A method for measuring flight trajectory of a target pellet in real time based on a multi-node orthogonal light curtain

Abstract

This invention relates to the field of photoelectric measurement technology and provides a real-time measurement method for the flight trajectory of a target based on a multi-node orthogonal light curtain. The method includes: deploying at least two measurement nodes along the target's flight path; each node forming a two-dimensional detection light curtain through an orthogonal linear array photoelectric sensor and a collimated light source; using a high-speed signal processing unit to process the shading pulse signals from all sensor channels in parallel, and extracting the timestamps and two-dimensional position coordinates of the target as it passes through each light curtain in real time; and reconstructing the target's three-dimensional spatial position, velocity vector, and flight attitude at each moment in real time using spatial intersection and time difference algorithms based on the node spatial coordinates, timestamps, and two-dimensional coordinates. This invention enables high-precision and low-latency synchronous measurement of all parameters, including position, velocity, and attitude, of a high-speed flying target, providing crucial real-time trajectory data support for applications such as high-repetition-rate laser target firing.

Description

Technical Field

[0001] This invention relates to the field of photoelectric measurement technology, and in particular to a method for real-time measurement of target trajectory based on multi-node orthogonal light curtains. Background Technology

[0002] In applications involving high-speed micro-targets, such as laser fusion, real-time, high-precision measurement of the high-speed flight trajectory of miniature targets in a vacuum environment is a crucial prerequisite for achieving precise laser aiming, energy coupling analysis, and physical process diagnosis. Existing technologies employ high-speed camera-based visual measurement systems that capture sequential images of the target's flight and then use complex image processing algorithms to calculate the trajectory. Other technologies utilize single-point photoelectric sensors to trigger timing based on changes in light intensity caused by the target passing through a single detection light curtain, thereby obtaining its passage time and limited positional information.

[0003] However, while high-speed camera systems can acquire two-dimensional image information, their data processing is complex and computationally intensive, resulting in poor real-time performance and making it difficult to meet the requirements for real-time online feedback control of each target shot in high-repetition-rate target practice. Furthermore, the strong background light interference generated during target practice significantly reduces the image signal-to-noise ratio, leading to insufficient measurement stability. In addition, while single-point photoelectric sensor systems have fast response times, they can only provide the time and one-dimensional position information of the target shot as it passes through a single plane, and cannot independently reconstruct the complete trajectory, velocity vector, and flight attitude of the target shot in three-dimensional space. Therefore, they cannot simultaneously meet the high requirements for high-precision, high-real-time, and full-parameter measurement of high-speed flying targets.

[0004] In view of this, a method for real-time measurement of target flight trajectory based on multi-node orthogonal light curtain is proposed. Summary of the Invention

[0005] This invention provides a method for real-time measurement of target trajectory based on multi-node orthogonal light curtains, which solves the problem of not being able to simultaneously meet the high requirements of high precision, high real-time performance and full parameter measurement of high-speed flying targets.

[0006] This invention provides a method for real-time measurement of target trajectory based on multi-node orthogonal light curtains, including: At least two measurement nodes are set up along the flight path of the target pellet. Each measurement node includes two sets of orthogonally arranged linear array photoelectric sensors and corresponding collimated light sources to form a two-dimensional detection light curtain covering the cross-section of the flight path. As the target pellet passes through the two-dimensional detection light curtain at each measurement node in sequence, pulse signals generated by the two sets of linear array photoelectric sensors at each measurement node due to the target pellet blocking the light are collected. Based on the collected pulse signals, all channel signals are processed in parallel using a high-speed signal processing unit to extract the timestamp of the target passing through each of the two-dimensional detection light curtains in real time, and to extract the two-dimensional position coordinates of the target on each of the two-dimensional detection light curtain planes. Based on the known fixed coordinates of each measurement node in three-dimensional space and the corresponding timestamps and two-dimensional position coordinates extracted from each node, the three-dimensional spatial position, velocity vector and flight attitude of the target at each moment are reconstructed in real time through spatial line intersection and time difference algorithms.

[0007] Furthermore, after deploying at least two measurement nodes along the flight path of the target pellet and before performing target pellet measurements, the method further includes: The calibrator is used to pass through the two-dimensional probe light curtain of each measurement node sequentially along a preset calibration trajectory; The calibration signal generated when the calibration object passes through each of the two-dimensional detection light curtains is recorded, and the spatial plane equation and relative position relationship of each of the two-dimensional detection light curtains in a unified coordinate system are calculated based on the geometric relationship between the calibration signal recorded at each node and the preset calibration trajectory.

