Media terminal publication task management method and device, electronic equipment and medium

Abstract

The application relates to the technical field of publication management, and provides a media terminal publication task management method and device, an electronic device and a medium. A three-dimensional topological model is generated by performing space model construction on a plane data set of a target building and a media terminal label data set, an optimal path sequence is generated by combining current position information of a mobile terminal and acquired task instructions for path planning, the mobile terminal is verified in position by combining a calibrated position coordinate after calibration, visual data sent by the mobile terminal after position verification is acquired to verify whether a construction personnel is in position, task guide data is generated according to the visual data verified by multiple modes and the task instructions, when operation logs are acquired, the operation logs are traced and analyzed by an abnormal quality inspection model according to the task guide data, and a quality inspection traceability report is generated. Efficient management of a publication task is realized by three-dimensional modeling, dynamic path planning, position verification and visual verification.

Description

Technical Field

[0001] This application relates to the field of uploading management technology, and in particular to a method, apparatus, electronic device, and medium for managing uploading tasks on a media terminal. Background Technology

[0002] The media terminal uploading process mainly includes three steps: uploading statistics, uploading arrangement, and result sorting. Uploading statistics involves counting and dividing the images, that is, counting and classifying the number and distribution of advertising images. Uploading arrangement involves removing the old images from the media frame and replacing them with new advertising images. Successful uploading is photographed, and unsuccessful uploading is marked. Result sorting involves classifying and organizing the reasons for uploading failures and the photos.

[0003] In existing methods of managing print submission tasks, submission statistics are completed by managers through a CMIS system, which integrates various data within the enterprise, providing a unified data management platform. Submission scheduling is handled by engineering personnel, who are responsible for the specific print replacement work and record any submission failures. However, the current management method has several drawbacks. First, the submitted photos can only be guaranteed to have been taken in a specific building complex, without a one-to-one correspondence with specific locations. This means that checking the print replacement results can only be done through spot checks or patrols, resulting in insufficient actual coverage and the risk of incorrect or missed prints. Second, photo sorting still relies on manual labor, which not only increases workload but also reduces efficiency. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, electronic device, and medium for managing uploading tasks on a media terminal, in order to solve the problems of low efficiency and quality in uploading work on media terminals.

[0005] The first aspect of this application provides a method for managing uploading tasks on a media terminal, the method comprising:

[0006] The target building floor plan dataset and the media terminal tag dataset in the target building are obtained according to the obtained task instructions. A three-dimensional spatial model is constructed based on the target building floor plan dataset and the media terminal tag dataset to obtain a three-dimensional topology model.

[0007] The current location information of the mobile terminal is obtained, and dynamic path planning is performed based on the current location information, the task instructions, and the three-dimensional topology model to obtain the optimal path sequence.

[0008] When the spatial location verification dataset sent by the mobile terminal is obtained, multimodal location calibration processing is performed on the spatial location verification dataset to obtain the calibration location coordinates;

[0009] The media terminal coordinates are obtained based on the calibration location coordinates and the optimal path sequence, and the location of the mobile terminal is verified based on the media terminal coordinates and the calibration location coordinates.

[0010] When the mobile terminal passes the location verification, the visual data of the media terminal sent by the mobile terminal is obtained, and the visual data is subjected to multimodal visual verification.

[0011] When the visual data passes the multimodal visual verification, the target task sub-instruction is obtained from the task instruction according to the media terminal coordinates, and the task guidance data of the media terminal is generated according to the target task sub-instruction and the visual data.

[0012] When the operation log sent by the mobile terminal is obtained, a multi-dimensional source tracing analysis is performed on the operation log based on the task guidance data and the preset abnormal quality inspection model to obtain a quality inspection source tracing report.

[0013] In an optional implementation, the step of constructing a three-dimensional spatial model based on the target building plan dataset and the media terminal tag dataset to obtain a three-dimensional topological model includes:

[0014] Perform planar graphic parsing processing on the target building planar dataset to obtain the advertising space metadata table within the target building;

[0015] Perform UTM coordinate transformation on the two-dimensional geographic coordinates in the ad slot metadata table to obtain a three-dimensional coordinate dataset for the ad slot;

[0016] The ad space 3D coordinate dataset is subjected to Delaunay triangulation to obtain the adjacency matrix between ad spaces;

[0017] The media terminal tag dataset is parsed to obtain the associated dataset between tag IDs and ad slot numbers;

[0018] The associated dataset is matched with the three-dimensional coordinate dataset of the ad slot through a preset tag mapping model to obtain the mapping relationship data between media terminal tags and three-dimensional space model;

[0019] A three-dimensional spatial model is constructed based on the ad space 3D coordinate dataset, the adjacency matrix, and the mapping relationship data to obtain a 3D topological model.

[0020] In an optional implementation, the step of performing dynamic path planning based on the current location information, the task instructions, and the three-dimensional topology model to obtain the optimal path sequence includes:

[0021] The current location information is calibrated to obtain the initial coordinates of the mobile terminal;

[0022] The task sub-instructions are sorted according to the task priority in the task instruction to obtain the task execution queue;

[0023] Path optimization is performed based on the initial coordinates, the adjacency matrix in the 3D topology model, and the task execution queue to obtain the optimal path sequence.

[0024] In an optional implementation, performing multimodal location calibration processing on the spatial location verification dataset to obtain calibration location coordinates includes:

[0025] The NFC tag is decoded to obtain the NFC tag ID, and the target coordinates are obtained by matching the NFC tag ID with the mapping relationship data in the three-dimensional topology model.

[0026] The IMU data is integrated with accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal;

[0027] The environmental video data is processed using the ORB-SLAM2 algorithm to obtain the three-dimensional point cloud data of the mobile terminal.

[0028] The target coordinates, the IMU displacement change, and the three-dimensional point cloud data are subjected to Kalman filtering to obtain the calibration position coordinates.

[0029] In an optional implementation, when the NFC tag ID fails to match the mapping relationship data in the three-dimensional topology model, the method further includes:

[0030] The IMU data is integrated with accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal;

[0031] The environmental video data is processed using the ORB-SLAM2 algorithm to obtain the three-dimensional point cloud data of the mobile terminal.

[0032] Visual features are extracted from the current frame image in the environmental video data to obtain two-dimensional image features;

[0033] The three-dimensional point cloud data and the two-dimensional image features are serialized and hashed to obtain a spatial fingerprint.

[0034] The spatial fingerprint, the IMU displacement change, and the three-dimensional point cloud data are subjected to Kalman filtering to obtain the calibration position coordinates.

