Personnel trajectory positioning and tracking method and system based on RFID and AI video identification

Abstract

The invention relates to the technical field of personnel positioning and tracking, and provides a personnel trajectory positioning and tracking method and system based on RFID and AI video identification, and the method comprises the steps: obtaining video data in an industrial environment; based on the video data, judging whether an event that the personnel target is shielded by the automation equipment occurs; if the event that the personnel target is shielded by the automation equipment occurs, acquiring preset motion information of the automation equipment; calculating a predicted reproduction space-time window of the personnel target based on the predetermined motion information and the state of the personnel target before the personnel target is shielded; and in the prediction reproduction space-time window, processing the video data, searching and identifying a personnel target so as to realize personnel track positioning and tracking. The method has the advantages of being capable of effectively coping with visual shielding of automation equipment on personnel targets, achieving rapid and accurate re-identification and trajectory tracking after shielding, and improving continuity and accuracy of personnel positioning and tracking.

Description

Technical Field

[0001] This invention relates to the technical field of personnel location tracking, specifically to a method and system for personnel trajectory location tracking based on RFID and AI video recognition. Background Technology

[0002] In modern industrial production and logistics warehousing environments, personnel tracking systems combining Radio Frequency Identification (RFID) and Artificial Intelligence (AI) video recognition technologies are typically deployed to improve management efficiency and operational safety. These systems acquire personnel's identity and approximate location by equipping them with RFID tags, while utilizing cameras throughout the area for visual capture and analysis to accurately record personnel movement trajectories. However, in practical applications, especially in complex scenarios where humans and automated equipment operate in close proximity, the reliability and accuracy of existing technologies face significant challenges. When multiple signal sources intertwine, visual obstruction is frequent, and environmental characteristics are constantly changing, the system often struggles to maintain continuous and accurate tracking of specific individuals. Beyond industrial and warehousing environments, personnel tracking systems based on the fusion of RFID and AI video recognition are increasingly being introduced into crowded public settings such as events and exhibitions. In these settings, the system uses RFID tags for personnel identification and area entry / exit statistics, while leveraging AI visual recognition technology for crowd flow monitoring and abnormal behavior detection, thus providing auxiliary support for maintaining order, emergency evacuation, and personnel dispatch. Although these application scenarios are not primarily focused on industrial automation, they still place higher demands on the accuracy and anti-interference capabilities of the system. In particular, under conditions of frequent occlusion, dense crowds, and drastic dynamic changes in the environment, existing technologies still face significant challenges in continuous tracking and identity verification.

[0003] In existing technologies, when a person is obscured by automated equipment, the system often struggles to accurately predict when and where they will reappear, leading to a break in the tracking chain. Even after the target reappears, the reliability of re-identification algorithms based on features such as body posture and gait is significantly reduced due to prolonged occlusion or complex environments, and tracking drift may even occur, erroneously shifting the tracking focus to other objects, thus failing to achieve continuous and accurate trajectory localization and tracking of a specific individual. Furthermore, in certain special work areas, such as cold storage facilities, environmental factors may cause a decline in camera image quality, further exacerbating the difficulty of person re-identification. Therefore, how to effectively address the visual occlusion of people by automated equipment and achieve rapid and accurate re-identification and trajectory tracking after occlusion is a critical problem that current technology urgently needs to solve.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for personnel trajectory positioning and tracking based on RFID and AI video recognition. It has the advantages of effectively dealing with the visual occlusion of personnel targets by automated equipment, and achieving fast and accurate re-identification and trajectory tracking after occlusion, thereby improving the continuity and accuracy of personnel positioning and tracking.

[0006] This application provides a method for locating and tracking personnel trajectories based on RFID and AI video recognition, including: Acquire video data in industrial environments; Based on video data, determine whether there are events where people are obscured by automated equipment; If an event occurs where a person is obscured by automated equipment, the predetermined motion information of the automated equipment is obtained. Based on the predetermined motion information and the state of the person target before it is occluded, the spatiotemporal window for the predicted reproduction of the person target is calculated. Within the predicted spatiotemporal window, video data is processed to search for and re-identify personnel targets in order to achieve personnel trajectory positioning and tracking.

[0007] The above solution can effectively address the visual occlusion of personnel targets by automated equipment, and achieve rapid and accurate re-identification and trajectory tracking after occlusion, thereby improving the continuity and accuracy of personnel positioning and tracking.

[0008] Optionally, this application also proposes a step of calculating the predicted spatiotemporal window of the person target based on the predetermined motion information and the state of the person target before being occluded, when the actual motion of the automated equipment deviates from the predetermined motion information and the predetermined motion information, and the person target moves during the occlusion period, including: Collect response data from various RFID readers covering the work area in an industrial environment; Analyze the response data to determine whether the automated equipment has experienced unexpected stillness corresponding to a deviation between the actual movement and the predetermined movement information, and determine the movement path of the personnel target during the occlusion period; If an unexpected standstill occurs, the time component of the predicted spatiotemporal window is modified based on the duration of the unexpected standstill. Based on the movement path of the personnel target, the spatial components of the predicted spatiotemporal window are corrected.

[0009] The above scheme further improves the accuracy of the prediction and reconstruction spatiotemporal window correction when automated equipment moves abnormally or when personnel move during occlusion, thus ensuring the robustness of tracking.

[0010] Optionally, this application also proposes a step for determining, based on video data, whether an event has occurred in which a person target is obscured by automated equipment, including: In video data, the intersection-union ratio (IUU) of the bounding boxes of people in consecutive image frames is used for analysis; When the intersection-over-union ratio of the bounding boxes of a person target in consecutive image frames is lower than a preset threshold, and the overlap between the last visible position of the person target and the bounding box of the automated device exceeds a preset occlusion overlap ratio, it is determined that an event has occurred in which the person target is occluded by the automated device.

[0011] The above scheme provides a specific method for determining events where personnel targets are occluded, improving the accuracy and automation of occlusion event recognition.

[0012] Optionally, this application also proposes a step of analyzing response data, determining whether the automated equipment has experienced unexpected stillness corresponding to a deviation between actual movement and predetermined movement information, and determining the movement path of the personnel target during the occlusion period when the number of moving targets within the occlusion area formed by the automated equipment is greater than one, including: Before a person is obscured by automated equipment, acquire the baseline response characteristics of the RFID tag corresponding to the person. Analyze the response data to determine whether the automated equipment has experienced unexpected stationary states corresponding to a deviation between the actual movement and the predetermined movement information, and extract the movement path characteristics from the response data; All movement path features are compared with the baseline response features to identify the target movement path features that match the baseline response features. The target movement path characteristics are determined as the movement path of the personnel target during the occlusion period.

[0013] The above solution solves the problem of accurately identifying the movement path of a specific person in situations with multiple targets occlusion, thus improving tracking accuracy in complex scenarios.

[0014] Optionally, this application also proposes a step of comparing all movement path features with reference response features to identify target movement path features that match the reference response features, including: The real-time response features of the mobile path characteristics are compared with the baseline response features; During the comparison process, it is determined whether the state of the real-time response feature changes from a state that matches the benchmark response feature to a state that does not match, and the matching state judgment result is obtained. If the matching status judgment result is yes, then analyze the response characteristic change process of the transition to the non-matching status, and determine whether the response characteristic change process is caused by the personnel target performing a specific operation action; If it is determined that the cause is the execution of a specific operation by a personnel target, the transformed real-time response feature is set as the updated baseline response feature, and the updated baseline response feature is compared with the movement path feature. All movement path features are compared with the baseline response features, or all movement path features are compared with the updated baseline response features. The movement path features that are identified as matching are determined as the target movement path features.

