A three-dimensional pose measurement method of a high-speed projectile in a large scene

Abstract

The application provides a three-dimensional position measurement method of a high-speed projectile in a large scene. Firstly, a plurality of general high-definition cameras with adjacent far optical axes and large optical axis included angles are used to synchronously acquire high-speed projectile movement images in multiple angles in combination with a Beidou high-precision timing technology, and then high-precision reconstruction of the three-dimensional coordinates of the high-speed projectile in the large scene and the multiple fields of view is completed based on a multi-optical axis fusion processing method, so that the problems in the prior art, such as the influence of ground impact vibration on the observation of the high-speed projectile by a traditional binocular vision measurement system and the failure of a single camera to run due to damage, are solved, and high-precision and high-reliability measurement of the three-dimensional position of the high-speed projectile is realized.

Description

Technical Field

[0001] This invention belongs to the field of graphics processing technology, specifically relating to a method for measuring the three-dimensional pose of a high-speed projectile in a large scene. Background Technology

[0002] The accurate detection of missile terminal trajectory parameters is challenging due to factors such as the vast geographical area of ​​the impact zone, harsh natural environment, complex scene interference, and the missile's high speed and small size. Therefore, developing a three-dimensional pose measurement method suitable for high-speed missiles in large-scale scenarios, which can effectively improve the accuracy and reliability of measurement results while meeting experimental requirements, is of great significance.

[0003] The literature “Huang Huimin. Research on 3D Reconstruction and Measurement Technology Based on Images [D]. University of Chinese Academy of Sciences (Xi’an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences), 2021.DOI:10.27605 / d.cnki.gkxgs.2020.000040.” proposes a 3D pose measurement method based on binocular stereo vision, which mainly uses the time difference between left and right images to estimate depth information and is divided into three steps: camera calibration, stereo matching, and 3D information recovery. The literature “Huang Shasha, Liu Qi, Qiao Jia. Measurement Method of Motion Attitude of High-Speed ​​Projectiles [J]. Journal of Projectiles, Rockets and Guidance, 2018, 38(04):73-76.DOI:10.15892 / j.cnki.djzdxb.2018.04.018.” uses a calibrated high-speed camera measurement system to take pictures of the projectile, and then performs intersection measurement on multiple marker points sprayed on the projectile to finally calculate the motion attitude parameters of the projectile. The literature “Zhuang Sufeng, Tu Dawei, Zhang Xu, et al. Research on underwater binocular stereo vision corresponding point matching and 3D reconstruction method [J]. Journal of Instrumentation, 2022, 43(05):147-154.DOI:10.19650 / j.cnki.cjsi,J2209215.” proposes a method for underwater corresponding point matching and 3D measurement, realizing the 3D reconstruction of underwater targets. The literature “Yan Xiaoyu, Lu Fanfan, Ge Lusheng, et al. Measurement research based on binocular stereo vision system [J]. Journal of Chongqing Technology and Business University (Natural Science Edition), 2023, 40(03):48-55.DOI:10.16055 / j.issn.1672-058X.2023.0003.007.” designs a measurement system method based on binocular stereo vision, and uses the BM stereo matching feature method to realize the measurement of target objects. The literature “Li Xiangping, Lu Junyong, Zhang Xiao, et al. Binocular visual measurement of sabot separation of discarding sabot [J]. Journal of National University of Defense Technology, 2023, 45(01): 150-156” built an object attitude measurement platform based on the principle of binocular vision, realizing the tracking and measurement of the marked points on the surface of the projectile and sabot.