[0035] In an optional implementation, the step of acquiring the visual data of the media terminal sent by the mobile terminal and performing multimodal visual verification on the visual data includes:

[0036] Multi-scale feature extraction processing is performed on the image frames in the visual data to obtain the advertising frame boundary information;

[0037] Geometric normal vectors are calculated based on the visual data to obtain the normal vectors of the advertising frame.

[0038] PnP calculation is performed based on the advertising frame boundary information to obtain camera pose parameters;

[0039] The boundary information of the advertising frame and the preset template boundary information are matched to obtain a matching degree index;

[0040] The angle between the camera pose parameters and the normal vector of the advertising frame is calculated to obtain the similarity.

[0041] The matching degree index and the similarity are weighted and fused to obtain a comprehensive verification score;

[0042] The comprehensive verification score is compared with a preset visual verification threshold to verify whether the visual data meets the preset guidance data synthesis requirements.

[0043] In an optional implementation, the step of performing multi-dimensional source tracing analysis on the operation log based on the task guidance data and a preset anomaly quality inspection model to obtain a quality inspection source tracing report includes:

[0044] The time deviation data, path deviation data, and verification score data in the operation log are normalized to obtain a standardized operation log dataset.

[0045] The standardized operation log dataset is traversed and processed by a preset anomaly inspection model to obtain an anomaly score set.

[0046] The abnormal score set is spatiotemporally correlated with the AR rendering trigger record in the task guidance data to obtain a spatiotemporally coupled abnormal task set.

[0047] The spatiotemporal coupling abnormal task set is subjected to multi-dimensional index aggregation processing to obtain the link index deviation matrix;

[0048] The deviation matrix of the process indicators is weighted and standardized to obtain the identifier of the process with the largest deviation.

[0049] The AR instruction rendering data in the task guidance data is correlated and mapped according to the identifier of the maximum deviation link to generate the quality inspection traceability report.

[0050] A second aspect of this application provides a media terminal uploading task management device, the device comprising:

[0051] The model building module is used to obtain the target building plan dataset and the media terminal label dataset in the target building according to the obtained task instructions, and to build a three-dimensional spatial model based on the target building plan dataset and the media terminal label dataset to obtain a three-dimensional topological model.

[0052] The path planning module is used to obtain the current location information of the mobile terminal, and to perform dynamic path planning based on the current location information, the task instructions and the three-dimensional topology model to obtain the optimal path sequence.

[0053] The coordinate calibration module is used to perform multimodal position calibration processing on the spatial position verification dataset sent by the mobile terminal when the spatial position verification dataset is obtained, so as to obtain the calibration position coordinates.

[0054] The location verification module is used to obtain the media terminal coordinates based on the calibration location coordinates and the optimal path sequence, and to verify the location of the mobile terminal based on the media terminal coordinates and the calibration location coordinates.

[0055] The visual verification module is used to acquire the visual data of the media terminal sent by the mobile terminal when the mobile terminal passes the location verification, and to perform multimodal visual verification on the visual data.

[0056] The task guidance module is used to obtain a target task sub-instruction from the task instruction according to the coordinates of the media terminal when the visual data passes the multimodal visual verification, and to generate task guidance data for the media terminal according to the target task sub-instruction and the visual data.

[0057] The traceability and quality inspection module is used to perform multi-dimensional traceability analysis on the operation logs sent by the mobile terminal when the operation logs are obtained, based on the task guidance data and the preset abnormal quality inspection model, so as to obtain a quality inspection and traceability report.

[0058] A third aspect of this application provides an electronic device, the electronic device 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 media terminal uploading task management method as described above.

[0059] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the media terminal uploading task management method described above.

[0060] In summary, this application includes at least the following beneficial technical effects:

[0061] 1. By constructing a three-dimensional topology model and performing dynamic path planning, operators can quickly locate the target media terminal, greatly shortening the search time and improving the efficiency of task execution.

[0062] 2. By utilizing multimodal position calibration and position verification steps, it can be ensured that the operator has indeed reached the designated media terminal position, avoiding task delays or erroneous execution due to incorrect position.

[0063] 3. Based on the coordinates of the media terminal and the task instructions, the generated task guidance data provides clear and specific operation guidance for operators, reducing the difficulty of operation and the error rate.

[0064] 4. By conducting multi-dimensional source tracing analysis of operation logs, problems in the operation can be identified and corrected in a timely manner, ensuring the quality of published tasks. At the same time, the generated quality inspection source tracing report also provides important reference for subsequent quality improvement. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0066] Figure 1 This is a flowchart of a media terminal uploading task management method provided in an embodiment of this application;

[0067] Figure 2 This is a terminal interaction diagram illustrating a media terminal uploading task management method provided in an embodiment of this application;

[0068] Figure 3 This is a functional block diagram of a media terminal uploading task management device provided in an embodiment of this application;

[0069] Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0070] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0071] The media terminal uploading task management method provided in this application embodiment is executed by a server, and correspondingly, the media terminal uploading task management device runs on the server. The server triggers the execution of the media terminal uploading task management method provided in this application embodiment according to the uploading task instruction, thereby guiding construction personnel to complete the uploading task through interaction between the server and the construction personnel's mobile terminal. The following description, from the server's perspective and in conjunction with the uploading task execution process, explains the media terminal uploading task management method provided in this application embodiment.

[0072] Please refer to the above. Figures 1 to 2 The media terminal uploading task management method provided in this application includes the following steps.

[0073] Step S1: Obtain the target building floor plan dataset and the media terminal tag dataset in the target building according to the obtained task instructions, and construct a three-dimensional spatial model based on the target building floor plan dataset and the media terminal tag dataset to obtain a three-dimensional topology model.

[0074] Upon receiving a task instruction from the administrator, the system retrieves the corresponding target building floor plan dataset and the media terminal tag dataset pre-set for the target building from the database based on the publication address recorded in the task instruction. The target building floor plan dataset typically includes, but is not limited to, the building's CAD drawings, floor plan designs, geographic information system data of the building, and information such as the location coordinates, size, and shape of each media terminal (i.e., advertising spaces with NFC tags, hereinafter collectively referred to as advertising spaces).

[0075] The system parses the floor plan data of the target building, including CAD files, GIS data, or drawings. This typically involves using graphics processing software or custom algorithms (such as OpenCV or Shapely libraries) for coordinate extraction, geometric analysis, and boundary recognition. This transforms the two-dimensional coordinate information into an ad placement metadata table. This table contains information such as the geographic coordinates, floor level, ad type, shape, and size of each ad placement.