[0015] The above scheme takes into account the situation where personnel targets may perform specific operations during occlusion, resulting in changes in response characteristics. By dynamically updating the baseline response characteristics, the accuracy of matching and recognition is improved.

[0016] Optionally, this application also proposes a step of analyzing the change process of response characteristics when the state changes into a mismatch, and determining whether the change process is caused by a person performing a specific task, including: Within the shielded area formed by the automated equipment, the real-time response characteristics, excluding those that transition to a mismatched state, are selected and recorded as the response characteristics of other movement paths. Based on the response characteristics of other movement paths, the background change characteristics characterizing global disturbances are determined; Background change features are separated from the response feature changes and used as local change features; Based on local change characteristics, determine whether the change process of response characteristics is caused by personnel performing specific operational actions.

[0017] The above scheme provides a method to distinguish between specific human actions and global environmental disturbances, making the judgment of changes in response characteristics more accurate and avoiding misjudgment.

[0018] Optionally, this application also proposes a step of determining background change characteristics characterizing global disturbances based on the response characteristics of other movement paths, including: Based on the historical response characteristics of each of the other mobile paths, the contribution index of each other mobile path is determined. Based on the contribution index, the changes in response characteristics of other mobile paths are weighted to generate background change characteristics.

[0019] The above scheme, by weighting the response features of different movement paths, more accurately extracts global background change features, thereby improving the robustness of background disturbance recognition.

[0020] Optionally, this application also proposes a step of determining the contribution index of each other mobile path based on their respective historical response characteristics, including: Identify change points in the time series of historical response characteristics for each of the other movement paths; Calculate the dispersion statistics based on the time series data after the change point; If no change point is identified, the dispersion statistics are calculated based on the latest time series data preset in the historical response features. Contribution index is determined based on the dispersion statistics.

[0021] The above scheme provides a specific method for determining contribution indicators, making the generation of background change features more refined and adaptive.

[0022] Optionally, this application also proposes a step for obtaining predetermined motion information of the automated equipment if an event occurs where a person target is obscured by the automated equipment, including: Based on the video frame corresponding to the event where a person is obscured by automated equipment, the identification number of the automated equipment is identified, and a query request containing the camera number, timestamp, obstruction location and equipment number is sent to the scheduling and control system of the automated equipment. After receiving a query request, the dispatch and control system queries the task database and returns the scheduled motion information of the automated equipment.

[0023] The above scheme clarifies the specific process for obtaining the predetermined motion information of automated equipment, ensuring the timeliness and accuracy of the data required for the calculation of the spatiotemporal window for prediction and reproduction.

[0024] Optionally, this application also proposes a personnel trajectory positioning and tracking system based on RFID and AI video recognition, used to perform personnel trajectory positioning and tracking based on RFID and AI video recognition, including: The video data acquisition module is used to acquire video data in industrial environments. The occlusion event detection module is used to determine, based on video data, whether an event has occurred in which a person or target is occluded by automated equipment. The motion information acquisition module is used to acquire the predetermined motion information of the automated equipment if an event occurs in which a person is obscured by the automated equipment. The reproduction window calculation module is used to calculate the predicted reproduction spatiotemporal window of the person target based on the predetermined motion information and the state of the person target before it is occluded. The personnel location and tracking module is used to process video data within the predicted recurrence time and space window, search for and re-identify personnel targets, so as to realize personnel trajectory location and tracking.

[0025] The above scheme provides a system for implementing the above method, enabling the method to be practically deployed and applied, and has good operability.

[0026] As can be seen from the above, the personnel trajectory positioning and tracking method and system based on RFID and AI video recognition provided in this application introduces the concept of predictive recurrence spatiotemporal window and combines it with the predetermined motion information of automated equipment. After the personnel target is occluded, it can accurately predict the time and spatial range of its reappearance, thereby achieving rapid re-identification and continuous tracking. It has the advantages of effectively dealing with the visual occlusion of personnel targets by automated equipment and achieving rapid and accurate re-identification and trajectory tracking after occlusion, thus improving the continuity and accuracy of personnel positioning and tracking. Attached Figure Description

[0027] Figure 1 This is a flowchart of a method for locating and tracking personnel based on RFID and AI video recognition, according to one embodiment of the present invention. Figure 2 This is one of the flowcharts of a method for locating and tracking personnel based on RFID and AI video recognition in another embodiment of the present invention; Figure 3 This is a second flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition, as described in another embodiment of the present invention. Figure 4 This is the third flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 5 This is the fourth flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 6 This is the fifth flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 7 This is the sixth flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 8 This is the seventh flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 9 This is the eighth flowchart of a method for locating and tracking personnel trajectories based on RFID and AI video recognition in another embodiment of the present invention; Figure 10 This is a system block diagram of a personnel trajectory positioning and tracking system based on RFID and AI video recognition according to another embodiment of the present invention; Explanation of reference numerals in the attached figures: 1. Personnel trajectory positioning and tracking system based on RFID and AI video recognition; 11. Video data acquisition module; 12. Occlusion event judgment module; 13. Motion information acquisition module; 14. Reproduction window calculation module; 15. Personnel positioning and tracking module. Detailed Implementation

[0028] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0029] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] Traditional personnel tracking systems based on RFID and AI video recognition are widely used in highly automated operations such as industrial production and logistics warehousing, and are also beginning to expand into densely populated public places such as gatherings to achieve large-scale crowd identification, area management, and trajectory tracking. However, in actual deployments, the tracking stability and identification accuracy of these systems in complex and dynamic environments still have significant shortcomings, especially under conditions of frequent target occlusion, mixed tag usage, or severe environmental interference.

[0031] Taking industrial production and logistics warehousing environments as an example, when target personnel coexist with automated equipment or goods with RFID tags and interact frequently, the tracking system is prone to identification interruptions. Specifically, at the work entrance, if target personnel and goods enter the identification area simultaneously, they may jointly trigger RFID identification. The system struggles to establish a high-confidence initial identity and video entity correspondence, leading to uncertain initial binding. During subsequent operations, target personnel are often completely obscured by large equipment (such as stacker cranes and automated guided vehicles), causing a break in the video tracking chain and making it difficult to continuously acquire their movement trajectory. Furthermore, when the RFID tools or tags carried by the target personnel are temporarily handed over to others, the location signal source may be misjudged as a continuation of the original target, further interfering with overall identification accuracy. In certain special areas, such as when personnel are wearing uniforms, in insufficient ambient light, or when the camera's shooting angle is limited, the quality of visual information deteriorates significantly, and the system may completely lose its ability to effectively identify individuals, failing to complete uninterrupted and accurate trajectory tracking.

[0032] For example, suppose a large automated logistics center has a personnel management system that combines RFID and video surveillance. When an employee enters the warehouse work area, their RFID tag interacts with the RFID reader at the entrance. However, if simultaneously a driverless forklift carrying large goods, also tagged with RFID, passes by, the reader may collect RFID signals from both the employee and the goods. In this case, the employee's silhouette might be obscured by tall goods in the entrance camera's view, preventing the system from accurately binding the employee's identity information to the visual entity in the video. During the employee's subsequent movement, their silhouette may be frequently obscured by roving automated guided vehicles (AGVs) and tall shelving columns, making it difficult for visual feature-based re-identification algorithms to maintain continuous tracking. If an employee hands a handheld data scanning terminal with an independent RFID tag to another employee during work, the system may incorrectly associate the tool's movement trajectory with the original employee, causing the generated trajectory to deviate from reality. Furthermore, when employees enter special work areas, such as cold storage, all personnel wear uniform cold-weather clothing, have highly similar visual characteristics, and the camera lenses at the entrance of the cold storage may condense fog due to temperature differences, resulting in a decrease in image quality. At this time, the video system will have difficulty identifying specific individuals, causing the tracking chain to be completely broken and making it impossible to obtain the complete activity path of the employees.