[0004] Traditional 3D pose measurement methods, in order to ensure image feature matching, involve overlapping information in the images, and the measurement accuracy depends on the image matching accuracy. This leads to the camera being positioned with a small optical axis angle, close to the optical axis. At the same time, high-speed projectiles can cause camera vibration and damage, affecting measurement accuracy and reliability. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a method for measuring the three-dimensional pose of high-speed projectiles in large-scale scenes. First, multiple general-purpose high-definition cameras with large optical axis angles and far adjacent optical axes are used, combined with BeiDou high-precision timing technology, to synchronously acquire multi-angle images of the high-speed projectile's motion. Then, based on a multi-optical axis fusion processing method, high-precision reconstruction of the three-dimensional coordinates of the high-speed projectile under multi-field-of-view conditions in large scenes is completed. This solves the problems of traditional binocular vision measurement systems being susceptible to ground impact vibrations and the inability to operate when a single camera is damaged, thus achieving high-precision and high-reliability measurement of the three-dimensional pose of high-speed projectiles.

[0006] A method for measuring the three-dimensional pose of a high-speed projectile in a large-scale scene, characterized by the following steps: Step 1: Deploy several general-purpose high-definition cameras with far adjacent optical axes and large optical axis angles to photograph the target. Before shooting, calibrate the internal and external parameters of the high-definition cameras. Far adjacent optical axes means that the distance between different high-definition cameras is more than 140m, and large optical axis angle means that the angle between the orientations of different high-definition cameras is more than 85°. Step 2: Utilize BeiDou high-precision time synchronization technology to unify the time code system of all high-definition cameras; Step 3: Use the synchronized high-definition cameras to capture images of the target, and acquire target image sequences. Use the LK optical flow method based on physical and spatiotemporal constraints to perform optical flow tracking on the image sequences acquired by each high-definition camera to obtain several target motion trajectories. The LK optical flow method based on physical and spatiotemporal constraints refers to adding constraints to the LK optical flow method according to the physical and spatiotemporal characteristics of the target. The physical constraint includes: a) the mapping vector on the y-axis of the vector from the target's motion termination center point to a certain starting optical flow point. Rather than x Mapping vectors on the axis The angle between b) The area of ​​the target in the image satisfies a certain threshold H; ,in, S 1 represents the total area of ​​the set of adjacent feature points in the image. S 0 represents the total area threshold of the feature point set. For a long strip-shaped target, its aspect ratio in the image satisfies the following condition: The initial range of the threshold H is 45°-135°. S The initial value is 10, and the threshold is...K The initial value is 4; Spatiotemporal constraints include: a) The target motion termination point is below the target motion starting point, satisfying... , This represents the ordinate value of the target's starting point in the image coordinate system. a) The vertical coordinate of the target's motion termination point in the image coordinate system; b) After the target terminates its motion, it will be accompanied by smoke and flames. The starting and ending points of the flames are detected. Filter out any subsequent interfering targets, i.e. ,in, T Indicates the current time of the image; Step 4: The improved Kalman filter method is used to estimate and predict the motion trajectory of each target to obtain the corrected target motion trajectory; the improved Kalman filter method refers to the Kalman filter method that defines its observation noise factor using the target pixel coordinates and the target motion time as parameters. Step 5: For each high-definition camera, construct its optical center and the plane containing the corrected target motion trajectory as its orbital plane. Calculate the intersection lines of each pair of orbital planes, and then use the least squares method to fit all the obtained plane intersection lines to obtain the initial target 3D motion trajectory. Use the minimum reprojection error method to perform nonlinear optimization on the initial target 3D trajectory to obtain the final target 3D motion trajectory. Then, based on the target's motion trajectory and geometric calculations, obtain the target's 3D attitude parameters.

[0007] Specifically, the calibration of the high-definition camera's intrinsic and extrinsic parameters in step 1 involves: using the Zhang Zhengyou calibration method to calibrate the high-definition camera's intrinsic parameters, and then using the PnP problem-solving method to calibrate the high-definition camera's extrinsic parameters.

[0008] Specifically, the method for deploying several general-purpose high-definition cameras with large optical axis angles in step 1 is as follows: four high-definition cameras are evenly deployed on a circumference 100-150m away from the target.