[0076] The actual location of a building is typically represented by latitude and longitude coordinates (e.g., WGS84 coordinate system), but modeling usually uses a planar coordinate system. This application uses a UTM coordinate system for modeling, therefore, the geographic coordinates of each ad space need to be transformed. Specifically, the following coordinate transformation algorithm is used to convert the two-dimensional geographic coordinates in the ad space metadata table to the UTM coordinate system suitable for three-dimensional spatial calculations, thereby obtaining the corresponding three-dimensional coordinate dataset for each ad space in the three-dimensional spatial model. The three-dimensional coordinate dataset includes, but is not limited to, the transformed planar coordinates and the floor height of the ad space. Simultaneously, to facilitate ad space management, all ad spaces in the target building are uniformly numbered when deploying media terminals. In the target building, the ad space number is the unique identification code of the media terminal. After obtaining the three-dimensional coordinates of each ad space, the three-dimensional coordinates and the corresponding ad space number are packaged together and stored in the three-dimensional coordinate dataset.

[0077]

[0078] in, These are the geographical latitude and longitude coordinates of the ad placement. The reference latitude and longitude (i.e., the center point of the target building). This is the Earth's average radius (approximately 6,378,137 meters). This refers to the height of the floor where the advertisement is located. , , (The three-dimensional coordinates of each ad placement after conversion)

[0079] In three-dimensional space, the positional relationships between ad placements are crucial. To construct the topological relationships of ad placements, the Delaunay triangulation algorithm can be used to process the three-dimensional coordinate dataset of ad placements. Delaunay triangulation generates non-overlapping and relatively uniform triangular meshes, helping to determine the adjacency relationships between ad placements. Specifically, the Delaunay triangulation algorithm divides the three-dimensional coordinate dataset into multiple triangular units, and by calculating the adjacency of each point (i.e., ad placement) within a triangular unit, the adjacency matrix A is obtained. nxn Among them, matrix A nxn China A ij =1 indicates that ad slot i and ad slot j are adjacent in space, A ij =0 indicates that ad slot i and ad slot j are not adjacent in space. Delaunay triangulation maximizes the minimum angle of each unit, avoiding the generation of weak triangles. The adjacency matrix obtained by Delaunay triangulation describes the spatial relationship between ad slots, providing a basis for subsequent path planning.

[0080] In the media terminal tag dataset, each ad slot is bound to a unique NFC tag. Therefore, when building a 3D spatial model and guiding construction personnel to perform the placement task, it is necessary to obtain the mapping relationship between the tag ID and the ad slot number for each ad slot. By parsing the media terminal tag dataset and summarizing the parsed data, a correlation dataset between tag IDs and ad slot numbers is generated. Using the ad slot numbers in the correlation dataset, a correlation relationship between tag IDs and 3D coordinates can be established, thus allowing the tag ID to identify the corresponding ad slot. Furthermore, by performing a matching algorithm on the correlation dataset and the 3D coordinate dataset with a pre-defined tag mapping model, the tag IDs are associated with the 3D coordinates of the ad slots, ensuring that each tag ID corresponds to the correct ad slot, ultimately obtaining the mapping relationship data between tags and the 3D spatial model.

[0081] In multi-level target buildings, the distance between advertising spaces and the time cost of switching floors are key factors in optimal route planning. In one alternative implementation, the edge weights between advertising spaces are calculated using the following formula based on scientific motion statistics.

[0082]

[0083] in, This represents the edge weight between ad slot i and ad slot j. Let i and j represent the three-dimensional coordinates of ad slot i and ad slot j, respectively. The preset walking speed (usually 1.2m / s). The time cost for switching floors (usually 20 seconds). This represents the floor height difference between ad slot i and ad slot j. By calculating the edge weights, the walking time and floor switching cost between ad slots are reflected. This provides crucial data support for subsequent path planning.

[0084] Based on the 3D coordinates, adjacency matrix, and mapping relationship data of the ad placements, a complete 3D spatial model (i.e., a 3D topological model) is generated using graphical modeling algorithms (e.g., topology analysis algorithms, 3D modeling tools, etc.). The 3D topological model describes the spatial relationships between ad placements, the position of each ad placement, and its mapping relationship with tags. The 3D topological model can be represented as follows: .in, The three-dimensional coordinates of each ad placement. This is an adjacency matrix between ad slots. The edge weights between ad slots.

[0085] Furthermore, the server performs mapping processing on the 3D topology model based on the task dataset in the task instructions, thereby obtaining a task-coordinate mapping table that records the ad slot coordinate information corresponding to each task. The task dataset contains multiple task sub-instructions, the ad slot number corresponding to each task sub-instruction, and the task deadline, among other information. Using the ad slot number in the task data, the corresponding 3D ad slot data and tag ID can be obtained from the 3D topology model. By summarizing the task sub-instructions with the corresponding 3D ad slot data and tag ID, the task-coordinate mapping table that records the ad slot coordinate information corresponding to each task is obtained.

[0086] Step S2: Obtain the current location information of the mobile terminal, and perform dynamic path planning based on the current location information, the task instruction, and the three-dimensional topology model to obtain the optimal path sequence.

[0087] It should be understood that the portable mobile terminal is a device integrating an IMU chip, which collects and uploads the mobile terminal's current IMU data (i.e., acceleration data and gyroscope data) to the server in real time according to preset time intervals. The server can use the IMU data to infer the movement status of the construction worker within a corresponding time period (e.g., movement vector and predicted position coordinates). The acceleration data provides the mobile terminal's acceleration information along various axes, provided by the IMU chip and accelerometer. By integrating the acceleration data, the velocity and displacement of the construction worker carrying the mobile terminal can be calculated. The gyroscope data provides the mobile terminal's rotational angular velocity, which helps in calculating the mobile terminal's rotation angle and direction.

[0088] When acquiring the current location information (usually GPS information or preset initial coordinates) transmitted back by the mobile terminal in response to the location information acquisition command, the current location information is estimated using an integration method, combined with real-time acquired and uploaded IMU data. Specifically, the velocity change is obtained through accelerometer acceleration data, and then the rotation angle of the mobile terminal is obtained by combining it with gyroscope data, ultimately calculating the relative displacement. To reduce errors, coordinate calibration is performed using the initial reference coordinates in the current location information (e.g., transmitted GPS data), and Kalman filtering technology is used to fuse data from multiple sensors to obtain more accurate current coordinates (i.e., initial coordinates). The position estimation formula for IMU data is shown below:

[0089]

[0090] in, This represents the acceleration measured by the accelerometer in three directions. This is the time interval, usually the sampling period. This represents the displacement along the corresponding axial direction.

[0091] The Kalman filter formula is shown below:

[0092]

[0093] in, It is the estimated state at the current moment (i.e., the calibrated coordinates). It is the state transition matrix, representing the dynamic model from the previous time step to the current time step. It controls the effect of inputs on the state. It is the Kalman gain, used to assess the confidence level of weighted measurements. It is a measurement (i.e., the output of GPS or IMU). It is the observation matrix, representing the mapping from the state space to the observation space.

