An unmanned aerial vehicle tracking method and system based on remote identification information

Abstract

The application discloses a kind of unmanned plane connection tracking method and system based on remote identification information, it is related to unmanned plane supervision technical field.The method includes: parsing the Remote ID message of target unmanned plane, obtains target state information;According to target state information generates prediction state sequence;Based on prediction state sequence, candidate linkage camera is filtered, and target linkage camera is determined;According to the current PTZ state of target linkage camera and target control parameter, generate head control instruction, and when meeting switching condition, execute camera connection switching.The system includes Remote ID receiving equipment, edge computing gateway, camera resource library, multiple linkage cameras and video linkage server.The application can realize the automatic, continuous video linkage tracking of target unmanned plane, improve the continuity, accuracy and response efficiency of target tracking.

Description

Technical Field

[0001] This invention relates to the field of drone monitoring technology, and in particular to a method and system for continuous drone tracking based on remote identification information. Background Technology

[0002] With the increasing application of drones in scenarios such as inspection, security, surveying, and emergency response, the demand for the detection, identification, tracking, and evidence collection of low-altitude targets is becoming increasingly prominent. In drone monitoring scenarios, it is necessary to not only grasp the identity and status information of the target drone, but also obtain continuous video footage of the target drone to meet the needs of real-time monitoring and subsequent traceability.

[0003] In existing technologies, one type of solution is mainly used to obtain the identity and location information of drone targets, enabling the detection and identification of drone targets; the other type of solution mainly relies on video surveillance equipment to observe and record targets, providing target image information. However, in practical applications, the former is usually difficult to directly form continuous visual tracking of target drones, while the latter is easily affected by factors such as camera coverage, target movement speed, scene changes, and equipment response capabilities, making it difficult to continuously and stably track drone targets.

[0004] Especially in scenarios where the target drone is continuously flying, its position is constantly changing, or it involves multiple camera coverage areas, existing technologies generally suffer from the following problems: difficulty in timely identifying the most suitable camera for the tracking task, difficulty in achieving smooth and continuous tracking transitions between multiple cameras, and a tendency for tracking interruptions, switching lags, or unstable control. Therefore, how to achieve continuous, stable, and automated video-linked tracking of target drones in drone surveillance scenarios has become a pressing technical problem to be solved in this field. Summary of the Invention

[0005] In existing technologies, it is difficult to effectively combine UAV target state perception with continuous video tracking, and in multi-camera scenarios, problems such as untimely target linkage camera identification, discontinuous camera switching, and insufficient tracking control stability easily arise. Therefore, this invention provides a UAV continuous tracking method and system based on remote identification information to achieve continuous, stable, and automated video linkage tracking of target UAVs.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] A method for continuous tracking of unmanned aerial vehicles (UAVs) based on remote identification information includes the following steps:

[0008] S1. Receive and parse the Remote ID message broadcast by the target UAV to obtain the target status information of the target UAV at multiple consecutive moments;

[0009] S2. Based on the target state information, generate a sequence of predicted states for multiple discrete prediction times within a preset prediction window in the future.

[0010] S3. Based on the predicted state sequence, retrieve candidate linked cameras, and for each candidate linked camera, determine the target control parameters and whether the visibility conditions are met at each discrete prediction time in the future.

[0011] S4. Based on the ratio of the number of times that meet the visibility conditions to the total number of discrete prediction times in each future discrete prediction time, determine the continuous observation capability parameters of each candidate linked camera; based on the maximum values ​​of the horizontal rotation adjustment, pitch adjustment, and zoom adjustment amounts corresponding to the target control parameters when each candidate linked camera is adjusted from the current PTZ state to the target control parameters, normalized by the corresponding maximum adjustment speed and the preset prediction window duration, determine the PTZ takeover cost parameters of each candidate linked camera; based on the continuous observation capability parameters and PTZ takeover cost parameters, determine the comprehensive score of each candidate linked camera, and determine the target linked camera based on the comprehensive score;

[0012] S5. Generate PTZ control commands based on the current PTZ status of the target-linked camera and the target control parameters, and send them to the target-linked camera;

[0013] S6. When the continuous observation capability parameter of the current target linkage camera is lower than the preset observation threshold, select the camera whose comprehensive score is not less than the preset score gain threshold from the candidate linkage cameras. Determine the camera with the highest comprehensive score in the selection results as the updated target linkage camera and send the corresponding PTZ control command to the updated target linkage camera.

[0014] Furthermore, the Remote ID message broadcast by the target drone is received and parsed through the following processing:

[0015] The Remote ID message actively broadcast by the target drone is received by the Remote ID receiving device, and the Remote ID message is decoded by the protocol at the edge computing gateway.

[0016] Based on the unique UAV identifier in the Remote ID message, multiple consecutive broadcast period messages belonging to the same target UAV are time-sequentially cached and associated to form an associated message sequence;

[0017] Longitude, latitude, altitude, speed, heading, and time parameters are extracted from the associated message sequence to form continuous target state information that can be directly called upon for subsequent prediction state sequence construction.

[0018] Furthermore, the predicted state sequence is obtained by extrapolating the short-time trajectory of the target UAV based on the target state information at multiple consecutive time points. The sequence of predicted states corresponding to each discrete prediction time. This is the number of discrete prediction times within the preset prediction window.

[0019] Furthermore, the candidate linked cameras are retrieved in the following manner:

[0020] The camera resource library is spatially proxied by using a three-dimensional spatial index to obtain a preset number of linked cameras that are closest to the target drone as initial candidate linked cameras.

[0021] For each initial candidate linkage camera, the following checks are performed in sequence: online status check, signal quality check, maximum effective viewing distance check, pitch angle range check, horizontal rotation range check, maximum available zoom range check, maximum horizontal rotation adjustment speed check, maximum pitch adjustment speed check, and maximum zoom adjustment speed check. Cameras that meet the linkage conditions are retained as candidate linkage cameras.

[0022] The three-dimensional spatial index is implemented based on the spatial database index.

[0023] Furthermore, for each candidate linked camera, the target control parameters are determined at each discrete prediction time in the future in the following manner:

[0024] For each candidate linked camera, the geographic coordinate parameters corresponding to the predicted state sequence are converted into local coordinate parameters with the candidate linked camera as the reference coordinate system, and the target azimuth angle, target elevation angle and target imaging magnification required for each discrete prediction time in the future are determined according to the local coordinate parameters.

[0025] The target control parameters include the target azimuth angle, the target elevation angle, and the required magnification for target imaging.