[0009] Specifically, step 2 involves receiving BeiDou satellite time and frequency information using a BeiDou high-precision timing module, deploying an NTP clock server near the high-definition camera storage terminal to use the received BeiDou satellite time and frequency information as a reference clock, synchronizing the time codes of all high-definition cameras, and configuring a laptop computer as a clock network manager to manage and monitor the clock server.

[0010] Specifically, the formula for defining the observation noise factor using the target pixel coordinates and target motion time as parameters in step 4 is as follows: (1) in, RRepresents the observation noise covariance matrix. R 0 represents the initial noise covariance matrix. The coordinates of the image center are... Let be the coordinates of the target at the initial moment of motion. for k The target coordinates at that moment k Indicates the current time of the image being processed. k 0 indicates the initial time of the image being processed.

[0011] Specifically, when performing nonlinear optimization of the initial target 3D trajectory using the minimum reprojection error method described in step 5, the objective function is set as follows: (2) in, f Represents the objective function value. Represents the calculated three-dimensional trajectory of the first... i The first high-definition camera track j The projection coordinates of each point in the image coordinate system Represents the third on the three-dimensional trajectory i The first high-definition camera track j The true projected coordinates of each point in the image coordinate system. n This indicates the number of points on the three-dimensional trajectory. K The number of high-definition cameras deployed. w i Indicates the first i The weighting factors are calculated using the following formula: (3) in, O ci P Indicates the first i The distance between the image point of a high-definition camera and the target point. O cj P Indicates the first j The distance between the image point of a high-definition camera and the target point.

[0012] Specifically, in step 5, obtaining the target's three-dimensional attitude parameters based on the target's motion trajectory and geometric calculations refers to the target's position, angle, and velocity. The position information is directly read from the three-dimensional motion trajectory; the angle is the angle between the three-dimensional motion trajectory and the y-axis and z-axis in space; and the velocity is the partial derivative of the three-dimensional motion distance with respect to time. The three-dimensional motion distance... s for: (4) in, , , These represent the position functions of the x-axis, y-axis, and z-axis relative to time, respectively. t The derivative; The speed of movement is .

[0013] The beneficial effects of this invention are: 1) By employing an improved Kalman filtering method related to the motion time of the high-speed projectile, the problem of ground camera vibration caused by the high-speed projectile's aerial motion and ground impact vibration caused by the projectile's landing is effectively solved, achieving high-precision measurement of the high-speed projectile's three-dimensional pose. 2) By using optical center-image plane mapping trajectory surfaces and obtaining the motion trajectory through trajectory surface intersection, the problem of inter-frame delay caused by time system deviations is effectively solved, achieving high-reliability measurement of the high-speed projectile's three-dimensional pose. This invention effectively improves the accuracy and reliability of measurement results while meeting experimental requirements. Attached Figure Description

[0014] Figure 1 This is a flowchart of a three-dimensional pose measurement method for a high-speed projectile in a large scene according to the present invention; Figure 2 This is a schematic diagram of the high-definition camera deployment according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the local area network time synchronization server setup according to an embodiment of the present invention; Figure 4 This is a 3D schematic diagram of multi-camera trajectories. Detailed Implementation

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments. The present invention includes, but is not limited to, the following embodiments.

[0016] This invention provides a method for measuring the three-dimensional pose of a high-speed projectile in a large-scale scene, mainly comprising two parts: First, using multiple general-purpose high-definition cameras with large optical axis distances and large optical axis angles, combined with BeiDou high-precision timing technology, multi-angle synchronous acquisition of high-speed projectile motion images is performed; then, based on a multi-optical axis fusion processing method, high-precision reconstruction of the three-dimensional coordinates of the high-speed projectile under multiple fields of view in a large-scale scene is completed. Figure 1 As shown, the specific implementation process is as follows: 1. High-definition camera setup and internal / external parameter calibration Several general-purpose high-definition cameras with large optical axis angles and adjacent distances were deployed to photograph the target, and the intrinsic and extrinsic parameters of the high-definition cameras were calibrated. "Adjacent distances" refers to a distance of more than 140m between different high-definition cameras, and "large optical axis angles" refers to an angle of more than 85° between the orientations of different high-definition cameras. Figure 2As shown, four high-definition cameras are arranged in a wide-angle, long-distance configuration, with two serving as primary cameras and two as backup cameras. The angle is approximately 90°, and the distance varies from approximately 100 to 150 meters depending on the target object. (See figure.) Oc For the coordinates of the high-definition camera's origin / optical center, Op Let the origin of the image coordinate system be... u、v The two axes of the image coordinate system p (u,v) Spatial coordinate points P Projection on the image coordinate system P (X,Y,Z) These are the actual spatial coordinates of the point.