[0094] Simultaneously, a task execution queue is generated by calculating task priorities. Task priorities can be sorted based on factors such as urgency and deadlines, ensuring the system executes the most urgent tasks first. The formula for calculating task priority scores is shown below:

[0095]

[0096] in, For the task Priority score. For the task The urgency level (which can be defined based on factors such as task type and task importance). The highest urgency value among all tasks. For the task The deadline. Current time. These are weighting coefficients used to balance the influence of different factors.

[0097] Finally, optimal path sequences are generated using path optimization algorithms (i.e., Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO)) based on the initial coordinates, spatial relationships in the adjacency matrix, and the task execution queue. The path optimization algorithms generate the shortest path for each task node based on the adjacency matrix between ad slots and the calculated path costs (e.g., walking time, floor switching costs, etc.). Specifically, the PSO algorithm initializes a particle swarm, calculates the total cost of a path represented by each particle, and the particles search for the optimal path in the path space. The fitness function of PSO is shown below:

[0098]

[0099] in, For path The fitness value represents the total cost of the path. To use the ad space To the ad space The distance. This represents the maximum distance. For the maximum time. The weights are distance and time, respectively.

[0100] Multiple candidate paths can be obtained through the fitness function of POS. ACO local optimization can be combined to update pheromones in the 3D topology model to guide the search for candidate paths, thereby selecting the optimal path sequence. The pheromone concentration reflects the quality of the candidate paths. After receiving the optimal path sequence from the server, the mobile terminal can synthesize an electronic map of the optimal path by combining it with the previously acquired 3D topology model. In a multi-task environment, the shortest path is found based on the connection relationships of ad placements and task priorities, reducing the working time of construction personnel and improving task execution efficiency.

[0101] Step S3: When the spatial location verification dataset sent by the mobile terminal is obtained, multimodal location calibration processing is performed on the spatial location verification dataset to obtain the calibration location coordinates.

[0102] Once the construction workers arrive at a certain advertising location, they use their mobile terminals to read the NFC tag on the media terminal. After obtaining the NFC tag, the mobile terminal triggers a preset video capture command, thereby acquiring environmental video data of the current space through the camera. The read NFC tag, IMU data, and environmental video data are then combined into a spatial location verification dataset, which is then sent to the server.

[0103] After receiving the spatial location verification dataset, the server demodulates the NFC tag's radio frequency signal and parses the protocol to obtain the corresponding NFC tag ID. Specifically, the mobile terminal's NFC chip (e.g., PN5180) receives a 13.56MHz electromagnetic wave signal from the NFC tag. The signal is demodulated into a subcarrier signal via load modulation and then decoded into binary data using Manchester encoding rules. The decoded data is parsed using the ISO / IEC 14443-3 protocol. The frame structure of this protocol includes a frame header, command code, data field, and checksum (i.e., CRC16-CCITT). The data is verified using the checksum to ensure the integrity of the tag information. If the verification fails, a retransmission mechanism is executed (up to 3 retries). Further, the NFC tag ID is extracted from the parsed data. The NFC tag ID is typically a unique identifier of 4 or 7 bytes. A task-coordinate mapping table obtained from the 3D topology model is used to perform a lookup and matching based on the NFC tag ID. The task-coordinate mapping table stores the tag ID of the advertising space and its corresponding 3D coordinate information.

[0104] Simultaneously, the IMU data from the mobile terminal is integrated using accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal. The integration process is consistent with the IMU data position calculation process in step S2, and will not be elaborated further here. Please refer to the IMU data position calculation process in step S2 for details.

[0105] Simultaneously, the environmental video data is processed using the ORB-SLAM2 algorithm to generate 3D point cloud data of the environment, containing the 3D coordinates of each feature point in the environment. ORB-SLAM2 is a vision-based SLAM (Simultaneous Localization and Mapping) algorithm that processes the video stream in real time, extracts feature points from the images, and uses these feature points for spatial localization and mapping. In each frame, the ORB-SLAM2 algorithm uses the ORB feature extraction method to detect feature points in the image and calculates the camera pose (i.e., the mapping relationship between the image and the 3D position and orientation in space) using the PnP algorithm. The minimization formula for the PnP problem is shown below:

[0106]

[0107] in, and The camera's rotation matrix and translation vector represent the camera's position and orientation. These are the coordinates of a three-dimensional point. These are the coordinates of a point in a two-dimensional image. This is the camera projection function.

[0108] When the NFC tag ID is used to search and match in the task-coordinate mapping table and obtain the corresponding 3D coordinates (i.e., target coordinates), a Kalman filter algorithm is used to fuse multi-sensor data (IMU data, visual data) to optimize positioning accuracy. The Kalman filter algorithm effectively eliminates integration errors in IMU data and further corrects position estimation by combining visual data. The execution process of the Kalman filter algorithm is the same as the Kalman filter formula execution process in step S2, except that the control input is replaced with the IMU displacement change, and the current observation value is replaced with target coordinates or 3D point cloud data. Further details are omitted here; please refer to the Kalman filter formula execution process in step S2 for more information. The calibrated position coordinates obtained through Kalman filtering achieve an accuracy of ±0.1 meters.

[0109] When a match is not found in the task-coordinate mapping table based on the NFC tag ID (i.e., the mapping relationship between the NFC tag ID and the 3D topology model fails to match), a Gaussian pyramid is constructed using the SIFT algorithm on the environmental video data, and Gaussian difference images at different scales are calculated. At each scale, the algorithm detects local extrema (i.e., points that have extrema at that scale and in adjacent scales). Extrema are potential keypoints, representing areas of significant variation in the image. The SIFT algorithm uses Taylor expansion to precisely locate each potential keypoint, further eliminating low-contrast and edge response points. The response of each keypoint is corrected using a Hessian matrix to ensure that only stable and valid keypoints are retained. A principal direction is assigned to each corrected keypoint to ensure rotation invariance. Direction assignment is accomplished by calculating the gradient orientation histogram of the region surrounding the keypoint. A principal direction is selected from the various directions, and all subsequent descriptors are rotated based on this direction to ensure that the features extracted from the same object are consistent under different rotation angles. The region surrounding the keypoint is divided into 4×4 sub-regions, and an 8-directional gradient histogram is calculated for each sub-region, ultimately generating a 128-dimensional feature vector.

[0110] Furthermore, efficient algorithms such as kd-trees are used for fast matching between descriptors. By calculating the Euclidean distance between descriptors, the most similar keypoint pairs between the current and reference images are found. For each matching descriptor pair, the ratio of the Euclidean distance between the nearest and second nearest neighbor descriptors is calculated. If the ratio is less than a predetermined threshold (usually 0.7), the matching pair is considered a high-confidence match and is retained. Through ratio testing, false matches caused by repetitive textures or noise can be effectively eliminated, retaining feature point pairs with high geometric and appearance consistency.

