading, and there is an urgent need to propose an effective solution to overcome these shortcomings. Summary of the Invention

[0005] The purpose of this application is to propose a data processing method, apparatus, computer equipment, and storage medium based on artificial intelligence, so as to solve the technical problems of low compliance and low efficiency in image uploading in the prior art.

[0006] Firstly, an artificial intelligence-based data processing method is provided, including: Receive the report data entered by the user on the device, and generate corresponding collection task queue data based on the report data; When it is detected that the acquisition page corresponding to the acquisition task queue data has been opened, the corresponding preview video stream is acquired based on the camera of the device. The preview video stream is subjected to frame extraction processing to obtain the corresponding target frame image; The target frame image is processed by image analysis based on a preset lightweight model to obtain the corresponding image analysis results. The analysis results are processed by a preset rule mapping engine to generate corresponding guidance information. The system processes the guidance information based on a preset feedback format and receives image data entered by the user after performing an image capture operation based on the guidance information. The image data is processed with a preset detection algorithm to obtain the corresponding detection results. If the detection result is that the image passes the refined detection, then the image data will be uploaded for processing.

[0007] Secondly, an artificial intelligence-based data processing device is provided, comprising: The first processing module is used to receive the report data entered by the user on the device and generate corresponding collection task queue data based on the report data. The acquisition module is used to acquire a corresponding preview video stream based on the camera of the device when it is detected that the acquisition page corresponding to the acquisition task queue data is opened; The extraction module is used to perform frame extraction processing on the preview video stream to obtain the corresponding target frame image; The analysis module is used to perform image analysis processing on the target frame image based on a preset lightweight model to obtain the corresponding image analysis results; The generation module is used to process the analysis results based on a preset rule mapping engine to generate corresponding guidance information. The second processing module is used to process the guidance information based on a preset feedback format, and to receive the image data entered by the user after performing an image acquisition operation based on the guidance information. The detection module is used to perform refined detection processing on the image data based on a preset detection algorithm to obtain the corresponding detection results; The upload module is used to upload the image data if the detection result is that the image passes the refined detection.

[0008] Thirdly, a computer device is provided, including 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 above-described artificial intelligence-based data processing method.

[0009] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the aforementioned artificial intelligence-based data processing method.

[0010] In the aforementioned scheme implemented by the data processing method, device, computer equipment, and storage medium based on artificial intelligence, the system first receives the report data entered by the user on the device and generates corresponding collection task queue data based on the report data. Then, when it is detected that the collection page corresponding to the collection task queue data is opened, the system captures the corresponding preview video stream based on the camera of the device. The preview video stream is then subjected to frame extraction processing to obtain the corresponding target frame image. Subsequently, the target frame image is subjected to image analysis processing based on a preset lightweight model to obtain the corresponding image analysis result. Subsequently, the analysis result is subjected to information generation processing based on a preset rule mapping engine to obtain the corresponding guidance information. The guidance information is further subjected to feedback processing based on a preset feedback form, and the system receives the image data entered by the user after performing an image collection operation based on the guidance information. In another step, the image data is subjected to refined detection processing based on a preset detection algorithm to obtain the corresponding detection result. If the detection result is that the refined detection is passed, the image data is uploaded. Based on the above automated processing flow, this application analyzes the preview video stream using a lightweight model and provides user feedback during the preview stage. This allows for timely detection and correction of problems during the acquisition process, improving the quality of the acquired data. This pre-emptive quality control method prevents users from acquiring unacceptable images, reducing subsequent re-acquisitions and improving acquisition efficiency. Furthermore, subsequent fine-grained detection algorithms are used to refine the image data acquired by users, ensuring that the data to be uploaded is valid and providing reliable input for subsequent evaluation and processing. Thus, this application forms a dual quality assurance system by combining a lightweight model with detection algorithms, effectively improving the compliance and accuracy of uploaded image data and reducing the number of image re-acquisitions, thereby increasing the processing efficiency of image uploads. Attached Figure Description

[0011] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is an exemplary system architecture diagram to which this application can be applied; Figure 2 This is a flowchart of an embodiment of the artificial intelligence-based data processing method according to this application; Figure 3 This is a schematic diagram of a structure of an embodiment of the artificial intelligence-based data processing apparatus according to this application; Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0014] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0015] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0016] like Figure 1 As shown, system architecture 100 may include terminal device 101, network 102, and server 103. Terminal device 101 may be a laptop 1011, tablet 1012, or mobile phone 1013. Network 102 is used as a medium to provide a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0017] Users can use terminal device 101 to interact with server 103 via network 102 to receive or send messages, etc. Various communication client applications can be installed on terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0018] Terminal device 101 can be various electronic devices with a display screen and support web browsing. In addition to laptops 1011, tablets 1012, or mobile phones 1013, terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer, and a desktop computer, etc.

[0019] Server 103 can be a server that provides various services, such as a backend server that provides support for the pages displayed on terminal device 101.

[0020] It should be noted that the artificial intelligence-based data processing method provided in the embodiments of this application is generally executed by a server / terminal device, and correspondingly, the artificial intelligence-based data processing device is generally set in the server / terminal device.