[0026] Furthermore, the process of determining the target control parameters and whether the visibility conditions are met for each candidate linked camera at each discrete prediction time in the future includes:

[0027] For the first The first candidate linked camera, in the future... At the discrete prediction time, the target UAV and the... Spatial distance between candidate linked cameras satisfy:

[0028] ;

[0029] In the formula, This serves as an index for candidate linked cameras. For the index of discrete prediction time, The target UAV calculated based on Vincenty's formula and the first The candidate linked camera in the first The ellipsoidal distance between discrete prediction times. For the target drone in the The height at each discrete prediction time. For the first The mounting height of each candidate linked camera; when determining spatial distance. Not greater than the The maximum effective line-of-sight distance of the candidate linked cameras, the target pitch angle is located at the th Within the pitch angle range of the candidate linked cameras, the required zoom level for target imaging is no greater than that of the first camera. When determining the maximum available zoom level of the candidate linked cameras, the first... The candidate linked camera in the first Each discrete prediction time satisfies the visibility condition; when calculating the ellipsoidal distance, the intermediate calculation results corresponding to the repeatedly called candidate linkage camera position parameters and the target UAV predicted position parameters are cached and reused.

[0030] Furthermore, the PTZ takeover cost parameters for each candidate linked camera are determined as follows:

[0031] The normalized values ​​for horizontal rotation, pitch, and magnification are obtained by dividing the horizontal rotation adjustment amount corresponding to the target azimuth angle by the product of the maximum horizontal rotation adjustment speed and the preset prediction window duration, the pitch adjustment amount corresponding to the target pitch angle by the product of the maximum pitch adjustment speed and the preset prediction window duration, and the magnification adjustment amount corresponding to the required magnification amount for target imaging by the product of the maximum magnification adjustment speed and the preset prediction window duration.

[0032] The maximum value among the normalized values ​​of horizontal rotation, pitch, and zoom is taken as the PTZ takeover cost parameter for the i-th candidate linked camera. ;

[0033] The preset prediction window duration is the time span corresponding to the preset prediction window.

[0034] Furthermore, the determination of the comprehensive score for each candidate linked camera based on the continuous observation capability parameter and the PTZ takeover cost parameter includes: No. Overall score of candidate linked cameras satisfy:

[0035] ;

[0036] In the formula, This serves as an index for candidate linked cameras. For the first Parameters of the continuous observation capability of each candidate linked camera. For the first PTZ takeover cost parameters for each candidate linked camera For the first Signal quality parameters of the candidate linked cameras For the first Mounting height adaptation parameters for each candidate linked camera For the first Zoom adaptation parameters for each candidate linked camera For the first Cruise status parameters of each candidate linked camera to These are the weights for the corresponding parameters;

[0037] The , , and Each by the first The signal quality of each candidate camera, the degree of matching between its mounting height and the target height, the degree of matching between its available zoom capability and the zoom required for target imaging, and its cruise status are obtained after normalization, and satisfy the following conditions:

[0038] ;

[0039] In the formula, Statement No. One weight parameter, The index of the weight parameter is , and ;

[0040] Based on the comprehensive score of each candidate linkage camera Sort the cameras in descending order and identify the candidate camera at the top of the sorting results as the target camera.

[0041] Furthermore, the pan-tilt control command generated based on the current PTZ state of the target-linked camera and the target control parameters is generated in the following manner:

[0042] The gimbal control commands include horizontal rotation control commands, pitch rotation control commands, and zoom control commands;

[0043] Based on the differences between the target azimuth angle and the current horizontal rotation angle, the differences between the target elevation angle and the current elevation angle, and the differences between the required magnification for target imaging and the current magnification, the horizontal rotation error, elevation error, and magnification error are constructed respectively.

[0044] PID calculations are performed on the horizontal rotation error, pitch error, and zoom error respectively to obtain horizontal rotation control commands, pitch rotation control commands, and zoom control commands. After limiting the amplitude and adapting the control protocol for the horizontal rotation control commands, pitch rotation control commands, and zoom control commands, the device control interface is called to execute them.

[0045] Before sending the PTZ control command, control the target-linked camera to pause patrol; after releasing the tracking control of the target-linked camera, control the target-linked camera to resume patrol.

[0046] The present invention also provides a drone follow-up tracking system based on remote identification information, comprising:

[0047] Remote ID receiving device, used to receive Remote ID messages broadcast by the target drone;

[0048] An edge computing gateway, connected to the Remote ID receiving device, is used to perform protocol decoding, timing caching, and association of the Remote ID messages to form target status information of the target drone at multiple consecutive moments.

[0049] The camera resource library is used to store the installation location, mounting height, maximum effective viewing distance, pitch angle range, horizontal rotation range, maximum available zoom, maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, maximum zoom adjustment speed, online status, signal quality, and cruise status of linked cameras.

[0050] Multiple linked cameras are used to receive PTZ control commands and perform corresponding horizontal rotation, pitch and zoom adjustments, and report the current PTZ status to the video linkage server;

[0051] The video linkage server, connected to the edge computing gateway, camera resource library, and multiple linkage cameras, includes:

[0052] The prediction module is used to generate a sequence of predicted states for multiple discrete prediction times within a preset prediction window based on the target state information.

[0053] The candidate filtering module is used to retrieve candidate linked cameras based on the predicted state sequence;

[0054] The parameter calculation module is used to determine the target control parameters for each candidate linked camera and to determine whether the visibility conditions are met.

[0055] The scoring and decision module is used to determine the continuous observation capability parameters, PTZ takeover cost parameters, and comprehensive scores of each candidate linked camera, and to determine the target linked camera based on the comprehensive score.

[0056] The control transmission module is used to generate PTZ control commands based on the current PTZ status of the target linkage camera and the target control parameters, and then send them to the target linkage camera.

[0057] The switching module is used to update the target linked camera when the continuous observation capability parameter of the current target linked camera is lower than the preset observation threshold, and to send the corresponding pan-tilt control command to the updated target linked camera.

[0058] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention forms continuous target state information by receiving and parsing the Remote ID message broadcast by the target UAV, and on this basis generates a predicted state sequence for multiple discrete prediction times within a future preset prediction window. This makes the selection of the linkage camera no longer limited to the static judgment at the current time, which is conducive to improving the accuracy and timeliness of the target linkage camera selection. By judging the visibility conditions of the candidate linkage cameras at each discrete prediction time in the future, and combining the continuous observation capability parameters and PTZ takeover cost parameters to determine the comprehensive score, it can take into account both continuous observation capability and equipment takeover cost, and reduce the risk of tracking interruption caused by improper selection of linkage cameras. The present invention generates PTZ control commands based on the current PTZ state of the target linkage camera and the target control parameters, and performs a succession switch when the continuous observation capability of the current target linkage camera decreases, thereby improving the continuity, stability and response speed of the target UAV video linkage tracking. Attached Figure Description

[0059] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the modular structure of the system of the present invention. Detailed Implementation

[0060] The technical solution of the present invention will be further described below with reference to the accompanying drawings. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Equivalent substitutions or modifications made by those skilled in the art to the technical features therein without departing from the technical concept of the present invention should all fall within the scope of protection of the present invention.