[0111] The keypoints in the filtered matching pairs are associated with 3D point cloud data (from environmental data generated by ORB-SLAM2). Each matching pair contains keypoints from the current image, keypoints from a reference image, and their D-SIFT descriptors. For example, a keypoint (x) in the current image... i y i ) corresponds to a point (X) in a 3D point cloud. i Y i Z i The associated 3D point data (X) i Y i Z i The 3D point cloud data and D-SIFT descriptors are converted to binary format and serialized. The serialized 3D point cloud data and descriptor data are concatenated into a binary stream. Finally, the SHA3-256 hash algorithm is applied to the concatenated data to generate a unique spatial fingerprint. The spatial fingerprint is then used to replace the target coordinates and perform a Kalman filter algorithm to obtain calibrated position coordinates with a positioning error reduced to ±0.1 meters.

[0112] Step S4: Obtain the media terminal coordinates based on the calibration location coordinates and the optimal path sequence, and verify the location of the mobile terminal based on the media terminal coordinates and the calibration location coordinates.

[0113] After obtaining the calibration location coordinates of the mobile terminal, the location of the mobile terminal is verified by combining the media terminal coordinates (i.e., verifying whether the construction personnel have reached the corresponding location). First, the ad slot that the current task needs to execute is selected from the optimal path sequence, and its coordinates are extracted. Assuming that the current task requires the construction personnel to execute a task located at ad slot k, then the media terminal coordinates are the coordinates corresponding to ad slot k. .

[0114] The distance error between the current mobile terminal's calibration location and the current media terminal's coordinates is calculated using the Euclidean distance formula. This distance error value will be used to determine whether the construction personnel have reached the target advertising space (i.e., the advertising space where the current task requires execution). The Euclidean distance formula is shown below:

[0115]

[0116] in, This represents the distance error between the current calibration location and the coordinates of the target media terminal. ( ) represents the three-dimensional coordinates of the calibration position. ( ) represents the three-dimensional coordinates of the target media terminal.

[0117] By comparing the calculated distance error With respect to the preset allowable error threshold To verify whether the construction workers have reached the target advertising position. If the verification is successful, the mobile terminal will receive the visual data acquisition command; when If the verification fails, the mobile terminal will remind the construction personnel to adjust their current location according to the preset location adjustment prompt method (for example, a preset SMS message prompt).

[0118] Step S5: When the mobile terminal passes the location verification, obtain the visual data of the media terminal sent by the mobile terminal, and perform multimodal visual verification on the visual data.

[0119] Upon receiving real-time image data (i.e., visual data) from the mobile terminal, multi-scale feature extraction algorithms, such as SIFT (Scale Invariant Feature Transform) or SURF (Speeded Robust Feature Transform), are used to extract keypoints and descriptors from the image. Keypoints and descriptors maintain the stability of edge information across different viewpoints and sizes. By detecting feature points (i.e., keypoints and descriptors) in the image and analyzing their local gradient information, the boundaries of the advertising frame are obtained. Typically, the advertising frame is rectangular or other regular shapes, and the boundary lines can be further extracted more precisely using Hough transform or other edge detection algorithms. Feature point detection can be performed using the following formula:

[0120]

[0121] in, Midpoint of the image The pixel value. This represents the gradient of the image along the x and y axes. Gradient calculation helps identify edges and feature points in the image. The extracted boundary information of the advertising frame is typically represented as a set of boundary points or a rectangle. Multi-scale feature extraction helps capture all important details in the image, especially under different angles and lighting conditions, ensuring the stability and reliability of the advertising frame's boundary information.

[0122] The normal vector of an advertising frame is a vector perpendicular to its surface. It is calculated using the vertices of a triangle or rectangle based on the frame's boundary information. For a rectangular advertising frame, the normal vector can be calculated using the cross product of the two sides of each vertex. Specifically, it is calculated using the two edge vectors... and The outer product is used to obtain the normal vector of the advertising frame.

[0123] The PnP algorithm calculates the camera's position and pose based on the 2D image coordinates and corresponding 3D spatial coordinates (derived from the advertising frame's boundary information and normal vectors). Specifically, the goal of the PnP algorithm is to calculate the camera's pose—its rotation matrix R and translation vector T—using a set of known 3D points and their projections into the image. PnP calculations help determine the camera's specific position and orientation relative to the advertising frame.

[0124] Image processing algorithms (e.g., Hough transform or edge matching) are used to compare the boundary information of the current advertisement frame with that of a preset template, and a matching degree index is calculated between the two boundary information. Boundary matching is a crucial step in determining whether the advertisement frame conforms to preset standards. By calculating the matching degree index, the similarity between the advertisement frame and the template can be assessed. Simultaneously, the angle between the normal vector in the camera pose and the normal vector of the advertisement frame is calculated using the dot product formula shown below. The size of this angle (i.e., the similarity) indicates the accuracy with which the camera is oriented towards the advertisement frame.

[0125]

[0126] in, Let be the camera's orientation normal vector. This is the normal vector of the advertisement frame. This refers to the angle between the camera and the advertising frame. By calculating this angle, the accuracy of the camera's alignment with the advertising frame can be quantified, ensuring the correctness of the position and the degree of alignment.

[0127] Furthermore, the matching score and similarity score are weighted and fused according to preset weights to calculate a comprehensive verification score. This weighted fusion comprehensively considers the boundary matching of the advertisement frame and the consistency of camera orientation, ensuring more accurate multimodal verification of the visual data. The comprehensive verification score is then compared with a preset visual verification threshold. If the comprehensive verification score is greater than the visual verification threshold, the visual data meets the preset requirements; otherwise, verification fails.

[0128] Step S6: When the visual data passes the multimodal visual verification, the target task sub-instruction is obtained from the task instruction according to the media terminal coordinates, and the task guidance data of the media terminal is generated according to the target task sub-instruction and the visual data.

[0129] The spatial relationship between task instructions and ad placements is matched using database queries or spatial indexing methods. By employing spatial coordinate matching, the task instructions search for matching sub-instructions based on the target ad placement, ensuring a correspondence between sub-instructions and actual ad placements. Sub-instructions define the specific tasks that installers must perform (e.g., ad installation, ad content modification, ad size adjustment). Based on the ad placement's boundary information, detailed operational guidelines, including ad position, size, and rotation angle, are generated to ensure the ad is correctly installed according to task requirements. ARKit's spatial anchoring technology overlays virtual content (e.g., ad models) onto the actual ad placement in the environment. Based on the sub-instructions, frame-by-frame animation data of the installation steps is generated, providing real-time AR guidance data (i.e., task guidance data) to installers. This AR guidance data includes a virtual display of the ad, adjustment steps, and required tools, helping installers perform operations according to task requirements.