[0021] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0022] Continue to refer to Figure 2 This document illustrates a flowchart of an embodiment of the AI-based data processing method according to this application. The order of steps in the flowchart can be changed, and some steps can be omitted, depending on different needs. The AI-based data processing method provided in this application can be applied to any scenario requiring data processing, and therefore can be applied to products in these scenarios, such as data processing products in the financial and insurance fields. The AI-based data processing method includes the following steps: Step S201: Receive the report data entered by the user on the device, and generate corresponding collection task queue data based on the report data.

[0023] In this embodiment, the data processing method based on artificial intelligence runs on an electronic device (e.g., Figure 1The server / terminal device shown can acquire user-entered claim data via wired or wireless connection. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra-wideband) connections, and other currently known or future-developed wireless connection methods. The implementing entity of this application may specifically be a data processing system, which can be simply referred to as the system. This application can be applied to vehicle claims scenarios in the financial insurance field. Claim data is typically entered by users through the insurance company's online channels after an insurance-related event (such as a vehicle accident, property damage, etc.). These online channels include the insurance company's official website, mobile front-ends (such as mini-programs, apps), etc. After encountering an insurance incident, users can actively open the relevant application of the insurance company using their device, enter the claim page, and fill in and submit claim data. Claim data may include the time and location of the accident, basic vehicle information (such as vehicle model and license plate number), accident type (such as collision, scratch, natural disaster, etc.), and a general description of the vehicle damage.

[0024] The specific implementation process of generating the corresponding collection task queue data based on the reported data will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0025] Step S202: When it is detected that the acquisition page corresponding to the acquisition task queue data is opened, the corresponding preview video stream is acquired based on the camera of the device.

[0026] In this embodiment, during or after the user enters the report data, the user will prepare for evidence collection (image collection) according to the system prompts. At this time, the collection interface will open to provide evidence for subsequent claims. The collection interface will clearly show the list of tasks that need to be completed corresponding to the collection task queue data, as well as the specific requirements and related prompts for each task.

[0027] Additionally, at each task node corresponding to the aforementioned data collection task queue, when a user enters the collection interface and prepares to collect data, the system automatically invokes the device's camera function to begin capturing a preview video stream. At this time, the user's device screen will display the vehicle's current view in real time, serving as the raw data collection source.

[0028] Step S203: Perform frame extraction processing on the preview video stream to obtain the corresponding target frame image.

[0029] In this embodiment, the specific implementation process of performing frame extraction processing on the preview video stream to obtain the corresponding target frame image will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0030] Step S204: Perform image analysis processing on the target frame image based on a preset lightweight model to obtain the corresponding image analysis results.

[0031] In this embodiment, the aforementioned lightweight model (hereinafter referred to as the model) is a lightweight convolutional neural network (CNN) or other small deep learning model embedded in the device (such as embedded in a mobile front-end), which needs to be generated through pre-training and optimization. During the model training process, a large amount of labeled image data can be used to teach the model how to recognize and judge various features of the image. Furthermore, in order to improve the model's running efficiency on the device, pruning and quantization are performed on the model. Pruning refers to removing some unimportant neural connections in the model to reduce the model's complexity; quantization converts the floating-point parameters in the model into low-precision integer parameters, reducing the model's computational load and memory usage.

[0032] Next, the extracted target frame image is input into the prepared lightweight model. The model performs layer-by-layer feature extraction and analysis on the input target frame image. For example, in the convolutional layer, low-level features such as edges and textures are extracted through convolution operations between the convolution kernel and the image; in the pooling layer, the features are downsampled to reduce the amount of data while retaining the main features; in the fully connected layer, the extracted features are comprehensively analyzed, and the results of the judgment on the degree of blur, exposure status, target integrity, and acquisition interval (i.e., image analysis results) are output. The entire analysis process is completed within millisecond latency, enabling the prediction of image quality before the subject triggers the acquisition command.

[0033] Local image analysis based on a lightweight model is a crucial step in moving quality control forward. By performing real-time analysis of the extracted frames locally, problems in the image can be identified promptly, such as blurriness, overexposure or underexposure, or incomplete targets. This allows for guidance through subsequent interactive feedback to adjust shooting parameters before the user triggers the acquisition command, improving the quality of the acquired data and preventing the acquisition of invalid or low-quality images.

[0034] Step S205: Based on a preset rule mapping engine, the analysis results are processed to generate information and obtain corresponding guidance information.

[0035] In this embodiment, the specific implementation process of generating information from the analysis results based on the preset rule mapping engine to obtain the corresponding guidance information will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0036] Step S206: Process the guidance information based on a preset feedback format, and receive the image data entered by the user after performing an image acquisition operation based on the guidance information.

[0037] In this embodiment, the specific implementation process of processing the guidance information based on the preset feedback form will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0038] Additionally, after providing guidance to the user, the user can adjust shooting angle, distance, and other parameters according to the system's guidance and prompts in preview mode before triggering the capture action. On the device, this is typically accomplished by clicking the capture button. Upon receiving the capture command, the system saves one frame from the current preview video stream or several frames after a short buffer, generating a still image. This still image is the final image data used for the claim.

[0039] The data acquisition process is a crucial step in transforming the preview video stream into static data usable for claims. After the preceding preview guidance and adjustments, the acquired static images at this stage generally meet the requirements in terms of image quality and composition, providing a solid foundation for subsequent secondary verification and uploading.

[0040] Step S207: Perform refined detection processing on the image data based on a preset detection algorithm to obtain the corresponding detection results.