[0017] The calibration of high-definition camera intrinsic and extrinsic parameters includes: calibration of the high-definition camera's intrinsic parameters using Zhang Zhengyou's calibration method, and then calibration of the high-definition camera's extrinsic parameters using the PnP problem-solving method. The process of extrinsic parameter calibration using the PnP problem-solving method is as follows: first, obtain a sufficient number (at least four pairs) of coordinate pairs corresponding to the pixel coordinates and the actual coordinates; then, according to... Given internal parameters K Solving for extrinsic parameters in the formula R and t ,in, Z express P w The distance to the optical center of a high-definition camera P uv Represents pixel coordinates. P w Represents world coordinate system coordinates. T Indicates participation from high-definition cameras R and t related, R Represents the extrinsic rotation matrix. t This represents the extrinsic parameter translation matrix.

[0018] 2. Timecode synchronization of high-definition cameras Utilizing BeiDou high-precision time synchronization technology, the time code system of all high-definition cameras is unified. This includes: First, installing a BeiDou satellite clock, which uses a BeiDou high-precision time synchronization module to receive BeiDou satellite time and frequency information. The BeiDou time synchronization antenna is placed in an unobstructed location to ensure accurate time and frequency information acquisition. Second, setting up a local area network (LAN) time synchronization server, which deploys an NTP clock server near the high-definition camera storage terminal to use the received BeiDou satellite time and frequency information as a reference clock to synchronize the time codes of all high-definition cameras. A laptop computer is configured as a clock network manager to manage and monitor the clock server. Specific details are as follows... Figure 3 As shown.

[0019] 3. Optical flow tracing based on the physical and spatiotemporal constraints of high-speed projectiles The target was captured by synchronized high-definition cameras, and each camera acquired a sequence of images. The LK optical flow method, based on physical and spatiotemporal constraints, was used to perform optical flow tracking on the image sequence acquired by each high-definition camera, resulting in several target motion trajectories. First, inter-frame differencing is performed on the acquired image sequence to obtain the initial image feature points. Based on the constant brightness principle and spatial consistency of the LK optical flow method, the following equation is established for 9 pixels centered on the first pixel: (5) in, and Represents the velocity component of a point. and Indicates brightness relative to the image x , y Components of the axis, This represents the brightness component relative to time, and the numbers in the subscripts indicate the pixel index. The least squares method is used to solve the problem. and This allows us to obtain the short-term motion state of the feature points.

[0020] Then, constraints are added to the optical flow map based on the physical and spatiotemporal characteristics of the target object itself, in order to effectively reduce the false detection rate of the target object's feature points in the optical flow map. Among these, the physical characteristic constraints include: a) the mapping vector on the y-axis of the vector from the target's motion termination center point to a certain starting optical flow point. Rather than x Mapping vectors on the axis The angle between a) Greater than a certain threshold H (depending on the physical characteristics of different targets); b) Based on the fact that projectiles generally maintain a certain physical shape and volume during motion, their area in the image is a certain value, and the area of ​​the target in the image satisfies... This can eliminate most of the interference from clouds and minor image changes, among which, S 1 represents the total area of ​​the set of adjacent feature points in the image. S 0 represents the total area threshold of the feature point set. For a long strip-shaped target, its aspect ratio in the image satisfies the following condition: The initial range of the threshold H is 45°-135°. S The initial value is 10, and the threshold is... K The initial value is 4.