[0130] Step S7: When the operation log sent by the mobile terminal is obtained, the operation log is analyzed in multiple dimensions according to the task guidance data and the preset abnormal quality inspection model to obtain a quality inspection and traceability report.

[0131] During the process of the construction workers carrying out the advertising placement task, the mobile terminal will record relevant data such as time, path, and verification score in real time. After each task sub-instruction is completed, the construction workers will be prompted to take a picture of the display effect of the current advertising space. The real-time recorded task-related data and display effect pictures will be combined into an operation log and uploaded to the server.

[0132] After receiving the operation log, the server normalizes the time deviation data, path deviation data, and verification score data in the log. Then, it performs the following standardization processing on the normalized data:

[0133]

[0134] in, Operation log data (time deviation, path deviation, verification score). For dataset The mean. For dataset The standard deviation. Through standardization, data is transformed into dimensionless values, ensuring that data from different dimensions can be analyzed on the same scale.

[0135] The pre-defined anomaly detection model uses the Isolation Forest algorithm to calculate anomaly scores on the standardized operation log dataset, thereby determining the presence of anomalies based on these scores. The Isolation Forest algorithm is used to determine whether a sample data point is an outlier by calculating its degree of isolation:

[0136]

[0137] in, Anomaly is the score for anomalies in a sample. It represents the total number of samples.

[0138] Based on the input historical data, the standardized operation log data is processed using a pre-defined anomaly detection model to obtain an anomaly score set. Anomaly source localization is performed using the following formula:

[0139]

[0140] in, For the first The operation data is in the first Values ​​in a dimension. For the first The mean of the dimension. For the first Standard deviation of the dimension. For the first Dimension weights.

[0141] The set of abnormal scores is spatiotemporally correlated with the AR rendering trigger records in the task guidance data to ensure that each abnormal score matches the corresponding task guidance and execution steps. The AR rendering trigger records in the task guidance data contain the task's trigger time, location, and corresponding operation instructions. This information is matched with the abnormal scores to determine the specific spatiotemporal location of each abnormality. Based on the spatiotemporal correlation, a spatiotemporally coupled abnormal task set is generated, recording the abnormal behavior of each task stage and its corresponding task guidance data.

[0142] A multi-dimensional index aggregation method is used to aggregate a set of spatiotemporally coupled anomaly tasks, calculating the deviation matrix for each task stage, reflecting the deviation between the performance of each stage and the standard. The aggregation process is based on multiple dimensions (e.g., time, path, validation score, etc.), summarizing the execution results of each stage and calculating the deviation value. Through multi-dimensional index aggregation, a comprehensive analysis of each stage of task execution can be performed, identifying weak points in the execution process.

[0143] The deviation matrix is ​​weighted, assigning different weights based on the importance of different stages or their impact on task execution. The weighted results are then standardized to ensure all deviation indicators are within the same scale. This allows the stage with the largest deviation to be identified. The stage with the largest deviation may be a bottleneck or source of error in the execution process.

[0144] Based on the identifier of the point of greatest deviation, the AR command rendering data in the task guidance data is associated with the corresponding task execution stage to generate a quality inspection traceability report. The report will include information on the point of greatest deviation, detailed content of the task guidance data, operating steps, and expected execution results.

[0145] This application applies to the field of publication management technology. It generates a 3D topology model by constructing a spatial model from a target building's planar dataset and a media terminal tag dataset. Combining the mobile terminal's current location information and acquired task instructions, it performs path planning to generate an optimal path sequence. The mobile terminal's location is verified using calibrated coordinates. Visual data sent by the verified mobile terminal is then used to verify the location of construction personnel. Based on the multimodal verified visual data and task instructions, task guidance data is generated. Upon obtaining the operation log, an anomaly analysis model is used to trace the log's source, generating a quality inspection report. This application achieves efficient and high-quality management of publication tasks through path planning, location verification, task guidance generation, and source analysis.

[0146] like Figure 2 The diagram shown is a functional block diagram of a media terminal uploading task management device provided in an embodiment of this application.

[0147] In some embodiments, the media terminal's uploading task management device 2 may include multiple functional modules composed of computer program segments. The computer programs of each program segment in the media terminal's uploading task management device 2 may be stored in the server's memory and executed by at least one processor to perform (see details). Figure 1 (Description) Functions of the media terminal's uploading task management method.

[0148] In this embodiment, the media terminal's uploading task management device 2 can be divided into multiple functional modules according to its functions. These functional modules may include: a model building module 21, a path planning module 22, a coordinate calibration module 23, a position verification module 24, a visual verification module 25, a task guidance module 26, and a traceability and quality inspection module 27. The module referred to in this invention is a series of computer program segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be detailed in subsequent embodiments.

[0149] The model building module 21 is used to obtain the target building plan dataset and the media terminal label dataset in the target building according to the obtained task instructions, and to build a three-dimensional spatial model based on the target building plan dataset and the media terminal label dataset to obtain a three-dimensional topological model.

[0150] In an optional implementation, the model building module 21 is specifically used for:

[0151] Perform planar graphic parsing processing on the target building planar dataset to obtain the advertising space metadata table within the target building;

[0152] Perform UTM coordinate transformation on the two-dimensional geographic coordinates in the ad slot metadata table to obtain a three-dimensional coordinate dataset for the ad slot;

[0153] The ad space 3D coordinate dataset is subjected to Delaunay triangulation to obtain the adjacency matrix between ad spaces;

[0154] The media terminal tag dataset is parsed to obtain the associated dataset between tag IDs and ad slot numbers;

[0155] The associated dataset is matched with the three-dimensional coordinate dataset of the ad slot through a preset tag mapping model to obtain the mapping relationship data between media terminal tags and three-dimensional space model;

[0156] A three-dimensional spatial model is constructed based on the ad space 3D coordinate dataset, the adjacency matrix, and the mapping relationship data to obtain a 3D topological model.

[0157] The path planning module 22 is used to obtain the current location information of the mobile terminal, and to perform dynamic path planning based on the current location information, the task instructions and the three-dimensional topology model to obtain the optimal path sequence.

[0158] In an optional implementation, the path planning module 22 is specifically used for:

[0159] The current location information is calibrated to obtain the initial coordinates of the mobile terminal;

[0160] The task sub-instructions are sorted according to the task priority in the task instruction to obtain the task execution queue;

[0161] Path optimization is performed based on the initial coordinates, the adjacency matrix in the 3D topology model, and the task execution queue to obtain the optimal path sequence.