[0041] In this embodiment, the specific implementation process of performing refined detection processing on the image data based on the preset detection algorithm to obtain the corresponding detection results will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0042] Step S208: If the detection result is that the image data passes the refined detection, then the image data is uploaded.

[0043] In this embodiment, if the generated detection result is found to pass the refined detection, it indicates that the collected image data conforms to the specifications. The image data will then enter the upload queue and be sent to the cloud server, thus completing the image data upload process. Specifically, the system will select an appropriate time to upload the image data to the cloud server based on network conditions and server status. During the upload process, the system will compress the image data to reduce data transmission volume and improve upload speed. Simultaneously, the system will employ encryption technology to encrypt the uploaded image data, ensuring its security during transmission. For example, HTTPS protocol will be used for data transmission, and AES encryption will be applied to the image data.

[0044] Data uploading is a crucial step in transmitting locally collected, valid data to the cloud server. Because strict quality control has been implemented on the image data, all uploaded images are valid, significantly reducing the upload of invalid, blurry, or duplicate images, thus saving bandwidth and applicant data costs. Furthermore, in weak network environments, data compression and encryption improve data transmission success rates and security, preventing process interruptions due to data transmission failures.

[0045] This application first receives the report data entered by the user on the device and generates corresponding collection task queue data based on the report data. Then, when it is detected that the collection page corresponding to the collection task queue data is opened, it captures the corresponding preview video stream based on the device's camera. The preview video stream is then subjected to frame extraction processing to obtain the corresponding target frame image. Next, the target frame image is processed for image analysis based on a preset lightweight model to obtain the corresponding image analysis result. Subsequently, the analysis result is processed for information generation based on a preset rule mapping engine to obtain corresponding guidance information. Further, the guidance information is processed for feedback based on a preset feedback format, and image data entered by the user after performing image collection operations based on the guidance information is received. Then, the image data is processed for refined detection based on a preset detection algorithm to obtain the corresponding detection result. If the detection result is a pass for refined detection, the image data is uploaded. Based on the above automated processing flow, this application analyzes the preview video stream using a lightweight model and provides guidance feedback to the user during the preview stage, enabling timely detection and correction of problems during the collection process and improving the quality of the collected data. This proactive quality control approach prevents users from collecting substandard images, reducing the number of subsequent re-collections and improving collection efficiency. Furthermore, subsequent fine-grained detection algorithms are used to refine the image data collected by users, ensuring that the uploaded data is valid and providing reliable input for subsequent evaluation and processing. Thus, this application establishes a dual quality assurance system through the combined use of a lightweight model and detection algorithms, effectively improving the compliance and accuracy of uploaded image data and reducing the number of image re-collections, thereby increasing image upload processing efficiency.

[0046] In some optional implementations, step S201, which involves generating corresponding data collection task queue data based on the reported data, includes the following steps: The reported data is parsed and processed based on a preset data parsing strategy to obtain corresponding parsed data.

[0047] In this embodiment, in the business scenario of vehicle accidents, the reported data may include the time and location of the accident, basic vehicle information (such as vehicle model and license plate number), accident type (such as collision, scratch, natural disaster, etc.), and a general description of the vehicle damage. The system parses and structures this raw reported data, transforming it into a format that a computer can understand and process, thereby obtaining the corresponding parsed data.

[0048] Call the preset rule base.

[0049] In this embodiment, the system internally pre-defines a series of data collection rules based on different accident types and vehicle damage conditions, according to actual business needs. For example, for collision accidents, it may be necessary to collect photos of the front, rear, side impact areas, and overall vehicle appearance; for scratch accidents, it is necessary to focus on collecting close-up and long-range photos of the scratched areas. Furthermore, the data collection rules are stored in a pre-defined database to form a rule base.

[0050] Based on the rule matching algorithm, the target data collection task rules that match the parsed data are retrieved from the rule base.

[0051] In this embodiment, based on the parsed data after parsing, the system will use a rule matching algorithm to find the corresponding collection task rules from the constructed rule base.

[0052] Based on the target collection task rules, a collection task queue data corresponding to the reported data is generated.

[0053] In this embodiment, after determining the target data collection task rules, a matching data collection task queue (or simply data collection task queue) is generated according to a certain logical order. Taking a vehicle accident as an example, if the accident report data shows that the front of the vehicle is damaged, the system may generate a data collection task sequence such as "front of vehicle including license plate -> close-up of damage -> distant environment".

[0054] By generating targeted data collection task sequences based on different report scenarios, the system ensures that the collected data comprehensively and accurately reflects the information required for the claim, guaranteeing the integrity and standardization of the claim's evidence chain. Simultaneously, the structured data collection task sequences help the operator to collect data systematically and orderly, improving collection efficiency. Furthermore, the generated task queue is stored in the system's database and pushed to users in real time via mobile front-end. Users can clearly see the list of tasks to be completed on the data collection interface, along with the specific requirements and relevant prompts for each task.

[0055] Based on the above processing flow, this application uses a data parsing strategy to analyze the reported data to obtain parsed data. Then, using a rule matching algorithm, it queries the rule base to find target data collection task rules that match the parsed data. Subsequently, based on these target data collection task rules, it automatically and intelligently generates data collection task queues corresponding to the reported data. Generating data collection task queues provides a clear direction and objective for subsequent data collection operations. By automatically generating data collection task queues based on reported data, the system ensures that the collected photos completely and accurately present the vehicle damage, providing strong evidentiary support for subsequent claims. At the same time, a reasonable task sequence design also helps improve data collection efficiency and reduce user operational complexity.