[0033] To address this, this application proposes a method for personnel trajectory positioning and tracking based on RFID and AI video recognition, combining... Figure 1 As shown, it includes: S1, acquire video data in the industrial environment; S2, based on video data, determines whether an event has occurred in which a person target is obscured by automated equipment; S3, if an event occurs in which a person target is obscured by automated equipment, then obtain the predetermined motion information of the automated equipment; S4, Based on the predetermined motion information and the state of the person target before being occluded, calculate the spatiotemporal window for the predicted reproduction of the person target; S5 processes video data within the predicted spatiotemporal window, searches for and re-identifies personnel targets to achieve personnel trajectory positioning and tracking.

[0034] The determination of whether a person is obscured by automated equipment involves real-time analysis of the video stream to identify situations where a person is visually blocked by automated equipment operating in an industrial environment. This can be achieved using image processing, target detection, or motion analysis techniques. For example, analyzing changes in the target bounding box, a decrease in target confidence, or the degree of overlap between the target and the equipment boundary is crucial for timely detection of potential visual interruptions in video tracking, thereby triggering subsequent trajectory recovery mechanisms. Furthermore, acquiring the planned motion information of the automated equipment refers to obtaining data such as the planned movement trajectory, speed, direction, or work plan of the automated equipment within a specific time period. This can be achieved by retrieving data from the equipment control system, scheduling system, or task management database, such as by querying the equipment's preset path plan or work instructions. This primarily provides kinematic evidence for predicting the possible location of a person after obstruction. Furthermore, calculating the predicted spatiotemporal window for the reappearance of personnel targets refers to predicting the time range and spatial area within which the personnel target will visually reappear, based on the known motion patterns of automated equipment and the last visible state of the personnel target before occlusion. This can be achieved using methods such as motion models, probabilistic prediction, or geometric extrapolation. For example, it can be achieved by combining the equipment's movement path with the personnel's last known speed and direction. Its main purpose is to narrow down the scope of subsequent video searches and improve the efficiency and accuracy of re-identification. Finally, searching for and re-identifying personnel targets refers to analyzing video data within the predicted spatiotemporal range to rediscover and confirm previously occluded personnel targets. This can be achieved using techniques such as target detection, feature matching, or multimodal fusion. For example, it can utilize deep learning models for target detection combined with RFID signals for identity verification. Its main purpose is to restore the interrupted tracking chain and ensure the continuity and integrity of personnel trajectories.

[0035] This application's solution uses video data from an industrial environment as a foundation to provide raw information for subsequent analysis and processing. When the system determines, based on this video data, that a person is occluded by automated equipment, the core logic of the solution is triggered. Once the occlusion event is identified, the system immediately acquires the predetermined motion information of the automated equipment causing the occlusion. This predetermined motion information, combined with the person's last visible state before occlusion, is used to calculate the predicted recurrence spatiotemporal window for the person. This spatiotemporal window defines the time period and spatial area in which the person may visually reappear. Subsequently, within this predicted recurrence spatiotemporal window, the system performs targeted processing on the video data, concentrating resources to search for and re-identify the person. In this way, even if the person temporarily disappears from the video, the system can use the known motion patterns of the automated equipment and the person's last state to make predictions, thereby quickly capturing and confirming the person's identity when they reappear. This achieves continuous location tracking of the person's trajectory, effectively overcoming the limitations of traditional video tracking in occluded scenarios.

[0036] In some preferred embodiments, this application is implemented as follows: First, multiple high-definition network cameras deployed within the industrial plant continuously acquire real-time video stream data, which is then transmitted to a central processing server for analysis. When the AI ​​visual analysis module on the server processes the video data, it detects personnel targets and automated equipment, such as unmanned transport vehicles or robotic arms, in real time. If the bounding box of a personnel target overlaps with the bounding box of an automated device, and the visibility or confidence of the personnel target significantly decreases, the system determines that the personnel target has been occluded by the automated device. Once such an occlusion event is identified, the system immediately sends a query request to the central scheduling and control system of the automated device to obtain motion information such as the device's predetermined operating path, speed, and stopping points for a future period. Simultaneously, the system records the personnel target's last known position, direction of movement, and speed before occlusion. Based on this predetermined motion information and the personnel target's state before occlusion, the system uses a motion prediction model, such as a prediction algorithm based on device path and personnel inertial motion, to calculate a predicted recurrence spatiotemporal window for the personnel target, which includes a time range and a spatial region. Subsequently, within this predicted spatiotemporal window, the system activates more sophisticated video analysis algorithms, such as feature-matching or deep learning-based re-identification models, to perform focused searches on the video data. Once a person is successfully identified within the prediction window, the system associates it with previous tracking chains, thereby achieving continuous location tracking of the person's trajectory.

[0037] Optional, combined Figure 2As shown, when the actual movement of the automated equipment deviates from the predetermined movement information, and the person target moves during the occlusion period, step S4, based on the predetermined movement information and the state of the person target before being occluded, calculates the predicted recurrence spatiotemporal window of the person target, including: S41 collects response data from various RFID readers covering the work area in an industrial environment; S42, Analyze the response data, determine whether the automated equipment has experienced an unexpected stationary state corresponding to a deviation between the actual movement and the predetermined movement information, and determine the movement path of the personnel target during the occlusion period; S43, if an unexpected stationary state occurs, then based on the duration of the unexpected stationary state, the time component of the predicted recurrence time-space window is modified. S44, based on the movement path of the personnel target, corrects the spatial components of the predicted spatiotemporal window.

[0038] Among them, the response data of the RFID reader refers to the electronic record generated by the RFID reader after receiving the RFID tag signal, which includes information such as the tag identification code, signal strength, and timestamp. This data can be collected in real time using an ultra-high frequency (UHF) RFID system or an active RFID system. Its purpose is to obtain the real-time location and status information of personnel and automated equipment within the obscured area. Unintended stationary movement refers to the automated equipment failing to move continuously according to the preset trajectory or speed during the planned movement process due to malfunctions, command errors, or external interference, and instead remaining stationary at a certain position. This can be determined by comparing the real-time location data of the automated equipment with the planned movement information; when the real-time position does not change for a long time or the change is much smaller than expected, it is judged to identify abnormal movement of the automated equipment and provide a basis for subsequent time correction. The movement path refers to the actual trajectory of the personnel target in space while obscured by the automated equipment. This can be obtained through continuous analysis of the RFID reader response data, combined with... Positioning techniques such as signal strength, time difference of arrival (TDOA), or angle of arrival (AOA) are used to determine the actual positional changes of a person during the occlusion period, providing a basis for subsequent spatial correction. Correcting the temporal component of the predicted recurrence time window involves adjusting the time range for the reappearance of the person based on the duration of the unexpected stationary state of the automated equipment. This can be done by adding the duration of the unexpected stationary state to the original predicted recurrence time range, or by recalculating the recurrence time point based on the time and duration of the stationary state. The aim is to ensure that the predicted recurrence time is consistent with the actual situation. Correcting the spatial component of the predicted recurrence time window involves adjusting the spatial range for the reappearance of the person based on the actual movement path of the person during the occlusion period. This can be done by superimposing the movement path of the person onto the original predicted recurrence spatial range, or by redetermining the recurrence location based on the endpoint of the movement path. The aim is to ensure that the predicted recurrence location is consistent with the actual situation.