[0021] Spatiotemporal constraints include: a) The target motion termination point is below the target motion starting point, satisfying... , This represents the ordinate value of the target's starting point in the image coordinate system. a) The vertical coordinate of the target's motion termination point in the image coordinate system; b) After the target terminates its motion, it will be accompanied by smoke and flames. The starting and ending points of the flames are detected. Filter out any subsequent interfering targets, i.e. ,in, T Indicates the current time of the image.

[0022] The projectile trajectory based on the optical flow diagram can be obtained using the method described above.

[0023] 4. Projectile trajectory prediction based on improved Kalman filter The trajectory obtained in step 3 is affected by image distortion and the impact vibration of the high-definition camera, and has a certain error. The Kalman filter method is used to predict the trajectory of the projectile. The estimated value of the previous moment and the observed value of the current moment are used to update the estimate of the state variables and obtain the estimated value of the current moment.

[0024] Considering the characteristics of the object of this invention, the motion process is treated as an acceleration process during the calculation. State equations for the discrete process are established. x k and observation equations z k As shown in the following formula: (6) In the formula, These represent the image coordinates of the target point, the velocity components of the target point's motion, and the acceleration components, respectively. k , k- 1 represents k Time and k -1 time; A It is the state transition coefficient matrix. H It is the observation coefficient matrix. It is state noise. It is observation noise.

[0025] Based on the state equation and the observation equation, the predicted state equation and the updated state equation during the entire motion process are as follows: (7) (8) in, , This represents the estimated values ​​of the state variables and the covariance matrix of the state variables under prior conditions. , This represents the estimated values ​​of the state variables and the covariance matrix of the state variables. Q , R Represent the process noise covariance matrix and the observation noise covariance matrix. It is Kalman gain.I It is an identity matrix.

[0026] The parameters in the formula are: (9) To address the impact of image distortion and camera shake in high-definition cameras, this invention improves the Kalman filtering method by defining the observation noise factor using the target pixel coordinates and target motion time as parameters, i.e., calculating the observation noise covariance matrix using the following formula. R : (10) in, R 0 represents the initial noise covariance matrix. The coordinates of the image center are... Let be the coordinates of the target at the initial moment of motion. for k The target coordinates at that moment k Indicates the current time of the image being processed. k 0 indicates the initial time of the image being processed.

[0027] 5. Three-dimensional pose measurement based on multi-plane intersection Target 3D Motion Trajectory Determination Based on Multi-Plane Intersection: Considering the characteristics of the target environment—a large scene area with few landmarks—an algorithm for surface intersection based on the optical center and the projectile trajectory in multiple high-definition cameras is established to address the issue of inconsistent frame correspondence times. The trajectory solution is simplified to multi-plane intersection calculation; theoretically, several planes will intersect at a straight line, but in practice… Figure 4 As shown in the figure, O c For the coordinates of the high-definition camera, O p Let the origin of the image coordinate system be... u、v The two axes of the image coordinate system l i(u,v) For the calculated first i The target image trajectory of a high-definition camera L i(X,Y,Z) For the calculated i-th three-dimensional trajectory, L (X,Y,Z) This represents the actual three-dimensional trajectory.

[0028] First, the optical center of the target and the plane containing its corrected trajectory are constructed as its orbital plane. The intersections of each pair of orbital planes are calculated. Then, the least squares method is used to fit all the obtained plane intersections to obtain the initial 3D target trajectory. Finally, the minimum reprojection error method is used to perform nonlinear optimization on the initial 3D target trajectory to obtain the final 3D target trajectory. The specific solution process is as follows: 1) Find the normal vectors of the two planes. The two planes are... The normal vectors of the two planes can be obtained as follows: , .in, These are the coefficients of the plane equation, which can be determined from the optical center. O The image plane trajectory is obtained.