[0056] In some optional implementations of this embodiment, step S203 includes the following steps: It invokes various preset frame skipping methods.

[0057] In this embodiment, to reduce data volume and computational load, the system performs frame extraction on the continuous preview video stream in a specific manner. The frame extraction method can include extracting one frame at fixed time intervals (e.g., every 0.5 seconds) (first frame extraction method), or intelligently extracting keyframes based on changes in the scene (second frame extraction method). For example, when the surrounding environment of the vehicle does not change significantly, the frame extraction interval can be appropriately extended; while when the vehicle moves or the scene changes significantly, the frame extraction interval is shortened to ensure that the extracted frame images accurately reflect the vehicle's state.

[0058] Select the target frame extraction method from all the frame extraction methods described.

[0059] In this embodiment, a method can be selected from the above frame extraction methods according to actual business needs to serve as the corresponding target frame extraction method.

[0060] The preview video stream is processed by extracting frames based on the target frame extraction method to obtain the corresponding frame images.

[0061] In this embodiment, the above-mentioned preview video stream can be processed by frame extraction according to the frame extraction implementation method corresponding to the selected target frame extraction method, and the extracted frame image is used as the corresponding target frame image.

[0062] The frame image is used as the target frame image.

[0063] This application utilizes a pre-defined set of frame extraction methods; then selects a target frame extraction method from all these methods; subsequently, it performs frame extraction processing on the preview video stream based on the target frame extraction method to obtain corresponding frame images; and finally, it uses these frame images as the target frame images. Based on this processing flow, this application effectively improves the intelligence of target frame image generation by selecting a target frame extraction method from multiple methods, then performing frame extraction processing on the preview video stream based on the target frame extraction method, and using the obtained frame images as the corresponding target frame images. Furthermore, the subsequent use of a lightweight model for image analysis processing of the target frame images effectively reduces the amount of data and computation required for image analysis, thereby improving the efficiency of image analysis.

[0064] In some alternative implementations, step S205 includes the following steps: Call the preset rule mapping engine.

[0065] In this embodiment, a rule mapping engine was pre-built based on actual business needs and data acquisition specifications. The rule mapping engine defines the mapping relationships between various multi-source data combinations (including multi-source sensor data and analysis results of images) and natural language instructions (or prompts / instructions). For example, in cases of severe image shaking, a prompt instruction of "Insufficient image clarity, please keep the device stable" is triggered when the accelerometer data exceeds a certain threshold and the confidence level of the local algorithm's image sharpness analysis is below a certain value. In cases of abnormal target subject proportions, a prompt instruction of "Too close / too far, please adjust the spacing" is triggered when the bounding box coordinate comparison results show that the target subject proportion exceeds a preset range. In cases of strong light spots obscuring key areas, a prompt instruction of "Severe reflection, please adjust the viewing angle" is triggered when the local algorithm's image exposure analysis determines that the image is overexposed and the key area is covered by strong light spots.

[0066] Acquire multi-source sensing data corresponding to the target frame image.

[0067] In this embodiment, the specific implementation process of acquiring the multi-source sensing data corresponding to the target frame image will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0068] The image analysis results are matched with the multi-source perception data based on the rule mapping engine to obtain the corresponding target prompt information.

[0069] In this embodiment, during the acquisition process, the system acquires multi-source sensing data and algorithm analysis results corresponding to the target frame image in real time, and inputs this data into the rule mapping engine. Then, the rule mapping engine matches and judges this data according to predefined rules, dynamically activating corresponding prompts and serving as guidance information.

[0070] The target prompt information is used as the guidance information.

[0071] This application utilizes a pre-defined rule mapping engine to acquire multi-source sensor data corresponding to the target frame image. Then, based on the rule mapping engine, it matches the image analysis results with the multi-source sensor data to obtain corresponding target prompt information. This target prompt information is subsequently used as guidance information. Based on this processing flow, this application, by using a rule mapping engine, transforms the multi-source sensor data and image analysis results corresponding to the target frame image into matching target prompt information. This allows the operator to easily understand and adjust according to the prompts, thereby enabling timely and accurate feedback based on the real-time acquired target frame image data, ultimately improving data acquisition efficiency and data quality.

[0072] In some optional implementations, acquiring the multi-source sensing data corresponding to the target frame image includes the following steps: Obtain the feature confidence threshold corresponding to the target frame image.

[0073] In this embodiment, when using a lightweight model (hereinafter referred to as the model) to analyze the target frame image, the lightweight model outputs the confidence level of various feature judgments. For example, for the judgment of the degree of image blur, the model will give a confidence value, indicating the reliability of the model's judgment. Furthermore, the system will collect these confidence threshold data, i.e., feature confidence thresholds, for subsequent rule judgments.

[0074] The target frame image is compared with the bounding box coordinates based on the preset target detection algorithm to obtain the corresponding bounding box coordinate comparison results.

[0075] In this embodiment, during image analysis, a target detection algorithm can be used to identify the target subject in the frame corresponding to the target frame image, and a bounding box is drawn for each target subject. Then, the system collects the coordinate information of these bounding boxes, and by comparing the bounding box coordinates between different frames or with a preset standard, determines the position and proportion of the target subject in the frame, and generates corresponding bounding box coordinate comparison results. For example, in the scenario of capturing vehicle license plates, a reasonable proportion range for the license plate in the frame is preset, and by comparing the actual captured license plate bounding box proportion, it is determined whether it meets the requirements.