[0061] In one embodiment, such as Figure 1 As shown, this invention provides a method for continuous tracking of unmanned aerial vehicles (UAVs) based on remote identification information. This method mainly includes processes such as Remote ID message reception and parsing, predicted state sequence generation, candidate linked camera retrieval, target control parameter determination, visibility condition judgment, continuous observation capability parameter determination, PTZ takeover cost parameter determination, comprehensive scoring, target linked camera determination, gimbal control command generation, and continuous tracking switching.

[0062] S1. Receive and parse the Remote ID message broadcast by the target UAV to obtain the target status information of the target UAV at multiple consecutive times.

[0063] Specifically, the receiving and parsing of the Remote ID message broadcast by the target UAV is achieved through the following processing: receiving the Remote ID message actively broadcast by the target UAV through a Remote ID receiving device, and performing protocol decoding on the Remote ID message at the edge computing gateway; based on the unique UAV identifier in the Remote ID message, performing time-series caching and association on multiple consecutive broadcast period messages belonging to the same target UAV; extracting longitude, latitude, altitude, speed parameters, heading parameters, and time parameters from the associated message sequence to form continuous target state information that can be directly called for subsequent prediction state sequence construction.

[0064] In this embodiment, the Remote ID receiving device receives Remote ID messages actively broadcast by the target drone during flight. An edge computing gateway is connected to the Remote ID receiving device and performs protocol decoding and data processing on the Remote ID messages. By combining Remote ID message reception with edge-side decoding, target state data can be continuously acquired during the drone's continuous flight, providing a foundation for constructing subsequent predicted state sequences.

[0065] Upon receiving the Remote ID message, the edge computing gateway decodes the message according to the data format corresponding to the Remote ID message, parsing the original broadcast message into computable structured data. Subsequently, based on the unique UAV identifier in the message, the edge computing gateway merges multiple consecutive broadcast period messages belonging to the same target UAV, and performs time-series caching and association according to time order, thereby forming a continuous message sequence belonging to the same target UAV.

[0066] After forming a continuous message sequence, the edge computing gateway extracts longitude, latitude, altitude, velocity, heading, and time parameters from the sequence. Longitude, latitude, and altitude characterize the spatial location of the target UAV at each moment; velocity characterizes the UAV's motion state at the corresponding moment; heading characterizes the UAV's flight direction; and time parameters characterize the temporal relationship between the state data at different moments. These extracted parameters constitute the target UAV's state information across multiple consecutive moments.

[0067] In one embodiment, the target state information is organized and stored in chronological order, enabling subsequent steps to directly read state data from multiple adjacent time points to construct a predicted state sequence corresponding to multiple discrete prediction times within a preset prediction window. That is, the continuous target state information output in step S1 serves as the input data source for step S2.

[0068] In some embodiments, to improve the availability of continuous target state information, the edge computing gateway can also perform basic data processing on the received Remote ID messages. This basic data processing may include deduplicating duplicate messages, rearranging messages with time-series anomalies, and removing or marking messages with missing fields. Through these processes, the temporal consistency and data reliability of continuous target state information can be improved.

[0069] After completing step S1 and obtaining the target state information of the target UAV at multiple consecutive time points, proceed to step S2.

[0070] S2. Based on the target state information, generate a sequence of predicted states for multiple discrete prediction times within a preset prediction window in the future.

[0071] Specifically, the predicted state sequence is the future state obtained by short-time trajectory extrapolation based on the target UAV's state information at multiple consecutive time points. The sequence of predicted states corresponding to each discrete prediction time. This is the number of discrete prediction times within the preset prediction window.

[0072] In this embodiment, the target state information output in step S1 includes at least the longitude, latitude, altitude, speed parameters, heading parameters, and time parameters of the target UAV at multiple consecutive moments. Step S2, based on the target state information, predicts the movement trend of the target UAV in the near future, obtaining predicted state data corresponding to multiple discrete prediction moments within a preset prediction window. This predicted state data is then arranged in chronological order to form a predicted state sequence, which can be directly used for subsequent candidate camera retrieval, target control parameter determination, and visibility condition assessment.

[0073] In one embodiment, the preset prediction window represents a preset time span extending into the future from the current moment; the plurality of discrete prediction moments represent a plurality of prediction moments selected at preset time intervals within the preset time span. Assume the current moment is... The preset prediction window duration is The number of discrete prediction times is , These represent the discrete prediction times relative to the current time. The time offset, then the future discrete prediction time can be expressed as , … The discrete prediction times can be set at equal time intervals or at non-equal time intervals.

[0074] In this embodiment, short-time trajectory extrapolation refers to predicting the position and state of the target UAV within a short period of time in the future based on the target state information of the target UAV at the current moment and multiple consecutive moments before the current moment. Since the prediction window is a short-time window, the spatial position and motion state of the target UAV at each discrete prediction moment in the future can be extrapolated and estimated by utilizing the recent motion continuity of the target UAV.

[0075] In one specific implementation, the longitude, latitude, altitude, speed, heading, and time parameters of the target UAV at multiple recent consecutive moments can be selected as trajectory extrapolation inputs. Based on these inputs, the edge computing gateway or video-linked server determines the target UAV's direction and trend of motion at the current moment, and, combining the position changes, speed changes, and time intervals between adjacent moments, calculates the position state corresponding to each discrete prediction moment in the future. Each prediction moment thus corresponds to a set of prediction state data; multiple sets of prediction state data arranged in chronological order form the prediction state sequence.

[0076] In some embodiments, short-time trajectory extrapolation can be implemented using an extrapolation method based on current velocity and heading parameters. The current velocity parameter can be used as a speed reference for the next short time, and the current heading parameter as a direction reference for the next short time. Combined with the current position parameter and the time interval between adjacent state points, the predicted longitude, predicted latitude, and predicted altitude for each discrete prediction time in the future can be calculated. For the altitude direction, the predicted altitude for each discrete prediction time in the future can also be extrapolated by considering the altitude change trend between the current and previous times.

[0077] In other embodiments, the speed and heading trends can be corrected using target state information from multiple recent consecutive moments, and then a short-term trajectory extrapolation can be performed based on the corrected speed and heading trends. Any method that can generate a sequence of predicted states for multiple discrete prediction moments in the future based on target state information from multiple consecutive moments can be used to implement this invention.

[0078] In this embodiment, the predicted state data corresponding to each discrete prediction time includes at least the predicted position parameter for that discrete prediction time; in some embodiments, it may further include the predicted velocity parameter, predicted heading parameter, and predicted time parameter corresponding to that discrete prediction time. The resulting predicted state sequence can reflect the short-term movement trend of the target UAV within a preset prediction window in the future, rather than just the position prediction result at a single moment.