[0029] 2) Solve for the direction vector of the intersection line. .

[0030] 3) Find a point on the line ,make x =0, therefore Solving for the given information, we can obtain the following results: .according to and The intersection line is obtained.

[0031] 4) The initial three-dimensional trajectory of the projectile contains image resolution errors. Based on the obtained initial three-dimensional trajectory, the reciprocal of the distance from the high-definition camera to the object point is converted into a weighting factor. w i And set the weighting factor w i The relevant objective function employs a minimum reprojection error method to nonlinearly optimize the initial 3D trajectory of the target, yielding the final 3D motion trajectory of the target. Among these, the weighting factors... w i Calculate using the following formula: (11) in, O ci P Indicates the first i The distance between the optical center of a high-definition camera and the target point O cj P Indicates the first j The distance between the optical center of a high-definition camera and the target point K This refers to the number of high-definition cameras deployed.

[0032] The objective function is as follows: (12) in, f Represents the objective function value. Represents the calculated three-dimensional trajectory of the first... i The first high-definition camera track j The projection coordinates of each point in the image coordinate system Represents the third on the three-dimensional trajectory i The first high-definition camera track j The true projected coordinates of each point in the image coordinate system.n This represents the number of points on the three-dimensional trajectory.

[0033] Three-dimensional pose calculation of a high-speed projectile: Three-dimensional pose refers to three-dimensional position information and three-dimensional attitude information. The target's three-dimensional motion trajectory has already been calculated, and the target's position information at any given moment can be read. Three-dimensional attitude parameters refer to the target's position, angle, and velocity. The position information is directly read from the three-dimensional motion trajectory. The angle is the angle between the three-dimensional motion trajectory and the y-axis and z-axis in space. The velocity is the partial derivative of the three-dimensional motion distance with respect to time. The three-dimensional motion distance... s for: (13) in , , These represent the position functions of the x-axis, y-axis, and z-axis relative to time, respectively. t The derivative of the velocity relative to the speed. t The function is .

Claims

1. A method for measuring the three-dimensional pose of a high-speed projectile in a large-scale scene, characterized in that... The steps are as follows: Step 1: Deploy several general-purpose high-definition cameras with far adjacent optical axes and large optical axis angles to photograph the target. Before shooting, calibrate the internal and external parameters of the high-definition cameras. Far adjacent optical axes means that the distance between different high-definition cameras is more than 140m, and large optical axis angle means that the angle between the orientations of different high-definition cameras is more than 85°. Step 2: Utilize BeiDou high-precision time synchronization technology to unify the time code system of all high-definition cameras; Step 3: Use the synchronized high-definition cameras to capture images of the target, and acquire target image sequences. Use the LK optical flow method based on physical and spatiotemporal constraints to perform optical flow tracking on the image sequences acquired by each high-definition camera to obtain several target motion trajectories. The LK optical flow method based on physical and spatiotemporal constraints refers to adding constraints to the LK optical flow method according to the physical and spatiotemporal characteristics of the target. The physical constraint includes: a) the mapping vector on the y-axis of the vector from the target's motion termination center point to a certain starting optical flow point. Rather than x Mapping vectors on the axis The angle between b) The area of ​​the target in the image satisfies a certain threshold H; ,in, S 1 represents the total area of ​​the set of adjacent feature points in the image. S 0 represents the total area threshold of the feature point set. For a long strip-shaped target, its aspect ratio in the image satisfies the following condition: The initial range of the threshold H is 45°-135°. S The initial value is 10, and the threshold is... K The initial value is 4; Spatiotemporal constraints include: a) The target motion termination point is below the target motion starting point, satisfying... , This represents the ordinate value of the target's starting point in the image coordinate system. a) The vertical coordinate of the target's motion termination point in the image coordinate system; b) After the target terminates its motion, it will be accompanied by smoke and flames. The starting and ending points of the flames are detected. Filter out any subsequent interfering targets, i.e. ,in, T Indicates the current time of the image; Step 4: The improved Kalman filter method is used to estimate and predict the motion trajectory of each target to obtain the corrected target motion trajectory; the improved Kalman filter method refers to the Kalman filter method that defines its observation noise factor using the target pixel coordinates and the target motion time as parameters. Step 5: For each high-definition camera, construct its optical center and the plane containing the corrected target motion trajectory as its orbital plane. Calculate the intersection lines of each pair of orbital planes, and then use the least squares method to fit all the obtained plane intersection lines to obtain the initial target 3D motion trajectory. Use the minimum reprojection error method to perform nonlinear optimization on the initial target 3D trajectory to obtain the final target 3D motion trajectory. Then, based on the target's motion trajectory and geometric calculations, obtain the target's 3D attitude parameters.