[0076] Acquire the acceleration sensor data corresponding to the device.

[0077] In this embodiment, the device is equipped with an accelerometer, and the system can acquire data from the accelerometer through the device's sensor interface. The accelerometer data reflects the device's acceleration changes in three dimensions, and analyzing this data can help determine the device's vibration status. For example, if the device's acceleration changes drastically in a certain direction, it indicates that the device may be vibrating.

[0078] The feature confidence threshold, the bounding box coordinate comparison results, and the acceleration sensor data are integrated to obtain the corresponding integrated data.

[0079] In this embodiment, the acquired feature confidence threshold, bounding box coordinate comparison results, and acceleration sensor data can be integrated and processed, and the resulting integrated data can be used as the corresponding multi-source sensing data.

[0080] The integrated data is used as the multi-source sensing data.

[0081] This application obtains a feature confidence threshold corresponding to the target frame image; simultaneously, it performs bounding box coordinate comparison processing on the target frame image based on a preset target detection algorithm to obtain the corresponding bounding box coordinate comparison result; and it obtains accelerometer data corresponding to the device; then, it integrates the feature confidence threshold, the bounding box coordinate comparison result, and the accelerometer data to obtain the corresponding integrated data; subsequently, the integrated data is used as the multi-source sensing data. Based on the above processing flow, this application can obtain information related to the acquisition process from multiple perspectives by performing multi-source data acquisition, enriching the data dimensions. Since a single data source may not be able to fully and accurately reflect various situations during the acquisition process, this application, by combining multi-source sensing data such as feature confidence threshold, bounding box coordinate comparison result, and accelerometer data related to the target frame image, can more comprehensively and accurately understand information such as image quality, target position, and device status during the acquisition process, which is beneficial for providing more sufficient and accurate data basis for subsequent rule mapping engine processing.

[0082] In some optional implementations of this embodiment, the feedback form includes text feedback or voice feedback; step S206 includes the following steps: Determine whether the feedback format is text feedback.

[0083] In this embodiment, the feedback format may include text feedback or voice feedback. Whether a feedback format is text feedback can be determined by performing content analysis on the feedback format.

[0084] If the feedback is in the form of text feedback, the guidance information will be displayed based on a preset preview layer.

[0085] In this embodiment, if the feedback is detected to be in text form, a text prompt will be used to provide feedback to the user. Specifically, the obtained guidance information will be fed back to the user by overlaying text onto the preview layer in real time. For example, on the device's camera preview interface, guidance information may be displayed in the form of a pop-up window or subtitles, such as "Image clarity is insufficient, please keep the device stable," "Distance too close / far, please adjust the spacing," "Severe reflection, please adjust the viewing angle," etc.

[0086] If the feedback is in the form of voice feedback, the guidance information is converted to obtain the corresponding voice prompt.

[0087] In this embodiment, if the feedback is detected to be in the form of voice feedback, then voice feedback will be used to provide feedback to the user. Specifically, speech synthesis technology is used to convert the obtained guidance information into voice prompts.

[0088] The voice prompts are played based on a preset playback strategy.

[0089] In this embodiment, the system can invoke the device's voice playback function to play prompts to the user in real time. For example, when the user is holding the device to collect data, voice prompts can more intuitively remind the user to make adjustments.

[0090] This application determines whether the feedback is in text format. If the feedback is text, the guidance information is displayed based on a preset preview layer. If the feedback is voice, the guidance information is converted to obtain a corresponding voice prompt. The voice prompt is then played based on a preset playback strategy. Based on this process, this application provides guidance information to the user through text or voice prompts, allowing the user to promptly understand any problems encountered during the data collection process and make standardized corrections according to the prompts. This interaction mode lowers the cognitive threshold for non-professionals, enabling users to complete data collection tasks more easily and accurately, thereby improving the quality and standardization of the collected data.

[0091] In some optional implementations of this embodiment, step S207 includes the following steps: The image data is subjected to clarity detection based on the detection algorithm to obtain the corresponding first detection data.

[0092] In this embodiment, the detection algorithm described above can be a locally built-in algorithm with image detection capabilities (such as a more accurate image quality assessment algorithm), or the detection algorithm can directly adopt the aforementioned lightweight model. Specifically, regarding image clarity, the detection algorithm can be used to determine whether the user-collected image data has issues such as blurriness or noise, and generate corresponding first detection data.

[0093] The image data is subjected to target integrity detection to obtain the corresponding second detection data.

[0094] In this embodiment, to determine the integrity of the target, a detection algorithm can be used to determine whether the damaged parts of the vehicle are fully displayed in the image data collected by the user, and corresponding second detection data can be generated.

[0095] Exposure detection is performed on the image data to obtain the corresponding third detection data.

[0096] In this embodiment, for the exposure status, a detection algorithm can be used to detect and determine whether the image data collected by the user is too bright or too dark, resulting in loss of detail, and generate corresponding third detection data.

[0097] The first detection data, the second detection data, and the third detection data are integrated and analyzed to generate corresponding detection results.