[0079] In some embodiments, before performing short-time trajectory extrapolation, basic preprocessing can be performed on the target state information obtained in step S1. This basic preprocessing may include correcting abnormal timestamp data, smoothing significantly abrupt position data, and interpolating missing state points. Through these processes, the impact of abnormal data on the short-time trajectory extrapolation results can be reduced, improving the continuity and availability of the predicted state data corresponding to each discrete prediction time in the future.

[0080] After completing step S2 and obtaining the prediction state sequence corresponding to multiple discrete prediction times within the future preset prediction window, execute step S3.

[0081] S3. Based on the predicted state sequence, retrieve candidate linked cameras, and for each candidate linked camera, determine the target control parameters and whether the visibility conditions are met at each discrete prediction time in the future.

[0082] The candidate linked cameras are retrieved in the following way: a spatial proximity search is performed on the camera resource library using a three-dimensional spatial index to obtain a preset number of linked cameras closest to the target UAV as initial candidate linked cameras; each initial candidate linked camera is sequentially checked for online status, signal quality, maximum effective line of sight, pitch angle range, horizontal rotation range, maximum available zoom, maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, and maximum zoom adjustment speed, and cameras that meet the linkage conditions are retained as candidate linked cameras; the three-dimensional spatial index is implemented based on the spatial database index.

[0083] In this embodiment, device information and control capability information of multiple associated cameras are pre-stored in the camera resource library, including at least the installation location, hanging height, maximum effective visible distance, pitch angle range, horizontal rotation range, maximum available zoom amount, maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, maximum zoom adjustment speed, online status, signal quality, and cruise status of the associated cameras. Step S3 first retrieves, according to the predicted state sequence output in step S2, the associated cameras adjacent to the spatial position of the target UAV in the camera resource library to form an initial candidate set of associated cameras for subsequent screening.

[0084] In one embodiment, one or more predicted position parameters in the position parameters of the target UAV at the current moment or the predicted state sequence can be used as the input for spatial proximity retrieval, and a nearest neighbor query is performed in the camera resource library through a three-dimensional spatial index to obtain multiple associated cameras closer to the target UAV. The preset quantity can be set according to the deployment density of associated cameras, computing resources, and real-time requirements in the supervised area.

[0085] After obtaining the initial candidate associated cameras, device availability checks and linkage capability checks are sequentially performed on each initial candidate associated camera. The device availability check at least includes a device online status check and a signal quality check; the linkage capability check at least includes a maximum effective visible distance check, a pitch angle range check, a horizontal rotation range check, a maximum available zoom amount check, and a maximum adjustment speed check in each direction. The associated cameras retained after the above item-by-item screening constitute the candidate set of associated cameras.

[0086] For each candidate associated camera, the target control parameters are determined at each future discrete prediction moment in the following manner: for each candidate associated camera, the geographical coordinate parameters corresponding to the predicted state sequence are converted into local coordinate parameters with the candidate associated camera as the reference coordinate system, and the target azimuth angle, target pitch angle, and target zoom amount required for imaging are respectively determined according to the local coordinate parameters; the target control parameters include the target azimuth angle, target pitch angle, and target zoom amount required for imaging.

[0087] In this embodiment, for each candidate associated camera, first, with the installation position of the candidate associated camera as the reference point, the geographical coordinate parameters corresponding to the target UAV at each future discrete prediction moment are converted into local coordinate parameters relative to the candidate associated camera.

[0088] In one embodiment, the local coordinate parameters can be represented using a local three-dimensional coordinate system with the installation position of the candidate linked camera as the origin. For example, an East-North-Sky local coordinate system or other equivalent local reference coordinate system can be used. By projecting the predicted position of the target UAV at each discrete prediction time into the local three-dimensional coordinate system, the horizontal position component, depth position component, and elevation component relative to the candidate linked camera can be obtained.

[0089] Based on the aforementioned local coordinate parameters, the target azimuth angle, target elevation angle, and required magnification for target imaging at each discrete prediction time in the future can be further determined. Specifically, the target azimuth angle characterizes the horizontal pointing angle of the target UAV relative to the candidate linked camera; the target elevation angle characterizes the vertical pointing angle of the target UAV relative to the candidate linked camera; and the required magnification for target imaging characterizes the magnification parameters required to achieve the preset imaging effect for the target UAV under the current target distance and imaging requirements.

[0090] In some embodiments, the required zoom level for target imaging can be determined based on the relative distance between the target UAV and the candidate linked camera, the expected display size of the target in the image, and the optical imaging capability of the linked camera. Any zoom level that can characterize the required zoom level of the candidate linked camera to achieve clear target imaging can be used as the required zoom level for target imaging in this invention.

[0091] The process of determining the target control parameters and whether the visibility conditions are met for each candidate linked camera at future discrete prediction times includes: for the first... The first candidate linked camera, in the future... At the discrete prediction time, the target UAV and the... Spatial distance between candidate linked cameras satisfy:

[0092] ;

[0093] In the formula, This serves as an index for candidate linked cameras. For the index of discrete prediction time, The target UAV calculated based on Vincenty's formula and the first The candidate linked camera in the first The ellipsoidal distance between discrete prediction times. For the target drone in the The height at each discrete prediction time. For the first The mounting height of each candidate linked camera.

[0094] When spatial distance is determined Not greater than the The maximum effective line-of-sight distance of the candidate linked cameras, the target pitch angle is located at the th Within the pitch angle range of the candidate linked cameras, the required zoom level for target imaging is no greater than that of the first camera. When determining the maximum available zoom level of the candidate linked cameras, the first... The candidate linked camera in the first Each discrete prediction time satisfies the visibility condition; When calculating the ellipsoidal distance, intermediate calculation results corresponding to the position parameters of the candidate linked cameras and the predicted position parameters of the target UAV that are repeatedly called are cached and reused.

[0095] In this embodiment, the spatial determination distance This is used to comprehensively characterize the actual separation degree between the target UAV and the candidate linked cameras in geospatial space. First, based on the predicted position parameters of the target UAV and the installation position parameters of the candidate linked cameras, the Vincenty formula is used to calculate the ellipsoidal distance between them on the Earth's ellipsoid; then, the distance is combined with the target UAV's position on the... The height of each discrete prediction time point Hanging high with candidate linkage cameras To obtain the spatial determination distance .

[0096] Subsequently, the spatial determination distance, target pitch angle, and target imaging magnification required are respectively compared with the first... The maximum effective viewing distance, pitch angle range, and maximum available zoom of each candidate linked camera are compared. When all of the above conditions are met, it indicates that the candidate linked camera is selected in the [number]th [period]. Each discrete prediction time can not only cover the target UAV in space, but also complete the imaging of the target UAV in terms of attitude range and optical capabilities. Therefore, it can be determined that it meets the visibility condition at the discrete prediction time.