[0162] The coordinate calibration module 23 is used to perform multimodal position calibration processing on the spatial position verification dataset sent by the mobile terminal when the spatial position verification dataset is obtained, so as to obtain the calibration position coordinates.

[0163] In an optional implementation, the coordinate calibration module 23 is specifically used for:

[0164] The NFC tag is decoded to obtain the NFC tag ID, and the target coordinates are obtained by matching the NFC tag ID with the mapping relationship data in the three-dimensional topology model.

[0165] The IMU data is integrated with accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal;

[0166] The environmental video data is processed using the ORB-SLAM2 algorithm to obtain the three-dimensional point cloud data of the mobile terminal.

[0167] The target coordinates, the IMU displacement change, and the three-dimensional point cloud data are subjected to Kalman filtering to obtain the calibration position coordinates.

[0168] In an optional implementation, when the NFC tag ID fails to match the mapping relationship data in the three-dimensional topology model, the coordinate calibration module 23 is further configured to:

[0169] The IMU data is integrated with accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal;

[0170] The environmental video data is processed using the ORB-SLAM2 algorithm to obtain the three-dimensional point cloud data of the mobile terminal.

[0171] Visual features are extracted from the current frame image in the environmental video data to obtain two-dimensional image features;

[0172] The three-dimensional point cloud data and the two-dimensional image features are serialized and hashed to obtain a spatial fingerprint.

[0173] The spatial fingerprint, the IMU displacement change, and the three-dimensional point cloud data are subjected to Kalman filtering to obtain the calibration position coordinates.

[0174] The location verification module 24 is used to obtain the media terminal coordinates based on the calibration location coordinates and the optimal path sequence, and to perform location verification on the mobile terminal based on the media terminal coordinates and the calibration location coordinates.

[0175] The visual verification module 25 is used to acquire the visual data of the media terminal sent by the mobile terminal when the mobile terminal passes the location verification, and to perform multimodal visual verification on the visual data.

[0176] In an optional implementation, the visual verification module 25 is specifically used for:

[0177] Multi-scale feature extraction processing is performed on the image frames in the visual data to obtain the advertising frame boundary information;

[0178] Geometric normal vectors are calculated based on the visual data to obtain the normal vectors of the advertising frame.

[0179] PnP calculation is performed based on the advertising frame boundary information to obtain camera pose parameters;

[0180] The boundary information of the advertising frame and the preset template boundary information are matched to obtain a matching degree index;

[0181] The angle between the camera pose parameters and the normal vector of the advertising frame is calculated to obtain the similarity.

[0182] The matching degree index and the similarity are weighted and fused to obtain a comprehensive verification score;

[0183] The comprehensive verification score is compared with a preset visual verification threshold to verify whether the visual data meets the preset guidance data synthesis requirements.

[0184] The task guidance module 26 is used to obtain a target task sub-instruction from the task instruction according to the coordinates of the media terminal when the visual data passes the multimodal visual verification, and to generate task guidance data for the media terminal according to the target task sub-instruction and the visual data.

[0185] The traceability and quality inspection module 27 is used to perform multi-dimensional traceability analysis on the operation log sent by the mobile terminal when the operation log is obtained, based on the task guidance data and the preset abnormal quality inspection model, so as to obtain a quality inspection and traceability report.

[0186] In an optional implementation, the traceability quality inspection module 27 is specifically used for:

[0187] The time deviation data, path deviation data, and verification score data in the operation log are normalized to obtain a standardized operation log dataset.

[0188] The standardized operation log dataset is traversed and processed by a preset anomaly inspection model to obtain an anomaly score set.

[0189] The abnormal score set is spatiotemporally correlated with the AR rendering trigger record in the task guidance data to obtain a spatiotemporally coupled abnormal task set.

[0190] The spatiotemporal coupling abnormal task set is subjected to multi-dimensional index aggregation processing to obtain the link index deviation matrix;

[0191] The deviation matrix of the process indicators is weighted and standardized to obtain the identifier of the process with the largest deviation.

[0192] The AR instruction rendering data in the task guidance data is correlated and mapped according to the identifier of the maximum deviation link to generate the quality inspection traceability report.

[0193] It should be understood that the various variations and specific embodiments of the methods provided in the above embodiments are also applicable to the media terminal uploading task management device of this embodiment. Through the foregoing detailed description of the media terminal uploading task management method, those skilled in the art can clearly understand the implementation method of the media terminal uploading task management device in this embodiment. For the sake of brevity, it will not be described in detail here.

[0194] like Figure 3 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this application.

[0195] In a preferred embodiment of the present invention, the electronic device 3 may include, but is not limited to, a memory 31, at least one processor 32, and at least one communication bus 33.

[0196] Those skilled in the art should understand that Figure 3 The structure of the electronic device 3 shown does not constitute a limitation of the embodiments of the present invention. The electronic device 3 may also include more or fewer other hardware or software than shown, or different component arrangements.

[0197] In some embodiments, the electronic device 3 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, programmable gate arrays, digital processors, and embedded devices.

[0198] It should be noted that the electronic device 3 is merely an example. Other existing or future electronic products that are suitable for this application should also be included within the scope of protection of this application and are incorporated herein by reference.

[0199] In some embodiments, the memory 31 stores a computer program that, when executed by the at least one processor 32, implements all or part of the steps in the media terminal uploading task management method described above. The memory 31 includes a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data. Further, the computer-readable storage medium may primarily include a program storage area and a data storage area, wherein the program storage area may store an operating system, at least one application program required for a function, etc.

[0200] In some embodiments, the at least one processor 32 is the control unit of the electronic device 3, connecting various components of the electronic device 3 via various interfaces and lines. It executes programs or modules stored in the memory 31 and calls data stored in the memory 31 to perform various functions of the electronic device 3 and process data. For example, when the at least one processor 32 executes the computer program stored in the memory 31, it implements all or part of the steps of the media terminal uploading task management method described in this application embodiment; or it implements all or part of the functions of the media terminal uploading task management device. The at least one processor 32 may be composed of integrated circuits, such as a single-packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips.

[0201] In some embodiments, the at least one communication bus 33 is configured to enable communication between the memory 31 and the at least one processor 32, etc. Although not shown, the electronic device 3 may also include a power supply (e.g., a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 32 via a power management device, thereby enabling functions such as charging, discharging, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 3 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0202] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause an electronic device (which may be a personal computer, electronic device, or network device, etc.) or processor to execute portions of the methods described in the various embodiments of this application.