[0098] In this embodiment, the image data is comprehensively evaluated and judged based on the detection data (first detection data, second detection data, and third detection data) obtained from the above-mentioned detection, according to multiple preset evaluation indicators or damage assessment standards. If the image data fails to meet the requirements in any evaluation indicator, the system will consider the image data unqualified and determine that the image data has not passed the refined detection, thereby generating a detection result of failing the refined detection. Only when the image data meets the requirements in all evaluation indicators will it be determined that the image data has passed the refined detection, thereby generating a detection result of passing the refined detection.

[0099] Furthermore, if image data fails the fine-grained inspection, the system will forcefully display a re-acquisition suggestion, informing the user that the currently acquired image data has a problem and needs to be re-acquired. The re-acquisition suggestion will clearly point out the problem, such as "Image clarity is insufficient, please retake the photo" or "Target is incomplete, please adjust the shooting angle and re-acquire." In this way, the system will prevent unqualified images from entering the upload queue, ensuring that only images that meet the quality requirements are uploaded to the cloud. This image processing method effectively improves the quality and usability of acquired data and reduces the transmission and processing of invalid data.

[0100] This application performs sharpness detection on the image data based on the aforementioned detection algorithm to obtain corresponding first detection data; simultaneously, it performs target integrity detection on the image data to obtain corresponding second detection data; and it performs exposure detection on the image data to obtain corresponding third detection data. Subsequently, the first, second, and third detection data are integrated and analyzed to generate corresponding detection results. Based on the above processing flow, this application achieves accurate and refined detection of image data from multiple dimensions by using a detection algorithm to perform sharpness detection, target integrity detection, and exposure detection, ensuring the accuracy of the generated detection results. This ensures that the uploaded image data is valid and provides reliable data input for subsequent processing. Furthermore, the refined detection and image analysis based on a lightweight model work together to form a dual quality assurance system, further improving the accuracy and reliability of the collected data.

[0101] In some alternative implementations, the user information obtained is subject to user consent and complies with relevant laws and policies.

[0102] Furthermore, any software tools or components not belonging to our company that appear in the embodiments of this application are merely illustrative examples and do not represent actual use.

[0103] In addition, this application has the following significant technological benefits and business value: 1. By moving compliance verification forward, the first-pass rate of image acquisition is significantly improved.

[0104] Existing methods often employ a post-processing quality control model of "acquisition-upload-cloud review," resulting in operators only receiving results after the action is completed, leading to a lengthy feedback loop. This application addresses this by using a simplified on-device model to perform real-time analysis of the camera preview video stream, intervening during the framing stage before the operator triggers the acquisition command. This "what you see is what you get" mechanism can simultaneously identify defects such as blurriness, overexposure, and missing targets, preventing the generation of inferior images. By shifting quality control from "post-event correction" to "pre-event prevention," it effectively avoids repeated rejections and re-acquisitions caused by non-compliant data, significantly improving the first-pass yield of image acquisition.

[0105] 2. Reduce the difficulty of operations for non-professionals and achieve standardization of operations.

[0106] To address the pain points of ordinary car owners lacking professional knowledge in damage assessment and struggling to master data collection standards, this application introduces an adaptive rule engine based on multi-source perception. The system instantly maps complex algorithmic indicators (such as IoU value and fuzziness score) into easily understandable natural language instructions (such as "Please take a step back" or "Please stabilize your device"), and provides visual guidance on the screen. This interactive assistance is equivalent to providing the applicant with a "virtual inspection expert" who can guide them to operate according to standardized composition, spacing, and perspective, thereby ensuring that the claim evidence obtained by non-professionals also possesses a high standard of usability.

[0107] 3. Improve transmission bandwidth and computing power utilization to enhance system robustness in weak network environments.

[0108] This application leverages edge computing capabilities to perform the vast majority of image quality assessment locally, allowing only compliant images to enter the upload queue. This mechanism yields dual benefits: at the network level, it significantly reduces the upload of invalid, blurry, or duplicate images, saving the applicant's bandwidth costs, and simultaneously resolving the issue of process interruptions caused by large-volume data transmission failures in weak network scenarios such as underground parking garages and highway sections; at the computing power level, it filters massive amounts of low-quality data, significantly reducing the concurrent processing pressure and storage overhead of cloud servers, allowing central computing power to focus more on high-value loss assessment and anti-fraud calculations.

[0109] 4. Reduce the time limit for claim assessment and support a fully automated closed loop process.

[0110] Through the task-oriented data collection process and dual verification mechanism of this application, it is ensured that every frame submitted to the remote end is clear, complete, and meets the "valid data" standards for damage assessment. This high-quality structured data input eliminates the need for manual pre-screening and cleaning, providing an ideal input source for the cloud-based "AI automatic damage assessment algorithm," thereby significantly improving the success rate and accuracy of automatic damage assessment. Ultimately, this effectively shortens the overall cycle from reporting a claim to compensation, improving the service efficiency and customer satisfaction of insurance institutions.

[0111] It should be understood that the sequence number of each step in the above embodiments does not imply 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 embodiments of the present invention.

[0112] It should be emphasized that, to further ensure the privacy and security of the aforementioned image data, the image data can also be stored in a blockchain node.