[0097] In some embodiments, since the position parameters of the same candidate linked camera are repeatedly called at multiple discrete prediction times, and some predicted position parameters of the same target UAV are also repeatedly used in multiple visual condition determinations, the corresponding intermediate calculation results can be written into the cache unit to reduce the computational burden caused by repeated distance calculation and coordinate transformation, and improve the real-time performance of step S3.

[0098] Through the above processing, step S3 obtains the set of candidate linked cameras that meet the linkage conditions, the target control parameters of each candidate linked camera at each discrete prediction time in the future, and the visibility condition judgment results of each candidate linked camera at each discrete prediction time in the future. The above results serve as the input for step S4.

[0099] S4. Determine the continuous observation capability parameters of each candidate linked camera based on the ratio of the number of times that meet the visibility conditions to the total number of discrete prediction times in each future discrete prediction time. Determine the PTZ takeover cost parameters of each candidate linked camera based on the maximum values ​​of the horizontal rotation adjustment, pitch adjustment, and zoom adjustment amounts corresponding to the target control parameters when each candidate linked camera is adjusted from the current PTZ state to the target control parameters, respectively, after normalization by the corresponding maximum adjustment speed and the preset prediction window duration. Determine the comprehensive score of each candidate linked camera based on the continuous observation capability parameters and the PTZ takeover cost parameters, and determine the target linked camera based on the comprehensive score.

[0100] For each candidate linked camera, the continuous observation capability parameters of that candidate linked camera within a preset prediction window are first determined based on the visibility condition judgment results corresponding to each discrete prediction time in the future. Let there be a total of [number missing] cameras within the preset prediction window. At the discrete prediction time, the first... Among the candidate linked cameras are: If the visibility condition is satisfied at the nth discrete prediction time, then the nth... Continuous observation capability parameters of candidate linked cameras It can be calculated using the following formula:

[0101] ;

[0102] In the formula, Indicates the first The number of discrete prediction moments in which a candidate linked camera meets the visibility condition within a preset prediction window. This represents the total number of discrete prediction times within the preset prediction window. Therefore, The value range is from 0 to 1. Continuous observation capability parameter. The larger the value, the stronger the ability of the candidate linked camera to continuously observe the target drone within the preset prediction window in the future.

[0103] The PTZ takeover cost parameters for each candidate linked camera are determined as follows: The horizontal rotation adjustment amount corresponding to the target azimuth angle is divided by the product of the maximum horizontal rotation adjustment speed and the preset prediction window duration; the pitch adjustment amount corresponding to the target pitch angle is divided by the product of the maximum pitch adjustment speed and the preset prediction window duration; and the zoom adjustment amount corresponding to the zoom required for target imaging is divided by the product of the maximum zoom adjustment speed and the preset prediction window duration, to obtain the normalized values ​​for horizontal rotation, pitch, and zoom. The maximum value among these three values ​​is taken as the first normalized value. The PTZ takeover cost parameters for each candidate linked camera; the preset prediction window duration is the time span corresponding to the preset prediction window.

[0104] In one embodiment, for the first For each candidate camera to be linked, the target control parameters corresponding to the earliest discrete prediction time within a preset prediction window that satisfies the visibility condition can be selected as the takeover target parameters for that candidate camera. Let the first... The index of the earliest discrete prediction time when a candidate linked camera meets the visibility condition within a future preset prediction window is: ,but It can be defined as: No. The candidate linked camera in the first Each discrete prediction time satisfies the visibility condition. ;

[0105] If the first If a candidate linked camera does not have a discrete prediction time that meets the visibility conditions within a preset prediction window in the future, then the candidate linked camera is considered not to have effective takeover conditions, and the candidate linked camera will not participate in the PTZ takeover cost parameter calculation, or its PTZ takeover cost parameter will be set to a preset upper limit value.

[0106] When the The earliest discrete prediction time at which a candidate linked camera satisfies the visibility condition exists. When the discrete prediction time is reached, the target azimuth angle, target elevation angle, and the required zoom for target imaging are denoted as follows: , and .

[0107] Let the first The current horizontal rotation angle, current pitch angle, and current zoom of each candidate camera are as follows: , and Then its horizontal rotation adjustment amount Pitch adjustment amount and variable adjustment amount They are respectively:

[0108] ;

[0109] Let the first The maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, and maximum zoom adjustment speed of the candidate linked cameras are respectively... , and The preset prediction window duration is Then the normalized values ​​of horizontal rotation can be obtained respectively. Pitch normalized value and variable normalization value :

[0110] ;

[0111] In one embodiment, the normalized value is subjected to interval constraint processing to ensure it falls between 0 and 1. Then, the maximum value among the three is taken as the first value. PTZ takeover cost parameters for each candidate linked camera :

[0112] ;

[0113] thus, The smaller the value, the higher the value. The lower the takeover cost required to adjust a candidate linked camera from its current PTZ state to the target control parameters.

[0114] In obtaining continuous observation capability parameters PTZ takeover cost parameters Then, by combining signal quality parameters, mounting height adaptation parameters, zoom adaptation parameters, and cruise status parameters, a comprehensive score was given to each candidate linkage camera. Overall score of candidate linked cameras satisfy:

[0115] ;

[0116] In the formula, This serves as an index for candidate linked cameras. For the first Parameters of the continuous observation capability of each candidate linked camera. For the first PTZ takeover cost parameters for each candidate linked camera For the first Signal quality parameters of the candidate linked cameras For the first Mounting height adaptation parameters for each candidate linked camera For the first Zoom adaptation parameters for each candidate linked camera For the first Cruise status parameters of each candidate linked camera to The weights are for the corresponding parameters.

[0117] The , , and Each by the first The signal quality of each candidate camera, the degree of matching between its mounting height and the target height, the degree of matching between its available zoom capability and the zoom required for target imaging, and its cruise status are obtained after normalization, and satisfy the following conditions:

[0118] ;

[0119] In the formula, Statement No. One weight parameter, The index of the weight parameter is , and ;

[0120] Based on the comprehensive score of each candidate linkage camera Sort the cameras in descending order and identify the candidate camera at the top of the sorting results as the target camera.

[0121] In one embodiment, signal quality parameters According to the The real-time signal quality of each candidate linked camera was obtained after normalization processing; mounting height adaptation parameters. According to the The matching degree between the mounting height of each candidate linked camera and the predicted altitude of the target drone was obtained after normalization; zoom adaptation parameters According to the The degree of fit between the maximum available zoom level of each candidate linked camera and the zoom level required for target imaging is obtained after normalization; cruise state parameters According to the The preset values ​​for the current cruise status of each candidate linked camera are determined. Weight parameters. to It can be pre-configured by the system management terminal, or normalized after setting the importance of each evaluation factor according to the actual application scenario.