[0039] In some preferred embodiments, when the actual movement of automated equipment deviates from the predetermined movement information, and the personnel target moves during the occlusion period, the calculation of the predicted recurrence spatiotemporal window for the personnel target can be specifically implemented as follows: First, in an industrial environment, such as a large automated warehouse, multiple ultra-high frequency (UHF) radio frequency readers can be deployed, evenly distributed at various key locations in the work area, such as aisles and shelving area entrances. When the RFID tag worn by personnel or on automated equipment enters the reader's reading range, the reader collects and uploads response data in real time. This data may include the tag's unique identifier, signal strength indication (RSSI) value, and a precise timestamp. This data is transmitted to a central data processing server for storage and preliminary processing. Then, the data processing server continuously analyzes this collected response data. For example, for automated equipment, the server can compare its current location information (obtained by parsing the response data of its RFID tag) with a pre-set movement plan (e.g., a predetermined path and speed obtained from a scheduling control system). If the RSSI value of an automated device remains stable for an extended period without expected changes, or if its position update frequency is significantly lower than expected, it can be determined that the device has experienced unexpected stagnation. Simultaneously, for obscured personnel targets, the server tracks changes in the RSSI value of their RFID tags and signal switching between different readers. Combining triangulation or fingerprint positioning algorithms, it estimates and records the continuous position points of the personnel target during the stagnation period in real time, thereby determining their movement path. Furthermore, if the system determines that an automated device has experienced unexpected stagnation, such as an AGV stopping at an intersection for 5 minutes, the system immediately obtains the duration of these 5 minutes. The previously predicted reappearance time window for the personnel target, for example, expected to reappear between 10 and 20 seconds after stagnation, will now be revised based on this 5-minute unexpected stagnation period. Specifically, the start and end times of the original predicted time window can be postponed by 5 minutes, making the new predicted reappearance time window between 5 minutes and 10 seconds and 5 minutes and 20 seconds, reflecting the impact of the automated device's stagnation on the personnel reappearance time. Finally, the system uses the determined movement path of the person during the occlusion period to correct the spatial components of the predicted reappearance spatiotemporal window. For example, if the person was located at point A before being occluded, the original predicted reappearance spatial window was an area near point A. However, during the occlusion period, RFID data revealed that the person actually moved to point B. Therefore, the system will adjust the center point or range of the predicted reappearance spatial window to near point B, rather than near point A, thus more accurately indicating the possible reappearance location of the person. Through these corrections, the system can more accurately predict the reappearance of the person, providing more precise guidance for subsequent video re-identification.

[0040] Optional, combined Figure 3 As shown, S2, based on video data, determines whether an event has occurred where a person target is obscured by automated equipment. The steps include: S21, In video data, the intersection-over-union ratio of the bounding boxes of people targets in consecutive image frames is analyzed; S22, when the intersection-union ratio of the bounding box of the person target in consecutive image frames is lower than a preset threshold, and the overlap between the last visible position of the person target and the bounding box of the automated device exceeds a preset occlusion overlap ratio, it is determined that an event has occurred in which the person target is occluded by the automated device.

[0041] The intersection-union ratio (IUGR) of bounding boxes for people in consecutive image frames refers to a quantitative assessment of the overlap between bounding boxes generated from two or more consecutive frames of images detected in a video stream. Specifically, it is the ratio of the intersection area to the union area of ​​two bounding boxes. Its purpose is to reflect changes in the position and shape of the person within a short period of time. A low IUGR may indicate that the person has moved rapidly, deformed, or been occluded. The preset threshold is a pre-set value used to determine whether the IUGR meets a specific condition. Specifically, it can be a critical value derived through statistical analysis of a large amount of real-world scene data, machine learning training, or expert experience. Its purpose is to distinguish between normal movement of the person and a critical state where occlusion or drastic changes may occur. The overlap between the last visible position of a person and the bounding box of an automated device refers to the degree of spatial overlap between the bounding box corresponding to the last clearly identifiable position of the person in the video before they disappear from the video frame or are completely occluded, and the bounding box corresponding to the current position of the automated device. Specifically, this can be quantified by calculating the ratio of the intersection area of ​​these two bounding boxes to the area of ​​the person's bounding box. Its purpose is to confirm whether the disappearance or change of the person is indeed related to the presence of the automated device, thereby ruling out other non-occlusion factors causing the visual target to disappear. The preset occlusion overlap ratio is a pre-set ratio used to determine whether the overlap between the person and the automated device meets the occlusion condition. Specifically, it can be a percentage or proportion determined through simulation or actual testing of different occlusion scenarios, combined with the system's sensitivity requirements for occlusion events. Its purpose is to ensure that only when there is a sufficient degree of visual occlusion between the person and the automated device is it judged as an occlusion event, avoiding false positives.

[0042] In some preferred embodiments, determining whether an event has occurred where a person target is occluded by automated equipment can be implemented as follows. First, the system uses a deep learning model (e.g., a real-time object detection model based on YOLOv8 or Faster R-CNN architecture) to analyze video data in the industrial environment frame by frame to identify and obtain the bounding box information of the person target and automated equipment in each frame. For consecutive video frames, such as the current frame and the previous frame, the system extracts the bounding box coordinates of the person target (e.g., [x1, y1, x2, y2]). Subsequently, the intersection-over-union (IoU) ratio of the bounding boxes of the person target in these two consecutive frames is calculated. The formula for calculating the intersection-over-union ratio is: IoU = (A∩B) / (A∪B), where A and B represent the areas of the two bounding boxes, respectively.

[0043] Specifically, when the calculated Intersection over Union (IoU) ratio is lower than a preset threshold (e.g., 0.3, which can be determined through training and validation on a large amount of normal motion and occlusion scene data), the system further examines the spatial relationship between the person target and the automated equipment. At this point, the system obtains the bounding box position of the person target that was last clearly detected before the IoU fell below the threshold, and also obtains the bounding box position of the automated equipment in the current frame. Next, it calculates the overlap between the bounding box of the last visible position of the person target and the bounding box of the automated equipment. The overlap can be defined as the ratio of the intersection area of ​​the two bounding boxes to the area of ​​the person target's bounding box. For example, if the person target's bounding box is P and the automated equipment's bounding box is M, the overlap can be calculated as (P∩M) / P. When this overlap exceeds a preset occlusion overlap ratio (e.g., 0.7, indicating that most of the person target's area is occluded by the automated equipment), the system determines that an event has occurred where the person target is occluded by the automated equipment. For example, if a worker suddenly disappears from a video, and their previous position highly overlaps with the bounding box of a moving AGV, and the intersection-union ratio (IU) of their bounding boxes decreases significantly before and after their disappearance, the system will accurately identify this as an occlusion event. This method effectively avoids misjudgments caused by non-occlusion situations such as a worker quickly moving out of the frame or sudden changes in lighting leading to target detection failure.

[0044] Optional, combined Figure 4 As shown, when the number of moving targets within the shielding area formed by the automated equipment is greater than one, step S42 analyzes the response data to determine whether the automated equipment has experienced an unexpected stationary state corresponding to a deviation between the actual movement and the predetermined movement information, and determines the movement path of the personnel target during the shielding period. This includes the following steps: S421, before the personnel target is obscured by automated equipment, acquire the reference response characteristics of the radio frequency tag corresponding to the personnel target; S422, Analyze the response data to determine whether the automated equipment has experienced unexpected stationary motion corresponding to a deviation between the actual motion and the predetermined motion information, and extract the movement path characteristics from the response data; S423, compare all movement path features with the baseline response features respectively, and identify the target movement path features that match the baseline response features; S424, the target movement path characteristics are determined as the movement path of the personnel target during the occlusion period.