2. The method for measuring the three-dimensional pose of a high-speed projectile in a large-scale scene as described in claim 1, characterized in that: The calibration of the high-definition camera's intrinsic and extrinsic parameters in step 1 specifically involves: using the Zhang Zhengyou calibration method to calibrate the high-definition camera's intrinsic parameters, and then using the PnP problem-solving method to calibrate the high-definition camera's extrinsic parameters.

3. The method for measuring the three-dimensional pose of a high-speed projectile in a large scene as described in claim 1, characterized in that: The specific method for deploying several general-purpose high-definition cameras with large optical axis angles in step 1 is as follows: Four high-definition cameras are evenly deployed on a circumference 100-150m away from the target.

4. The method for measuring the three-dimensional pose of a high-speed projectile in a large scene as described in claim 1, characterized in that: The specific process of step 2 is as follows: Use the Beidou high-precision timing module to receive Beidou satellite time and frequency information, deploy an NTP clock server near the high-definition camera storage terminal to use the received Beidou satellite time and frequency information as the reference clock to complete the time code synchronization of all high-definition cameras, and configure a laptop computer as a clock network manager to manage and monitor the clock server.

5. The method for measuring the three-dimensional pose of a high-speed projectile in a large scene as described in claim 1, characterized in that: The specific formula for defining the observation noise factor using the target pixel coordinates and target motion time as parameters in step 4 is as follows: （1） in, R Represents the observation noise covariance matrix. R 0 represents the initial noise covariance matrix. The coordinates of the image center are... Let be the coordinates of the target at the initial moment of motion. for k The target coordinates at that moment k Indicates the current time of the image being processed. k 0 indicates the initial time of the image being processed.

6. The method for measuring the three-dimensional pose of a high-speed projectile in a large scene as described in claim 1, characterized in that: When performing nonlinear optimization of the initial target 3D trajectory using the minimum reprojection error method described in step 5, the objective function is set as follows: （2） in, f Represents the objective function value. Represents the calculated three-dimensional trajectory of the first... i The first high-definition camera track j The projection coordinates of each point in the image coordinate system Represents the third on the three-dimensional trajectory i The first high-definition camera track j The true projected coordinates of each point in the image coordinate system. n This represents the number of points on the three-dimensional trajectory. K The number of high-definition cameras deployed. w i Indicates the first i The weighting factors are calculated using the following formula: （3） in, O ci P Indicates the first i The distance between the image point of a high-definition camera and the target point. O cj P Indicates the first j The distance between the image point of a high-definition camera and the target point.

7. The method for measuring the three-dimensional pose of a high-speed projectile in a large scene as described in claim 1, characterized in that: The three-dimensional attitude parameters of the target obtained in step 5, based on the target's motion trajectory and geometric calculations, refer to the target's position, angle, and velocity. The position information is directly read from the three-dimensional motion trajectory; the angle is the angle between the three-dimensional motion trajectory and the y-axis and z-axis in space; and the velocity is the partial derivative of the three-dimensional motion distance with respect to time. The three-dimensional motion distance... s for: （4） in, , , These represent the position functions of the x-axis, y-axis, and z-axis relative to time, respectively. t The derivative; The speed of movement is .

Images