[0122] After obtaining the comprehensive scores for each candidate linked camera, the candidates are sorted in descending order according to their comprehensive scores, and the candidate linked camera with the highest score is determined as the target linked camera.

[0123] Through the above processing, step S4 outputs the continuous observation capability parameters, PTZ takeover cost parameters, comprehensive score, and the finally determined target camera for each candidate linked camera. These results serve as input for step S5.

[0124] S5. Generate PTZ control commands based on the current PTZ status of the target linkage camera and the target control parameters, and send them to the target linkage camera.

[0125] The generation of gimbal control commands based on the current PTZ state of the target-linked camera and the target control parameters is performed as follows: The gimbal control commands include horizontal rotation control commands, pitch rotation control commands, and zoom control commands; horizontal rotation error, pitch error, and zoom error are constructed based on the differences between the target azimuth angle and the current horizontal rotation angle, the differences between the target pitch angle and the current pitch angle, and the differences between the zoom amount required for target imaging and the current zoom amount; PID calculations are performed on the horizontal rotation error, pitch error, and zoom error to obtain the horizontal rotation control commands, pitch rotation control commands, and zoom control commands; after limiting and adapting the horizontal rotation control commands, pitch rotation commands, and zoom control commands to the control protocol, the device control interface is called for execution; before sending the gimbal control commands, the target-linked camera is controlled to pause cruise; after releasing the tracking control of the target UAV, the target-linked camera is controlled to resume cruise.

[0126] Once the target camera is identified, its current PTZ state is first obtained. The current PTZ state includes at least the current horizontal rotation angle, current pitch angle, and current zoom level. The current PTZ state can be obtained through methods such as proactive reporting by the target camera, polling by the video linkage server, or real-time reading from the device control interface.

[0127] The control error is constructed based on the difference between the target control parameters and the current PTZ state. Specifically, the difference between the target azimuth angle and the current horizontal rotation angle constitutes the horizontal rotation error, the difference between the target pitch angle and the current pitch angle constitutes the pitch error, and the difference between the required zoom for target imaging and the current zoom constitutes the zoom error.

[0128] In one embodiment, the horizontal rotation error can be separately... Pitch error and zoom error Represented as:

[0129] ;

[0130] In the formula, , , These represent the target azimuth angle, target elevation angle, and the required magnification for target imaging, respectively. , , These represent the current horizontal rotation angle, current pitch angle, and current zoom level, respectively.

[0131] PID calculations are performed on the horizontal rotation error, pitch error, and zoom error respectively to obtain the horizontal rotation control command, pitch rotation control command, and zoom control command. In one embodiment, for any control channel, the PID control output... It can be represented as:

[0132] ;

[0133] In the formula, For the control error of the corresponding channel, , , These are the proportional, integral, and derivative coefficients, respectively. The proportional term is used to improve the control response speed, the integral term is used to reduce steady-state error, and the derivative term is used to suppress overshoot and oscillation during the adjustment process.

[0134] After receiving horizontal rotation control commands, pitch rotation control commands, and zoom control commands, the control commands are subjected to amplitude limiting processing. This amplitude limiting processing refers to constraining the output amplitude, output rate, or control step size of the control commands based on the capabilities of the target-linked camera, in order to prevent the control commands from exceeding the equipment's allowable range.

[0135] In one embodiment, the control commands after amplitude limiting also need to be adapted to the control protocol. The video linkage server can convert the standardized control commands into a command format that the target linkage camera can recognize and execute, based on the device type and communication protocol corresponding to the target linkage camera. In some embodiments, the pan-tilt control commands output after the control protocol adaptation can adopt an absolute position control method or a relative incremental control method.

[0136] After completing the amplitude limiting process and control protocol adaptation, the device control interface is invoked to send the generated gimbal control commands to the target linkage camera. Upon receiving the corresponding horizontal rotation control commands, pitch rotation control commands, and zoom control commands, the target linkage camera performs horizontal rotation adjustment, pitch adjustment, and zoom adjustment to gradually bring the camera's observation attitude closer to the target control parameters, thereby achieving continuous linkage tracking of the target UAV.

[0137] In one embodiment, before sending the PTZ control command, the target-linked camera is first controlled to pause its patrol, to avoid the patrol and tracking tasks operating simultaneously on the PTZ control channel. Once the target-linked camera releases its tracking control of the target UAV, it is then controlled to resume patrol, allowing the device to return to normal patrol and monitoring status.

[0138] Through the above processing, step S5 outputs the PTZ control command sent to the target linkage camera, the updated PTZ status after the control is executed, and the current linkage tracking result. The above results are further used as the input basis for step S6.

[0139] S6. When the continuous observation capability parameter of the current target linkage camera is lower than the preset observation threshold, select the camera whose comprehensive score is not less than the preset score gain threshold from the candidate linkage cameras. Determine the camera with the highest comprehensive score in the selection results as the updated target linkage camera and send the corresponding PTZ control command to the updated target linkage camera.

[0140] The trigger condition for step S6 is that the continuous observation capability parameter of the current target linkage camera is lower than a preset observation threshold. The preset observation threshold is used to characterize the minimum capability requirement for the current target linkage camera to continue to maintain stable observation of the target UAV within a preset prediction window in the future. When the continuous observation capability parameter of the current target linkage camera is lower than the preset observation threshold, it indicates that the number of discrete prediction moments that meet the visibility conditions within the preset prediction window is small, and its ability to continue to undertake the tracking task decreases. At this time, the candidate linkage camera succession switching process is initiated.

[0141] After initiating the handover process, cameras whose overall score differs from the overall score of the current target camera from the candidate linked camera set are selected. These cameras have a score gain threshold not less than a preset threshold. In one embodiment, the overall score of the current target camera is set to... , No. The overall score of the candidate linked cameras is The preset score gain threshold is Candidate cameras that meet the following conditions will be added to the switching candidate set:

[0142] ;

[0143] The preset scoring gain threshold is used to avoid frequent switching of linked cameras due to only minor fluctuations in the overall score. An actual switch is only triggered when the candidate linked camera has a sufficient scoring advantage over the current target linked camera.

[0144] After the above screening of candidate linked cameras, the camera with the highest comprehensive score in the screening results is determined as the updated target linked camera. In one embodiment, if multiple candidate linked cameras meet the preset score gain threshold condition, the one with the highest comprehensive score is selected as the updated target linked camera; if the comprehensive scores are the same, they can be further sorted and selected according to the priority order of higher continuous observation capability parameter, lower PTZ takeover cost parameter, or higher signal quality parameter.

[0145] After identifying the updated target linked camera, this updated target linked camera is designated as the new target linked camera, and a corresponding PTZ control command is sent to it. The PTZ control command is generated in the same way as the PTZ control command generated for the target linked camera in step S5, that is, based on the current PTZ state and target control parameters of the updated target linked camera, control error is constructed, PID calculation is performed, amplitude limiting processing and control protocol adaptation are performed, and finally, the device control interface is called to execute.