[0045] Among them, the baseline response characteristics of RFID tags refer to the unique signal patterns or fingerprints generated by the RFID tags worn by personnel targets in a specific RFID reader network when they are not obstructed by automated equipment. Specifically, this may include, but is not limited to, combinations of parameters such as signal strength indication (RSSI), signal phase, time of arrival (TOA), and angle of arrival (AOA), as well as the trends of these parameters changing over time or location. Its purpose is to establish a reliable reference baseline for subsequent identification and comparison. Movement path characteristics refer to the RFID signal feature sequences of all possible moving targets (including personnel targets and other moving targets) parsed from the RFID reader response data within the area obstructed by automated equipment. Specifically, this can be a comprehensive description of the changes in signal strength, frequency, phase, etc. of multiple RFID tags over a period of time, reflecting the target's movement trajectory in space. Its purpose is to extract all potential movement trajectory information from complex response data, providing a data basis for subsequent screening. Target movement path characteristics refer to the RFID signal feature sequences of specific movement paths that are identified as matching the personnel target after comparison with the baseline response characteristics from all movement path characteristics. Its purpose is to accurately identify the actual movement trajectory of the personnel target from multiple moving targets.

[0046] In some preferred embodiments, specifically, when a worker wearing an RFID tag enters a narrow passage frequently traversed by automated forklifts, the system continuously collects data such as the signal strength, signal phase, and signal arrival time of the worker's RFID tag before the forklift obstructs the worker's path, storing this data as the worker's baseline response characteristics. When the forklift obstructs the worker, and other RFID-tagged mobile devices or goods are present in the passage, the RFID reader receives mixed response data from multiple RFID tags. At this point, the system analyzes this response data in real time, first determining if the forklift has come to an unexpected stop. Simultaneously, using signal processing algorithms, it parses and clusters this mixed response data into multiple independent movement path features, each representing a potential moving target. Subsequently, the system compares each of these parsed movement path features with the previously stored baseline response characteristics of the worker's RFID tag. The comparison process can employ methods such as signal pattern matching, correlation analysis, or machine learning classification. For example, if the signal strength change pattern and phase change trend of a certain movement path feature are highly consistent with the worker's baseline response characteristics, then this movement path feature will be identified as the target movement path feature matching the worker. Ultimately, the identified target movement path feature was determined to be the worker's actual movement path during the period when the forklift blocked it, thus providing an accurate basis for subsequent trajectory correction.

[0047] Optional, combined Figure 5 As shown, step S423, which compares all movement path features with the baseline response features to identify the target movement path features that match the baseline response features, includes: S4231, compare the real-time response features of the mobile path features with the baseline response features; S4232, During the comparison process, determine whether the state of the real-time response feature changes from a state that matches the benchmark response feature to a state that does not match, and obtain the matching state judgment result. S4233, if the matching state judgment result is yes, then analyze the response characteristic change process of the transition to the non-matching state, and determine whether the response characteristic change process is caused by the personnel target performing a specific operation action; S4234, If it is determined that the action was caused by a person performing a specific task, the transformed real-time response feature is set as the updated baseline response feature, and the updated baseline response feature is compared with the movement path feature. S4235, compare all movement path features with the baseline response features, or compare all movement path features with the updated baseline response features, and identify the movement path features that match as the target movement path features.

[0048] The real-time response characteristics of the movement path feature refer to the signal characteristics collected by the RFID reader from the RFID tag at a specific time point or time period while the person is obscured by automated equipment. Specifically, these can be the Received Signal Strength Indication (RSSI), phase, frequency offset, signal-to-noise ratio, or a combination thereof, reflecting the instantaneous RFID signal performance of the person within the obscured area. The baseline response characteristics refer to the signal characteristics collected and recorded by the RFID reader from the RFID tag in a stable state before the person is obscured by automated equipment. Specifically, these can be the average RSSI value, phase stability, or specific frequency response pattern of the RFID tag in an interference-free or low-interference environment before the person enters the obscured area, providing an initial reference standard for comparison and identification. The matching status judgment result refers to the judgment on the degree of similarity or difference between the real-time response characteristics and the baseline response characteristics by comparing them. Specifically, this can be based on a preset similarity threshold, judging whether the Euclidean distance, cosine similarity, or correlation coefficient between the real-time response characteristics and the baseline response characteristics are within an acceptable range, aiming to determine whether the real-time response characteristics still conform to the expected baseline pattern. The response characteristic change process refers to the trajectory or pattern of various parameters changing over time during the transition from a state matching the baseline response characteristic to a state of mismatch. Specifically, it can be the continuous rise, fall, or fluctuation trend of the real-time RSSI value, phase, or frequency within a short period of time. Its purpose is to provide a data basis for analyzing the causes of characteristic changes. Specific work actions refer to work-related behaviors of personnel targets within the obstruction area of ​​automated equipment that may cause significant changes in the response characteristics of their RFID tags. Specifically, it can be personnel targets picking up or putting down metal tools, putting on or removing protective equipment, changing body posture (e.g., squatting or standing), or operating equipment with electromagnetic interference. Its purpose is to distinguish response characteristic changes caused by personnel behavior from environmental interference. Updated baseline response characteristics refer to the real-time response characteristics that have changed after it is determined that the response characteristics have changed due to personnel targets performing specific work actions. Specifically, it can be the signal characteristics presented by the RFID tag in the new stable state after personnel targets perform specific work actions, such as the new average RSSI value or phase pattern after picking up metal tools. Its purpose is to adapt to changes in personnel target behavior and maintain the accuracy of subsequent comparisons.

[0049] In some preferred embodiments, this solution is implemented as follows. Assume an employee, E17, wearing an RFID tag in an automated warehouse. Before entering an area blocked by a large automated stacker crane, the system has acquired the tag's baseline response characteristics, such as an average received signal strength indication of -60 dBm and a stable phase of 180 degrees. After E17 enters the blocked area, the system continuously collects the tag's real-time response characteristics. At a certain moment, the system detects a change in E17's real-time response characteristics, such as a sudden drop in received signal strength indication to -75 dBm and significant phase fluctuations, which no longer match the baseline response characteristics. The system determines that the real-time response characteristics have changed from a match to a mismatch. At this point, the system analyzes this change in response characteristics. For example, by analyzing the rate, duration, and pattern of change in received signal strength indication and phase, the system can identify this as a typical signal attenuation and interference pattern caused by a person picking up a metal tool (e.g., a heavy wrench). If the system determines that this change is caused by E17 performing a specific task (picking up a metal tool), it sets the current changed real-time response characteristics (e.g., the new average received signal strength indication of -75dBm and fluctuating phase pattern) as the updated baseline response characteristics. Subsequently, the system continues to compare these updated baseline response characteristics with E17's subsequent movement path characteristics. Simultaneously, for other moving targets within the obstructed area that have not performed specific tasks, the system still uses their original baseline response characteristics for comparison. Ultimately, the system identifies all movement path characteristics that match either the original or updated baseline response characteristics and determines them as E17's movement path during the obstruction. In this way, even if E17 picks up or puts down a tool during obstruction, the system can continuously and accurately track its position.

[0050] Optional, combined Figure 6 As shown, the analysis in S4233 transforms into a response characteristic change process of a mismatched state. The steps to determine whether the response characteristic change process is caused by the personnel performing a specific task include: A1, select the real-time response features within the shielded area formed by the automated equipment, excluding those that change to a mismatched state, and record them as the response features of other movement paths; A2, Based on the response characteristics of other movement paths, determine the background change characteristics that characterize global disturbances; A3, extracting background change features from the response feature change process, as local change features; A4. Based on local change characteristics, determine whether the change process of response characteristics is caused by personnel performing specific work actions.