[0203] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0204] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0205] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for managing uploading tasks on a media terminal, characterized in that, The method includes: The target building floor plan dataset and the media terminal tag dataset in the target building are obtained according to the obtained task instructions. A three-dimensional spatial model is constructed based on the target building floor plan dataset and the media terminal tag dataset to obtain a three-dimensional topology model. The current location information of the mobile terminal is obtained, and dynamic path planning is performed based on the current location information, the task instructions, and the three-dimensional topology model to obtain the optimal path sequence. When the spatial location verification dataset sent by the mobile terminal is obtained, multimodal location calibration processing is performed on the spatial location verification dataset to obtain the calibration location coordinates; The media terminal coordinates are obtained based on the calibration location coordinates and the optimal path sequence, and the location of the mobile terminal is verified based on the media terminal coordinates and the calibration location coordinates. When the mobile terminal passes the location verification, the visual data of the media terminal sent by the mobile terminal is obtained, and the visual data is subjected to multimodal visual verification. When the visual data passes the multimodal visual verification, the target task sub-instruction is obtained from the task instruction according to the media terminal coordinates, and the task guidance data of the media terminal is generated according to the target task sub-instruction and the visual data. When the operation log sent by the mobile terminal is obtained, a multi-dimensional source tracing analysis is performed on the operation log based on the task guidance data and the preset abnormal quality inspection model to obtain a quality inspection source tracing report.

2. The media terminal uploading task management method according to claim 1, characterized in that, The step of constructing a three-dimensional spatial model based on the target building plan dataset and the media terminal tag dataset to obtain a three-dimensional topological model includes: Perform planar graphic parsing processing on the target building planar dataset to obtain the advertising space metadata table within the target building; Perform UTM coordinate transformation on the two-dimensional geographic coordinates in the ad slot metadata table to obtain a three-dimensional coordinate dataset for the ad slot; The ad space 3D coordinate dataset is subjected to Delaunay triangulation to obtain the adjacency matrix between ad spaces; The media terminal tag dataset is parsed to obtain the associated dataset between tag IDs and ad slot numbers; The associated dataset is matched with the three-dimensional coordinate dataset of the ad slot through a preset tag mapping model to obtain the mapping relationship data between media terminal tags and three-dimensional space model; A three-dimensional spatial model is constructed based on the ad space 3D coordinate dataset, the adjacency matrix, and the mapping relationship data to obtain a 3D topological model.

3. The media terminal uploading task management method according to claim 1, characterized in that, The step of performing dynamic path planning based on the current location information, the task instructions, and the three-dimensional topology model to obtain the optimal path sequence includes: The current location information is calibrated to obtain the initial coordinates of the mobile terminal; The task sub-instructions are sorted according to the task priority in the task instruction to obtain the task execution queue; Path optimization is performed based on the initial coordinates, the adjacency matrix in the 3D topology model, and the task execution queue to obtain the optimal path sequence.

4. The media terminal uploading task management method according to claim 3, characterized in that, The spatial location verification dataset includes NFC tags, IMU data, and environmental video data. Multimodal location calibration processing is performed on the spatial location verification dataset to obtain calibrated location coordinates, including: The NFC tag is decoded to obtain the NFC tag ID, and the target coordinates are obtained by matching the NFC tag ID with the mapping relationship data in the three-dimensional topology model. The IMU data is integrated with accelerometer and gyroscope data to obtain the IMU displacement change of the mobile terminal; The environmental video data is processed using the ORB-SLAM2 algorithm to obtain the three-dimensional point cloud data of the mobile terminal. The target coordinates, the IMU displacement change, and the three-dimensional point cloud data are subjected to Kalman filtering to obtain the calibration position coordinates.

5. The media terminal uploading task management method according to claim 1, characterized in that, The step of acquiring the visual data of the media terminal sent by the mobile terminal and performing multimodal visual verification on the visual data includes: Multi-scale feature extraction processing is performed on the image frames in the visual data to obtain the advertising frame boundary information; Geometric normal vectors are calculated based on the visual data to obtain the normal vectors of the advertising frame. PnP calculation is performed based on the advertising frame boundary information to obtain camera pose parameters; The boundary information of the advertising frame and the preset template boundary information are matched to obtain a matching degree index; The angle between the camera pose parameters and the normal vector of the advertising frame is calculated to obtain the similarity. The matching degree index and the similarity are weighted and fused to obtain a comprehensive verification score; The comprehensive verification score is compared with a preset visual verification threshold to verify whether the visual data meets the preset guidance data synthesis requirements.

6. The media terminal uploading task management method according to claim 1, characterized in that, The step of performing multi-dimensional source tracing analysis on the operation log based on the task guidance data and the preset anomaly quality inspection model to obtain a quality inspection source tracing report includes: The time deviation data, path deviation data, and verification score data in the operation log are normalized to obtain a standardized operation log dataset. The standardized operation log dataset is traversed and processed by a preset anomaly inspection model to obtain an anomaly score set. The abnormal score set is spatiotemporally correlated with the AR rendering trigger record in the task guidance data to obtain a spatiotemporally coupled abnormal task set. The spatiotemporal coupling abnormal task set is subjected to multi-dimensional index aggregation processing to obtain the link index deviation matrix; The deviation matrix of the process indicators is weighted and standardized to obtain the identifier of the process with the largest deviation. The AR instruction rendering data in the task guidance data is correlated and mapped according to the identifier of the maximum deviation link to generate the quality inspection traceability report.

7. A media terminal uploading task management device, characterized in that, The device includes: The model building module is used to obtain the target building plan dataset and the media terminal label dataset in the target building according to the obtained task instructions, and to build a three-dimensional spatial model based on the target building plan dataset and the media terminal label dataset to obtain a three-dimensional topological model. The path planning module is used to obtain the current location information of the mobile terminal, and to perform dynamic path planning based on the current location information, the task instructions and the three-dimensional topology model to obtain the optimal path sequence. The coordinate calibration module is used to perform multimodal position calibration processing on the spatial position verification dataset sent by the mobile terminal when the spatial position verification dataset is obtained, so as to obtain the calibration position coordinates. The location verification module is used to obtain the media terminal coordinates based on the calibration location coordinates and the optimal path sequence, and to verify the location of the mobile terminal based on the media terminal coordinates and the calibration location coordinates. The visual verification module is used to acquire the visual data of the media terminal sent by the mobile terminal when the mobile terminal passes the location verification, and to perform multimodal visual verification on the visual data. The task guidance module is used to obtain a target task sub-instruction from the task instruction according to the coordinates of the media terminal when the visual data passes the multimodal visual verification, and to generate task guidance data for the media terminal according to the target task sub-instruction and the visual data. The traceability and quality inspection module is used to perform multi-dimensional traceability analysis on the operation logs sent by the mobile terminal when the operation logs are obtained, based on the task guidance data and the preset abnormal quality inspection model, so as to obtain a quality inspection and traceability report.

8. An electronic device, characterized in that, The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the media terminal uploading task management method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the media terminal uploading task management method according to any one of claims 1 to 6.

Images