[0146] In one embodiment, before the updated target-linked camera begins the tracking task, the current target-linked camera can be controlled to exit the tracking control state, and then the updated target-linked camera can be controlled to enter the tracking control state, so as to achieve a smooth handover of the linked tracking task between the two linked cameras. In other embodiments, a brief overlap control phase can also be set, so that the current target-linked camera and the updated target-linked camera can jointly track the target drone for a short period of time, thereby reducing the risk of image loss during the switching moment.

[0147] By using step S6, when the current target linkage camera's continuous observation capability declines in the future, the linkage tracking task can be promptly transferred to a better candidate linkage camera, thereby improving the continuity and stability of target UAV video linkage tracking.

[0148] like Figure 2 As shown, another embodiment of the present invention also provides a drone follow-up tracking system based on remote identification information, comprising:

[0149] Remote ID receiving device, used to receive Remote ID messages broadcast by the target drone;

[0150] An edge computing gateway, connected to the Remote ID receiving device, is used to perform protocol decoding, timing caching, and association of the Remote ID messages to form target status information of the target drone at multiple consecutive moments.

[0151] The camera resource library is used to store the installation location, mounting height, maximum effective viewing distance, pitch angle range, horizontal rotation range, maximum available zoom, maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, maximum zoom adjustment speed, online status, signal quality, and cruise status of linked cameras.

[0152] Multiple linked cameras are used to receive PTZ control commands and perform corresponding horizontal rotation, pitch and zoom adjustments, and report the current PTZ status to the video linkage server;

[0153] The video linkage server, connected to the edge computing gateway, camera resource library, and multiple linkage cameras, includes:

[0154] The prediction module is used to generate a sequence of predicted states for multiple discrete prediction times within a preset prediction window based on the target state information.

[0155] The candidate filtering module is used to retrieve candidate linked cameras based on the predicted state sequence;

[0156] The parameter calculation module is used to determine the target control parameters for each candidate linked camera and to determine whether the visibility conditions are met.

[0157] The scoring and decision module is used to determine the continuous observation capability parameters, PTZ takeover cost parameters, and comprehensive scores of each candidate linked camera, and to determine the target linked camera based on the comprehensive score.

[0158] The control transmission module is used to generate PTZ control commands based on the current PTZ status of the target linkage camera and the target control parameters, and then send them to the target linkage camera.

[0159] The switching module is used to update the target linked camera when the continuous observation capability parameter of the current target linked camera is lower than the preset observation threshold, and to send the corresponding pan-tilt control command to the updated target linked camera.

[0160] In this embodiment, the Remote ID receiving device, edge computing gateway, camera resource library, multiple linked cameras, and video linkage server work together. The Remote ID receiving device and edge computing gateway acquire the target drone's status information. The video linkage server performs target prediction, candidate linked camera screening, target linked camera determination, gimbal control, and seamless switching based on the target status information. The linked cameras then execute corresponding tracking control, thereby achieving continuous video linkage tracking of the target drone.

[0161] In summary, this invention constructs a complete processing chain for target UAVs, encompassing state perception, trajectory prediction, selection of linked cameras, gimbal control, and seamless switching. This results in a UAV continuous tracking solution with clear sequential connections, well-defined module coordination, and continuous operation. This solution is suitable for low-altitude surveillance scenarios involving fixed-point cameras, as well as applications requiring continuous observation and rapid response to UAV targets, such as patrol security, emergency response, and protection of key areas. It demonstrates excellent engineering adaptability and significant potential for widespread application.

[0162] Without contradiction, the various embodiments and technical features described in this specification can be combined with each other.

[0163] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. For those skilled in the art, various modifications, alterations, or equivalent substitutions can be made without departing from the technical concept of this application, and such modifications, alterations, or equivalent substitutions should all fall within the scope of protection defined by the claims of this application.

Claims

1. A method for continuous tracking of unmanned aerial vehicles (UAVs) based on remote identification information, characterized in that, Includes the following steps: S1. Receive and parse the Remote ID message broadcast by the target UAV to obtain the target status information of the target UAV at multiple consecutive moments; S2. Based on the target state information, generate a sequence of predicted states for multiple discrete prediction times within a preset prediction window in the future. S3. Based on the predicted state sequence, retrieve candidate linked cameras, and for each candidate linked camera, determine the target control parameters and whether the visibility conditions are met at each discrete prediction time in the future. For each candidate linked camera, the target control parameters for each discrete prediction time in the future are determined as follows: For each candidate linked camera, the geographic coordinate parameters corresponding to the prediction state sequence are converted into local coordinate parameters with the candidate linked camera as the reference coordinate system, and the target azimuth angle, target elevation angle, and target imaging zoom are determined according to the local coordinate parameters for each discrete prediction time in the future; the target control parameters include the target azimuth angle, target elevation angle, and target imaging zoom. The process of determining the target control parameters and whether the visibility conditions are met for each candidate linked camera at future discrete prediction times includes: For the The first candidate linked camera, in the future... At the discrete prediction time, the target UAV and the... Spatial distance between candidate linked cameras satisfy: ； In the formula, This serves as an index for candidate linked cameras. For the index of discrete prediction time, The target UAV calculated based on Vincenty's formula and the first The candidate linked camera in the first The ellipsoidal distance between discrete prediction times. For the target drone in the The height at each discrete prediction time. For the first The mounting height of each candidate linked camera; when determining spatial distance. Not greater than the The maximum effective line-of-sight distance of the candidate linked cameras, the target pitch angle is located at the th Within the pitch angle range of the candidate linked cameras, the required zoom level for target imaging is no greater than that of the first camera. When determining the maximum available zoom level of the candidate linked cameras, the first... The candidate linked camera in the first Each discrete prediction time satisfies the visibility condition; when calculating the ellipsoidal distance, the intermediate calculation results corresponding to the repeatedly called candidate linkage camera position parameters and the target UAV predicted position parameters are cached and reused; S4. Based on the ratio of the number of times that meet the visibility conditions to the total number of discrete prediction times in each future discrete prediction time, determine the continuous observation capability parameters of each candidate linked camera; based on the maximum values ​​of the horizontal rotation adjustment, pitch adjustment, and zoom adjustment amounts corresponding to the target control parameters adjusted from the current PTZ state to the maximum values ​​of the corresponding maximum adjustment speed and the preset prediction window duration, determine the PTZ takeover cost parameters of each candidate linked camera; the determination of the PTZ takeover cost parameters of each candidate linked camera is as follows: divide the horizontal rotation adjustment amount corresponding to the target azimuth angle by the product of the maximum horizontal rotation adjustment speed and the preset prediction window duration, divide the pitch adjustment amount corresponding to the target pitch angle by the product of the maximum pitch adjustment speed and the preset prediction window duration, and divide the zoom adjustment amount corresponding to the zoom amount required for target imaging by the product of the maximum zoom adjustment speed and the preset prediction window duration to obtain the normalized values ​​of horizontal rotation, pitch, and zoom; take the maximum value among the normalized values ​​of horizontal rotation, pitch, and zoom as the first value. PTZ takeover cost parameters for candidate linked cameras The preset prediction window duration is the time span corresponding to the preset prediction window; the comprehensive score of each candidate linkage camera is determined based on the continuous observation capability parameter and the PTZ takeover cost parameter, and the target linkage camera is determined based on the comprehensive score; S5. Generate PTZ control commands based on the current PTZ status of the target-linked camera and the target control parameters, and send them to the target-linked camera; S6. When the continuous observation capability parameter of the current target linkage camera is lower than the preset observation threshold, select the camera whose comprehensive score is not less than the preset score gain threshold from the candidate linkage cameras. Determine the camera with the highest comprehensive score in the selection results as the updated target linkage camera and send the corresponding PTZ control command to the updated target linkage camera.