[0051] The response characteristic change process refers to the continuous numerical or pattern-wise change of the real-time response characteristic of a personnel target RFID tag over time, during the transition from a state matching the baseline response characteristic to a state of mismatch. This can be recorded and characterized using time series data, differential sequences, or trend analysis, with the aim of capturing the dynamic process leading to the mismatch. The response characteristics of other movement paths refer to the real-time response data generated by the RFID tags of all moving objects (e.g., other personnel, other automated equipment, or goods with RFID tags) within the shielded area formed by automated equipment, excluding the personnel target whose response characteristic change process is currently being analyzed. This can be recorded using indicators such as RF signal strength, phase, frequency offset, or signal-to-noise ratio, with the aim of providing a reference benchmark reflecting the overall changes in the current environment. The background change characteristics characterizing global disturbances refer to changes in environmental factors in an industrial environment caused by the actions of non-specific personnel targets that have a general impact on the response characteristics of all or most RFID tags, such as the movement of large equipment, fluctuations in the electromagnetic environment, or... Overall changes in temperature and humidity can be extracted from the response characteristics of other movement paths using methods such as statistical averaging, trend line fitting, or principal component analysis. The purpose is to quantify and identify these common environmental disturbances. Local change characteristics refer to the unique changes caused by the human target's own behavior (especially specific work actions) after removing or separating global disturbances (i.e., background change characteristics) from the response characteristics of the RFID tag. These can be obtained using methods such as difference calculation, residual analysis, or specific filter processing. The purpose is to highlight and isolate signal changes directly related to the human target's work actions. Specific work actions refer to standardized operations that the human target may perform during the period of obstruction, which will have a significant and identifiable impact on the response characteristics of the RFID tag it wears or carries. Examples include carrying heavy objects, operating handheld devices, changing body posture, or interacting with specific objects in the environment. These can be identified and judged using pre-trained models, expert rule bases, or feature pattern libraries built based on historical data. The purpose is to distinguish the differences in response characteristics caused by the human target's normal movement and the performance of work actions.

[0052] In some preferred embodiments, when the system detects that the real-time response characteristics of a personnel target RFID tag change from a matched state to a mismatched state, in order to accurately determine whether this change is caused by the personnel target performing a specific work action, the following specific implementation can be carried out: First, the system continuously monitors all active RFID tag signals within the shielded area formed by the automated equipment. When a mismatch occurs in the target personnel's tag signal, the system selects and records the real-time response characteristics of all other RFID tags besides the target personnel tag from these active signals, such as their Received Signal Strength Indication (RSSI) values, phase information, or frequency offsets, and uses this data as the response characteristics of other movement paths. This data can be organized into a time series, reflecting the dynamic changes of non-target tags in the area. Next, based on the response characteristics of these other movement paths, the system can determine the background change characteristics characterizing global disturbances. For example, the response characteristics of these other movement paths can be aggregated, such as calculating their moving average, median, or performing principal component analysis, to extract the prevalent environmental noise or interference patterns in the area. For example, if a large forklift starts in the area, its electromagnetic interference may cause a general decrease in the RSSI values ​​of all tags; this general downward trend is the background change characteristic. Subsequently, the system separates the previously determined background change features from the response characteristic changes of the target personnel's RFID tag, thus obtaining local change features. This can be achieved by calculating the difference between the real-time response changes of the target personnel's tag and the background change features, or by adaptive filtering, to eliminate the influence of environmental factors on the target tag signal, ensuring that the remaining signal changes reflect the personnel's own behavior. Finally, based on these separated local change features, the system determines whether the response characteristic change process is caused by the personnel performing a specific task. For example, a classification model, such as a support vector machine or neural network, can be pre-trained, taking the time series of local change features as input and outputting a judgment result on whether it is a specific task. Alternatively, a rule base can be established; for example, if the local change features exhibit fluctuations of a specific amplitude or a specific pattern (such as the periodic attenuation of the signal when carrying heavy objects) within a short period of time, it is determined to be a specific task. In this way, the system can accurately distinguish whether the change in the personnel's response characteristics is caused by environmental changes or their own task actions.

[0053] Optional, combined Figure 7 As shown, A2 determines the background change characteristics characterizing a global disturbance based on the response characteristics of other movement paths, including the following steps: A21, based on the historical response characteristics of each of the other mobile paths, determine the contribution index of each other mobile path; A22, based on the contribution index, weights the changes in response characteristics of other movement paths to generate background change characteristics.

[0054] Among them, the historical response characteristics of other mobile paths refer to the patterns or trends formed by the system continuously recording and analyzing the radio frequency signal response data of non-target mobile entities (such as other automated equipment, goods with RFID tags, etc.) over a period of time. Specifically, it can include a set of parameters that change over time, such as signal strength, signal arrival time, number of reads, and signal phase at different times and locations of these mobile paths. Its purpose is to provide a basis for assessing the potential impact of these non-target mobile entities on the overall radio frequency environment.

[0055] The contribution index refers to the numerical value that quantifies the impact of each non-target mobile entity on the overall changes in the radio frequency environment based on the historical response characteristics of other mobile paths. Specifically, it can be calculated by analyzing the stability, activity, interaction frequency with the radio frequency reader, or dwell time in a specific area of ​​the historical response characteristics. Its purpose is to distinguish the weight of different non-target mobile entities on the changes in the background signal, so as to give more reasonable consideration in subsequent processing.

[0056] Among them, background change characteristics refer to the comprehensive characteristics generated by weighting the response characteristic changes of other movement paths within the shielded area, excluding personnel targets, which can characterize the global or universal disturbances in the radio frequency environment. Specifically, it can be obtained by multiplying the response characteristic changes of each other movement path by its corresponding contribution index and then accumulating or averaging them. Its purpose is to accurately reflect the interference that is prevalent in the environment, so as to effectively separate the local changes caused by the specific operation actions of personnel targets from their response characteristics.

[0057] Optional, combined Figure 8 As shown, the steps by which A21 determines the contribution index of each other mobile path based on their respective historical response characteristics include: A211, Identify change points in the time series of historical response characteristics for each of the other movement paths; A212, Calculate the dispersion statistics based on the time series data after this change point; A213, If the change point is not identified, the dispersion statistic is calculated based on the latest time series data preset in the historical response characteristics; A214, based on this dispersion statistic, determine the contribution index.

[0058] In this context, a change point refers to a point in time series data where the characteristics of the data undergo a significant change. This can be identified using statistical or machine learning methods. Dispersion statistics are statistical measures of the degree of fluctuation or dispersion of a set of data. These can be implemented using statistical indicators such as variance, standard deviation, mean absolute deviation, and interquartile range. The preset latest time series data refers to the most recent set of data selected from historical response characteristics to reflect current data characteristics when change points have not been identified. This can be determined using methods such as a fixed time window, a fixed number of data points, or a moving average. The contribution index is a numerical value used to quantify the impact of each other movement path on global disturbances or background changes. This can be implemented using the reciprocal of the dispersion statistics, the normalized result of the dispersion statistics, or a weighted value obtained by mapping the dispersion statistics through a preset function.

[0059] In some preferred embodiments, the specific process for determining the contribution index of each other movement path can be as follows. First, for each other movement path, such as a stationary shelf or a continuously moving AGV, the system continuously collects its radio frequency response (RF) characteristics and forms a time series. To identify whether there are significant changes in this time series, the Cumulative Sum Control Chart (CUSUM) algorithm can be used. This algorithm identifies a change point by calculating the cumulative deviation between the response characteristic value and the target mean, and when the cumulative deviation exceeds a preset threshold. For example, if the response characteristic of a shelf has been stable around a certain RSSI value for a period of time, but suddenly its RSSI value begins to deviate continuously due to environmental interference or the deployment of new equipment nearby, the CUSUM algorithm can detect this change in time and mark the change point.

[0060] Secondly, once a change point is identified, the system calculates its dispersion statistics based on the time series data following that change point. For example, it can calculate the standard deviation of this data. The standard deviation directly reflects the degree of data fluctuation; the larger the standard deviation, the more unstable the response characteristics of that movement path. If no change point is identified, the system calculates its standard deviation based on the latest time series data preset in the historical response characteristics, such as the latest response characteristic data for a preset duration.