[0113] The blockchain referred to in this application is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0114] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results. Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0115] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These 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 methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0116] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0117] Further reference Figure 3 As a response to the above Figure 2 To implement the method shown, this application provides an embodiment of a data processing device based on artificial intelligence, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0118] like Figure 3 As shown, the artificial intelligence-based data processing device 300 described in this embodiment includes: a first processing module 301, a data acquisition module 302, an extraction module 303, an analysis module 304, a generation module 305, a second processing module 306, a detection module 307, and an upload module 308. Wherein: The first processing module 301 is used to receive the report data entered by the user on the device and generate corresponding collection task queue data based on the report data. The acquisition module 302 is used to acquire a corresponding preview video stream based on the camera of the device when it is detected that the acquisition page corresponding to the acquisition task queue data is opened; The extraction module 303 is used to perform frame extraction processing on the preview video stream to obtain the corresponding target frame image; The analysis module 304 is used to perform image analysis processing on the target frame image based on a preset lightweight model to obtain the corresponding image analysis results; The generation module 305 is used to perform information generation processing on the analysis results based on a preset rule mapping engine to obtain corresponding guidance information; The second processing module 306 is used to process the guidance information based on a preset feedback format, and to receive the image data entered by the user after performing an image acquisition operation based on the guidance information. The detection module 307 is used to perform refined detection processing on the image data based on a preset detection algorithm to obtain the corresponding detection results; The upload module 308 is used to upload the image data if the detection result is that the image passes the refined detection.

[0119] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here.

[0120] In some optional implementations of this embodiment, the first processing module 301 includes: The parsing submodule is used to parse and process the reported data based on a preset data parsing strategy to obtain the corresponding parsed data; The first submodule is used to invoke the preset rule base; The query submodule is used to query the rule base and find the target collection task rules that match the parsed data based on the rule matching algorithm. The generation submodule is used to generate collection task queue data corresponding to the reported data based on the target collection task rules.

[0121] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here.

[0122] In some optional implementations of this embodiment, the extraction module 303 includes: The second submodule is used to invoke various preset frame extraction methods; The filtering submodule is used to filter out the target frame extraction method from all the frame extraction methods. The processing submodule is used to perform frame extraction processing on the preview video stream based on the target frame extraction method to obtain the corresponding frame image; The first determining submodule is used to use the frame image as the target frame image.

[0123] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here.

[0124] In some optional implementations of this embodiment, the generation module 305 includes: The third submodule is used to invoke the preset rule mapping engine; The acquisition submodule is used to acquire multi-source sensing data corresponding to the target frame image; The matching submodule is used to match the image analysis results with the multi-source perception data based on the rule mapping engine to obtain the corresponding target prompt information; The second determining submodule is used to use the target prompt information as the guidance information.

[0125] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here.

[0126] In some optional implementations of this embodiment, the acquisition submodule includes: The first acquisition unit is used to acquire the feature confidence threshold corresponding to the target frame image; The processing unit is used to perform bounding box coordinate comparison processing on the target frame image based on a preset target detection algorithm to obtain the corresponding bounding box coordinate comparison result. The second acquisition unit is used to acquire acceleration sensor data corresponding to the device; An integration unit is used to integrate the feature confidence threshold, the bounding box coordinate comparison result, and the acceleration sensor data to obtain corresponding integrated data. A determining unit is used to use the integrated data as the multi-source sensing data.

[0127] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here. In some optional implementations of this embodiment, the second processing module 306 includes: The judgment submodule is used to determine whether the feedback format is text feedback. The display submodule is used to display the guidance information based on a preset preview layer if the feedback is in the form of text feedback. The conversion submodule is used to convert the guidance information to obtain the corresponding voice prompt if the feedback is in the form of voice feedback. The playback submodule is used to process the voice prompts based on a preset playback strategy.

[0128] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here.

[0129] In some optional implementations of this embodiment, the detection module 307 includes: The first detection submodule is used to perform clarity detection on the image data based on the detection algorithm to obtain corresponding first detection data; The second detection submodule is used to perform target integrity detection on the image data to obtain corresponding second detection data; The third detection submodule is used to perform exposure detection on the image data to obtain the corresponding third detection data; The analysis submodule is used to integrate and analyze the first detection data, the second detection data, and the third detection data to generate corresponding detection results.

[0130] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based data processing method in the aforementioned embodiments, and will not be repeated here. To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed]. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0131] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected via a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing 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 (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0132] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0133] The memory 41 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), 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, magnetic disk, optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as the hard disk or memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Of course, the memory 41 may also include both the internal storage unit and its external storage device of the computer device 4. In this embodiment, the memory 41 is typically used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions for data processing methods based on artificial intelligence. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or will be output.

[0134] In some embodiments, the processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions stored in the memory 41 or to process data, for example, to execute computer-readable instructions of the artificial intelligence-based data processing method.

[0135] The network interface 43 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 4 and other electronic devices.

[0136] Compared with the prior art, the embodiments of this application have the following beneficial effects: In this embodiment, the application analyzes the preview video stream using a lightweight model and provides user guidance and feedback during the preview stage. This allows for timely detection and correction of problems during the acquisition process, improving the quality of the acquired data. This pre-emptive quality control method prevents users from acquiring unacceptable images, reducing subsequent re-acquisitions and improving acquisition efficiency. Furthermore, subsequent fine-grained detection algorithms are used to refine the image data acquired by the user, ensuring that the data to be uploaded is valid, thus providing reliable input for subsequent evaluation and processing. Therefore, this application forms a dual quality assurance system by combining a lightweight model and detection algorithms, effectively improving the compliance and accuracy of uploaded image data and reducing the number of image re-acquisitions, thereby improving image upload processing efficiency.