2. The method for continuous tracking of unmanned aerial vehicles based on remote identification information according to claim 1, characterized in that, The process of receiving and parsing the Remote ID message broadcast by the target UAV is achieved through the following steps: receiving the Remote ID message actively broadcast by the target UAV through a Remote ID receiving device, and performing protocol decoding on the Remote ID message at the edge computing gateway; based on the unique UAV identifier in the Remote ID message, performing time-series caching and association on multiple consecutive broadcast period messages belonging to the same target UAV to form an associated message sequence; extracting longitude, latitude, altitude, speed parameters, heading parameters, and time parameters from the associated message sequence to form continuous target state information that can be directly called for subsequent prediction state sequence construction.

3. The method for continuous tracking of unmanned aerial vehicles based on remote identification information according to claim 2, characterized in that, The predicted state sequence is obtained by extrapolating the short-time trajectory of the target UAV based on its state information at multiple consecutive time points. The sequence of predicted states corresponding to each discrete prediction time. This is the number of discrete prediction times within the preset prediction window.

4. The method for continuous tracking of unmanned aerial vehicles based on remote identification information according to claim 3, characterized in that, The candidate linked cameras are retrieved in the following way: a spatial proximity search is performed on the camera resource library using a three-dimensional spatial index to obtain a preset number of linked cameras closest to the target UAV as initial candidate linked cameras; each initial candidate linked camera is sequentially checked for online status, signal quality, maximum effective line of sight, pitch angle range, horizontal rotation range, maximum available zoom, maximum horizontal rotation adjustment speed, maximum pitch adjustment speed, and maximum zoom adjustment speed, and cameras that meet the linkage conditions are retained as candidate linked cameras; the three-dimensional spatial index is implemented based on the spatial database index.

5. The method for continuous tracking of unmanned aerial vehicles based on remote identification information according to claim 1, characterized in that, The comprehensive score for each candidate linked camera is determined based on continuous observation capability parameters and PTZ takeover cost parameters, including: the first... Overall score of candidate linked cameras satisfy: ； In the formula, This serves as an index for candidate linked cameras. For the first Parameters of the continuous observation capability of each candidate linked camera. For the first PTZ takeover cost parameters for each candidate linked camera For the first Signal quality parameters of the candidate linked cameras For the first Mounting height adaptation parameters for each candidate linked camera For the first Zoom adaptation parameters for each candidate linked camera For the first Cruise status parameters of each candidate linked camera to As the weight of the corresponding parameter, the , , and Each by the first The signal quality of each candidate camera, the degree of matching between its mounting height and the target height, the degree of matching between its available zoom capability and the zoom required for target imaging, and its cruise status are obtained after normalization, and satisfy the following conditions: ； In the formula, Statement No. One weight parameter, The index of the weight parameter, and ; Based on the comprehensive score of each candidate linkage camera Sort the cameras in descending order and identify the candidate camera at the top of the sorting results as the target camera.

6. The method for continuous tracking of unmanned aerial vehicles based on remote identification information according to claim 1, characterized in that, The generation of PTZ control commands based on the current PTZ state of the target-linked camera and the target control parameters is performed as follows: The PTZ control commands include horizontal rotation control commands, pitch rotation control commands, and zoom control commands. Horizontal rotation error, pitch error, and zoom error are constructed based on the differences between the target azimuth angle and the current horizontal rotation angle, the differences between the target pitch angle and the current pitch angle, and the differences between the required zoom amount for target imaging and the current zoom amount. PID calculations are performed on the horizontal rotation error, pitch error, and zoom error to obtain the horizontal rotation control commands, pitch rotation control commands, and zoom control commands. After limiting and adapting the horizontal rotation control commands, pitch rotation commands, and zoom control commands to the control protocol, the device control interface is invoked for execution. Before sending the PTZ control commands, the target-linked camera is controlled to pause cruise. After releasing the tracking control of the target-linked camera, the target-linked camera is controlled to resume cruise.

7. A drone follow-up tracking system based on remote identification information, used to execute a drone follow-up tracking method based on remote identification information as described in any one of claims 1-6, characterized in that, include: Remote ID receiving device, used to receive Remote ID messages broadcast by the target drone; An edge computing gateway, connected to the Remote ID receiving device, is used to perform protocol decoding, timing caching, and association of the Remote ID messages to form target status information of the target drone at multiple consecutive moments. The camera resource library stores the installation location, mounting height, maximum effective viewing distance, tilt angle range, horizontal rotation range, maximum available zoom, maximum horizontal rotation adjustment speed, maximum tilt adjustment speed, maximum zoom adjustment speed, online status, signal quality, and cruise status of the linked cameras; multiple linked cameras receive PTZ control commands and execute corresponding horizontal rotation, tilt, and zoom adjustments, and report the current PTZ status to the video linkage server; The video linkage server, connected to the edge computing gateway, camera resource library, and multiple linkage cameras, includes: a prediction module for generating a prediction state sequence of multiple discrete prediction times within a preset prediction window based on the target state information; a candidate filtering module for retrieving candidate linkage cameras based on the prediction state sequence; a parameter calculation module for determining target control parameters for each candidate linkage camera and determining whether the visibility conditions are met; a scoring decision module for determining the continuous observation capability parameters, PTZ takeover cost parameters, and comprehensive score of each candidate linkage camera, and determining the target linkage camera based on the comprehensive score; a control sending module for generating PTZ control commands based on the current PTZ state and target control parameters of the target linkage camera and sending them to the target linkage camera; and a switching module for updating the target linkage camera when the continuous observation capability parameters of the current target linkage camera are lower than a preset observation threshold, and sending the corresponding PTZ control commands to the updated target linkage camera.

Images