[0061] Finally, based on the calculated dispersion statistics, a contribution index can be determined. For example, the standard deviation can be normalized, and its reciprocal can be used as the contribution index. Specifically, if the standard deviation of the response characteristics of a certain movement path is σ, then its contribution index can be calculated as 1 / (1+σ). Thus, the smaller the standard deviation, the higher the contribution index, indicating that the response characteristics of the movement path contribute more reliably to the background changes; conversely, the larger the standard deviation, the lower the contribution index. In this way, the system can dynamically evaluate the stability of each non-target movement path and adjust its weight in the generation of background change characteristics accordingly, thereby more accurately separating the response characteristic changes caused by specific operational actions of personnel targets.

[0062] Optional, combined Figure 9 As shown, if an event occurs in S3 where a person target is obscured by automated equipment, the steps for obtaining the predetermined motion information of the automated equipment include: S31, based on the video frame corresponding to the event where a person is obscured by an automated device, identify the identification number of the automated device, and send a query request containing the camera number, timestamp, obstruction location and device number to the scheduling and control system of the automated device; S32, after receiving the query request, the dispatch control system queries the task database and returns the scheduled motion information of the automated equipment.

[0063] The identification number of the automated equipment refers to the code used to uniquely identify the automated equipment, which may include serial number, asset number, or internal identification code, etc., and its purpose is to accurately identify the automated equipment that has experienced an occlusion event; the scheduling and control system refers to the central control platform responsible for managing and coordinating the operation of the automated equipment, which can be a centralized software system or a distributed control network, and its purpose is to receive query requests and provide equipment operation data; the task database refers to the collection that stores the various task plans, operating status, and historical trajectory data of the automated equipment, which can be a relational database, a non-relational database, or a distributed file system, and its purpose is to provide the data source required for the scheduling and control system to query; the planned motion information refers to the data such as the motion trajectory, speed, stopping point, and task target that the automated equipment is scheduled to execute within a specific time period, which may include path coordinate sequence, timestamp, speed curve, or task instruction set, and its purpose is to describe the expected behavior of the automated equipment during the occlusion period.

[0064] In some preferred embodiments, when the system detects that an employee's silhouette is completely obscured by a moving AGV, a process for acquiring the AGV's predetermined motion information is triggered. Specifically, the system immediately analyzes the video frame that caused the obscuration event, using a pre-trained image recognition model to identify the AGV's unique identification number from its appearance features (e.g., vehicle color, specific markings, or the shape of the lidar on the roof), such as "AGV-007". Simultaneously, the system records the corresponding camera number (e.g., "Cam-A3"), the precise timestamp (e.g., "2023-10-26 14:35:12.345"), and the obscuration location information (e.g., the center point of the obscured area expressed in pixel coordinates or relative scene coordinates) where the last visible position of the employee coincides with the AGV's bounding box. Subsequently, the system encapsulates this information into a query request and sends it to the AGV's central scheduling and control system via a network interface (e.g., TCP / IP protocol). This scheduling and control system can be deployed on a separate server cluster and maintains a distributed task database containing all AGV task plans and real-time statuses. Upon receiving a query request, the dispatch control system quickly retrieves relevant task records for that AGV within a specific time period from the task database, based on the AGV's identification number, timestamp, and obstruction location provided in the request. For example, it might find that AGV-007 is scheduled to perform a transport task from "Area A Shelf 12" to "Area B Loading / Unloading Port" between 14:35:00 and 14:36:00, accompanied by a detailed path point sequence, preset speed curve, and estimated arrival time. The dispatch control system packages this detailed pre-planned motion information (e.g., a series of timestamped path coordinates and corresponding speed values) and returns it to the personnel trajectory positioning and tracking system via the network. In this way, the system can obtain the precise movement trajectory of the AGV during the obstruction period, thus providing an accurate data foundation for subsequent personnel trajectory prediction and tracking.

[0065] A personnel trajectory positioning and tracking system based on RFID and AI video recognition is used to perform personnel trajectory positioning and tracking based on RFID and AI video recognition. The personnel trajectory positioning and tracking system 1 based on RFID and AI video recognition includes: Video data acquisition module 11 is used to acquire video data in an industrial environment; The occlusion event judgment module 12 is used to determine, based on video data, whether an event occurs in which a person target is occluded by automated equipment; The motion information acquisition module 13 is used to acquire the predetermined motion information of the automated equipment if an event occurs in which a person target is obscured by the automated equipment. The reproduction window calculation module 14 is used to calculate the predicted reproduction spatiotemporal window of the person target based on the predetermined motion information and the state of the person target before it is occluded. The personnel location and tracking module 15 is used to process video data within the predicted reproduction time and space window, search for and re-identify personnel targets, so as to realize personnel trajectory location and tracking.

[0066] The video data acquisition module refers to the hardware or software unit responsible for collecting visual information from the industrial environment. It can be implemented using devices such as high-definition network cameras, industrial-grade video capture cards, or video streaming servers. Its purpose is to provide raw data input for subsequent video analysis. The occlusion event judgment module is a logical processing unit used to analyze video data to identify situations where personnel targets are occluded by automated equipment. It can be implemented using software modules based on image processing and computer vision algorithms, such as by analyzing the intersection-union ratio and overlap of the target bounding box. Its purpose is to promptly detect critical events that may interrupt personnel tracking. The motion information acquisition module refers to the interface or processing unit used to acquire the predetermined motion trajectory and status data of relevant automated equipment when an occlusion event occurs. It can be implemented by a software interface or protocol converter that communicates with the scheduling and control system of automated equipment. Its purpose is to provide key motion parameters for predicting the reproducibility of personnel positions. The reappearance window calculation module refers to a calculation unit that predicts the spatiotemporal range in which a person may reappear based on known motion information and the state of the person before being occluded. It can be implemented using a software module that integrates prediction algorithms and state estimation models. Its purpose is to narrow down the search range for subsequent people and improve the efficiency of re-identification. The personnel location and tracking module refers to a comprehensive processing unit that performs deep processing on video data within a predicted spatiotemporal range to search for and re-identify personnel targets. It can be implemented using a software system that combines deep learning models and multi-target tracking algorithms. Its purpose is to quickly and accurately restore continuous tracking of personnel after they are occluded.

[0067] In some preferred embodiments, this application is implemented as follows: The video data acquisition module can consist of multiple high-resolution IP cameras deployed in the industrial site. These cameras are connected to a central video server via Ethernet. The server is responsible for receiving, storing, and preprocessing the video stream, such as performing frame synchronization and preliminary encoding. The occlusion event judgment module can be a software application running on a high-performance computing unit (e.g., an industrial PC equipped with a GPU). It integrates a deep learning-based object detection model (such as YOLO or Faster R-CNN) to identify personnel targets and automated equipment in the video in real time. It determines the occurrence of occlusion events by calculating the intersection-over-union ratio and overlap degree of the bounding boxes of personnel targets and automated equipment. The motion information acquisition module can be a software interface that communicates with the factory's Manufacturing Execution System (MES) or automated equipment scheduling system via ModbusTCP or OPCUA protocol. When an occlusion event is triggered by the judgment module, this interface sends a query request to the scheduling system to obtain the current task plan, preset path, and speed information of a specific automated equipment (such as an AGV or robotic arm). The replay window calculation module can be an algorithm service running on the same high-performance computing unit. It receives device motion data from the motion information acquisition module and the last visible position and velocity information of the personnel target from the occlusion event judgment module. Using state estimation algorithms such as Kalman filtering or particle filtering, combined with the device motion model, it predicts the spatial region and time point where the personnel target may appear in the future, forming a three-dimensional spatiotemporal prediction window. The personnel positioning and tracking module can be a software component integrating multi-target tracking algorithms (such as DeepSORT or ByteTrack) and re-identification models (such as feature matching based on ReID networks). Within the spatiotemporal range provided by the replay window calculation module, it prioritizes processing specific regions of the video stream, searches for visual entities with features similar to the occluded personnel target, and performs identity reconfirmation, thereby recovering and continuing the personnel trajectory tracking. Through the configuration of these specific components, the system can achieve continuous and accurate trajectory tracking of personnel in complex industrial environments, even with frequent occlusion.