[0137] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based data processing method described above.

[0138] Compared with the prior art, the embodiments of this application have the following main advantages: In this embodiment, the application analyzes the preview video stream using a lightweight model and provides user guidance and feedback during the preview stage. This allows for timely detection and correction of problems during the acquisition process, improving the quality of the acquired data. This pre-emptive quality control method prevents users from acquiring unacceptable images, reducing subsequent re-acquisitions and improving acquisition efficiency. Furthermore, subsequent fine-grained detection algorithms are used to refine the image data acquired by the user, ensuring that the data to be uploaded is valid, thus providing reliable input for subsequent evaluation and processing. Therefore, this application forms a dual quality assurance system by combining a lightweight model and detection algorithms, effectively improving the compliance and accuracy of uploaded image data and reducing the number of image re-acquisitions, thereby improving image upload processing efficiency.

[0139] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0140] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A data processing method based on artificial intelligence, characterized in that, Includes the following steps: Receive the report data entered by the user on the device, and generate corresponding collection task queue data based on the report data; When it is detected that the acquisition page corresponding to the acquisition task queue data has been opened, the corresponding preview video stream is acquired based on the camera of the device. The preview video stream is subjected to frame extraction processing to obtain the corresponding target frame image; The target frame image is processed by image analysis based on a preset lightweight model to obtain the corresponding image analysis results. The analysis results are processed by a preset rule mapping engine to generate corresponding guidance information. The system processes the guidance information based on a preset feedback format and receives image data entered by the user after performing an image capture operation based on the guidance information. The image data is processed with a preset detection algorithm to obtain the corresponding detection results. If the detection result is that the image passes the refined detection, then the image data will be uploaded for processing.

2. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of generating corresponding data collection task queue data based on the reported data specifically includes: The reported data is parsed and processed based on a preset data parsing strategy to obtain corresponding parsed data; Call the preset rule base; Based on the rule matching algorithm, the target data collection task rules that match the parsed data are retrieved from the rule base; Based on the target collection task rules, a collection task queue data corresponding to the reported data is generated.

3. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of performing frame extraction processing on the preview video stream to obtain the corresponding target frame image specifically includes: Invokes multiple preset frame skipping methods; Select the target frame extraction method from all the frame extraction methods described; The preview video stream is processed by extracting frames based on the target frame extraction method to obtain the corresponding frame images. The frame image is used as the target frame image.

4. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of processing the analysis results based on a preset rule mapping engine to obtain corresponding guidance information specifically includes: Invoke the preset rule mapping engine; Acquire multi-source sensing data corresponding to the target frame image; Based on the rule mapping engine, the image analysis results are matched with the multi-source perception data to obtain the corresponding target prompt information; The target prompt information is used as the guidance information.

5. The data processing method based on artificial intelligence according to claim 4, characterized in that, The step of acquiring multi-source sensing data corresponding to the target frame image specifically includes: Obtain the feature confidence threshold corresponding to the target frame image; The target frame image is compared with the bounding box coordinates based on the preset target detection algorithm to obtain the corresponding bounding box coordinate comparison result. Acquire acceleration sensor data corresponding to the device; The feature confidence threshold, the bounding box coordinate comparison result, and the acceleration sensor data are integrated to obtain the corresponding integrated data. The integrated data is used as the multi-source sensing data.

6. The data processing method based on artificial intelligence according to claim 1, characterized in that, The feedback format includes text feedback or voice feedback; the step of processing the guidance information based on the preset feedback format specifically includes: Determine whether the feedback format is text feedback. If the feedback is in the form of text feedback, the guidance information will be displayed based on a preset preview layer. If the feedback is in the form of voice feedback, the guidance information is converted to obtain the corresponding voice prompt. The voice prompts are played based on a preset playback strategy.

7. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of performing refined detection processing on the image data based on a preset detection algorithm to obtain the corresponding detection results specifically includes: Based on the detection algorithm, the image data is subjected to clarity detection to obtain the corresponding first detection data; Perform target integrity detection on the image data to obtain the corresponding second detection data; Exposure detection is performed on the image data to obtain the corresponding third detection data; The first detection data, the second detection data, and the third detection data are integrated and analyzed to generate corresponding detection results.

8. A data processing device based on artificial intelligence, characterized in that, include: The first processing module is used to receive the report data entered by the user on the device and generate corresponding collection task queue data based on the report data. The acquisition module is used to acquire a corresponding preview video stream based on the camera of the device when it is detected that the acquisition page corresponding to the acquisition task queue data is opened; The extraction module is used to perform frame extraction processing on the preview video stream to obtain the corresponding target frame image; The analysis module is used to perform image analysis processing on the target frame image based on a preset lightweight model to obtain the corresponding image analysis results; The generation module is used to process the analysis results based on a preset rule mapping engine to generate corresponding guidance information. The second processing module is used to process the guidance information based on a preset feedback format, and to receive the image data entered by the user after performing an image acquisition operation based on the guidance information. The detection module is used to perform refined detection processing on the image data based on a preset detection algorithm to obtain the corresponding detection results; The upload module is used to upload the image data if the detection result is that the image passes the refined detection.

9. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor executes the computer-readable instructions to implement the steps of the data processing method based on artificial intelligence as described in any one of claims 1 to 7.

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 data processing method based on artificial intelligence as described in any one of claims 1 to 7.

Images