[0068] Through the above technical solution, this application provides a personnel trajectory positioning and tracking system based on RFID and AI video recognition. This system effectively solves the problem of traditional methods struggling to coordinate the various functional modules in practical applications by modularizing and organically integrating functions such as video data acquisition, occlusion event judgment, motion information acquisition, replay window calculation, and personnel positioning and tracking. This systematic design makes data flow and processing smoother, enabling efficient information transmission and response between modules, thereby improving the efficiency and accuracy of personnel trajectory positioning and tracking. Especially in complex scenarios in industrial environments where personnel targets are frequently obscured by automated equipment, the system can maintain continuous tracking of personnel, improving tracking reliability.

[0069] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for locating and tracking personnel trajectories based on RFID and AI video recognition, characterized in that, include: Acquire video data in industrial environments; Based on video data, determine whether there are events where people are obscured by automated equipment; If an event occurs in which a person is obscured by automated equipment, then the predetermined motion information of the automated equipment is obtained; Based on the predetermined motion information and the state of the person target before being occluded, the predicted spatiotemporal window for the person target is calculated. Within the predicted spatiotemporal window, the video data is processed to search for and re-identify personnel targets in order to achieve personnel trajectory positioning and tracking.

2. The method for personnel trajectory positioning and tracking based on RFID and AI video recognition according to claim 1, characterized in that, When the actual movement of the automated equipment deviates from the predetermined movement information, and the person target moves during the occlusion period, the step of calculating the predicted recurrence spatiotemporal window of the person target based on the predetermined movement information and the state of the person target before being occluded includes: Collect response data from various RFID readers covering the work area in an industrial environment; Analyze the response data to determine whether the automated equipment has experienced an unexpected stationary state corresponding to a deviation between the actual movement and the predetermined movement information, and determine the movement path of the personnel target during the occlusion period; If the unexpected stillness occurs, the time component of the predicted recurrence time-space window is corrected based on the duration of the unexpected stillness. Based on the movement path of the personnel target, the spatial components of the predicted spatiotemporal window are corrected.

3. The method for personnel trajectory positioning and tracking based on RFID and AI video recognition according to claim 1, characterized in that, The step of determining whether an event has occurred where a person target is obscured by automated equipment based on video data includes: In video data, the intersection-union ratio (IUU) of the bounding boxes of people in consecutive image frames is used for analysis; When the intersection-over-union ratio of the bounding boxes of a person target in consecutive image frames is lower than a preset threshold, and the overlap between the last visible position of the person target and the bounding box of the automated device exceeds a preset occlusion overlap ratio, it is determined that an event has occurred in which the person target is occluded by the automated device.

4. The method for personnel trajectory positioning and tracking based on RFID and AI video recognition according to claim 2, characterized in that, When the number of moving targets within the shielding area formed by the automated equipment is greater than one, the step of analyzing the response data, determining whether the automated equipment has experienced an unexpected stationary state corresponding to a deviation between the actual movement and the predetermined movement information, and determining the movement path of the personnel target during the shielding period includes: Before a person is obscured by automated equipment, acquire the baseline response characteristics of the RFID tag corresponding to the person. Analyze the response data to determine whether the automated equipment has experienced an unexpected stationary state corresponding to a deviation between the actual movement and the predetermined movement information, and parse the movement path characteristics from the response data; All of the movement path features are compared with the baseline response features to identify the target movement path features that match the baseline response features; The target movement path features are determined as the movement path of the personnel target during the occlusion period.

5. A method for locating and tracking personnel based on RFID and AI video recognition according to claim 4, characterized in that, The step of comparing all the movement path features with the baseline response features to identify the target movement path features that match the baseline response features includes: The real-time response features of the mobile path features are compared with the baseline response features; During the comparison process, it is determined whether the state of the real-time response feature changes from a state that matches the benchmark response feature to a state that does not match, and a matching state judgment result is obtained. If the matching status judgment result is yes, then analyze the response characteristic change process of the transition to the non-matching status, and determine whether the response characteristic change process is caused by the personnel target performing a specific operation action; If it is determined that the cause is the execution of a specific operation by a personnel target, the transformed real-time response feature is set as the updated baseline response feature, and the updated baseline response feature is compared with the movement path feature. All of the aforementioned movement path features are compared with the baseline response features, or all of the aforementioned movement path features are compared with the updated baseline response features. The movement path features that are identified as matching are determined as the target movement path features.

6. A method for locating and tracking personnel based on RFID and AI video recognition according to claim 5, characterized in that, The analysis of the response characteristic change process in the transition to a mismatched state, and the step of determining whether the response characteristic change process is caused by the execution of a specific task action by a person, includes: Within the shielded area formed by the automated equipment, the real-time response characteristics, excluding those that transition to a mismatched state, are selected and recorded as the response characteristics of other movement paths. Based on the response characteristics of other movement paths, the background change characteristics characterizing global disturbances are determined; The background change features are separated from the response feature change process and used as local change features; Based on the aforementioned local change characteristics, it is determined whether the change in response characteristics is caused by the execution of specific operational actions by the personnel target.

7. A method for locating and tracking personnel based on RFID and AI video recognition according to claim 6, characterized in that, The step of determining the background change characteristics characterizing the global disturbance based on the response characteristics of other movement paths includes: Based on the historical response characteristics of each of the other mobile paths, the contribution index of each other mobile path is determined. Based on the contribution index, the changes in response characteristics of other movement paths are weighted to generate background change characteristics.

8. A method for locating and tracking personnel based on RFID and AI video recognition according to claim 7, characterized in that, The step of determining the contribution index of each other mobile path based on their respective historical response characteristics includes: Identify change points in the time series of historical response characteristics for each of the other movement paths; Calculate the dispersion statistics based on the time series data after the change point; If the change point is not identified, the dispersion statistic is calculated based on the latest time series data preset in the historical response features. The contribution index is determined based on the aforementioned dispersion statistics.

9. A method for locating and tracking personnel based on RFID and AI video recognition according to claim 1, characterized in that, The step of obtaining predetermined motion information of the automated equipment if an event occurs in which a person target is obscured by the automated equipment includes: Based on the video frame corresponding to the event where a person is obscured by automated equipment, the identification number of the automated equipment is identified, and a query request containing the camera number, timestamp, obstruction location and equipment number is sent to the scheduling and control system of the automated equipment. After receiving the query request, the scheduling and control system queries the task database and returns the scheduled motion information of the automated equipment.

10. A personnel trajectory positioning and tracking system based on RFID and AI video recognition, used to perform personnel trajectory positioning and tracking based on RFID and AI video recognition, characterized in that, include: The video data acquisition module is used to acquire video data in industrial environments. The occlusion event detection module is used to determine, based on video data, whether an event has occurred in which a person or target is occluded by automated equipment. The motion information acquisition module is used to acquire the predetermined motion information of the automated equipment if an event occurs in which a person target is obscured by the automated equipment. The reproduction window calculation module is used to calculate the predicted reproduction spatiotemporal window of the person target based on the predetermined motion information and the state of the person target before being occluded. The personnel location and tracking module is used to process the video data within the predicted reproduction time window, search for and re-identify personnel targets, so as to realize personnel trajectory location and tracking.

Images