[0008] Furthermore, the formation of the two-dimensional detection light curtain covering the cross-section of the flight path includes: At the measurement node, a light-shielding plate is moved along the preset detection area of ​​the two-dimensional detection light curtain, and the output signal of the linear array photoelectric sensor is monitored in real time. Adjust at least one of the following based on the changing characteristics of the output signal: the luminous intensity of the collimated light source, the beam angle, or the operating parameters of the linear array photoelectric sensor, until the two-dimensional detection light curtain has uniform and consistent shading response characteristics within the preset detection area.

[0009] Furthermore, the acquisition of pulse signals generated by the two sets of linear array photoelectric sensors in each measurement node due to target blockage includes: During the acquisition process, the signal-to-noise ratio and pulse waveform of the output signal of each linear array photoelectric sensor are monitored; The signal amplifier gain of the corresponding sensor channel is dynamically adjusted based on the monitored signal-to-noise ratio, and the sampling clock frequency and trigger threshold of the high-speed signal processing unit are dynamically adjusted based on the monitored pulse waveform width.

[0010] Furthermore, the real-time extraction of the timestamp of the target pellet passing through each of the two-dimensional detection light curtains includes: It receives pulse signals generated by two sets of orthogonal linear array photoelectric sensors at each measurement node; The moment when the rising edge of each group of pulse signals reaches a preset threshold is taken as the first candidate moment, and the width of each group of pulse signals is calculated. Determine whether the time difference between two first candidate times corresponding to the same node is less than a preset threshold. If the time difference is less than a preset threshold, then the timestamp corresponding to the current node is calculated based on the two first candidate times. If the time difference is greater than or equal to a preset threshold, a timestamp corresponding to the current node is generated based on the width of the pulse signal and the measurement data of adjacent nodes.

[0011] Furthermore, the extraction of the two-dimensional position coordinates of the target pellet on each of the two-dimensional detection light curtain planes includes: The sensitive pixel range blocked by the target pellet is determined based on the pulse signals from the two sets of sensor channels. Based on the query of the pre-stored lookup table in the sensitive pixel interval, the initial coordinates with sub-pixel precision are obtained; Apply pre-stored correction parameters to the initial coordinates to obtain one-dimensional coordinates; By synthesizing two orthogonal one-dimensional coordinates, the two-dimensional position coordinates of the target on the two-dimensional detection light curtain plane are obtained.

[0012] Furthermore, the spatial line intersection and time difference algorithm adopts a recursive solution process based on time series constraints: Based on the two-dimensional position coordinates extracted from a measurement node and the spatial equation of the light curtain plane, a spatial straight line characterizing the possible position of the target pellet is determined. The spatial straight lines of subsequent measurement nodes are introduced in sequence, and spatiotemporal consistency is verified with the currently constructed target motion state estimation. When new node data is introduced, the optimal estimate of the target's three-dimensional position and velocity is updated by utilizing the velocity continuity constraint of the target's motion. The recursive solution process is executed after each acquisition of data from a new measurement node to achieve real-time incremental reconstruction of the trajectory.

[0013] Furthermore, the reconstruction of the three-dimensional spatial position of the target at each moment specifically involves: In the recursive solution process, the three-dimensional spatial position of the target is expressed as a function that changes with time; An optimization problem is constructed and solved with the objective of minimizing the weighted sum of squared vertical distances from the spatial lines of each node to the function at the corresponding timestamp; the weights in the weighted sum of squares are dynamically allocated based on the confidence level of the timestamps extracted from each node and the residuals of the two-dimensional position coordinates.

[0014] Furthermore, the reconstructing of the velocity vector of the target pellet specifically involves: After obtaining the sequence of the three-dimensional spatial position of the target pellet changing with time, the position sequence is differentiated over time to calculate the instantaneous velocity sequence. An adaptive filtering algorithm based on a kinematic model is used to smooth the instantaneous velocity sequence in order to suppress measurement noise and meet physical constraints, and output the smoothed velocity vector.

[0015] Furthermore, the reconstructing of the target's flight attitude specifically involves: Based on the surface geometric features of the target pellet, a geometric relationship model is established between the surface marker points and their projections on at least two non-parallel light curtain planes; Obtain the two-dimensional position coordinates of different parts of the target surface passing through the light curtain, measured at similar times by different measurement nodes; The two-dimensional position coordinates are substituted into the geometric relationship model for inverse solution to calculate the roll, pitch and yaw angles of the target.

[0016] As can be seen from the above technical solutions, the present invention has the following advantages: This invention employs a method that deploys at least two measurement nodes along the target's flight path. Each node forms a two-dimensional detection light curtain using orthogonal linear array photoelectric sensors and collimated light sources. A high-speed signal processing unit processes the shading pulse signals from all sensor channels in parallel, extracting the timestamps and two-dimensional position coordinates of the target as it passes through each light curtain in real time. Based on the node spatial coordinates, timestamps, and two-dimensional coordinates, a spatial linear intersection and time difference algorithm is used to reconstruct the target's three-dimensional spatial position, velocity vector, and flight attitude at each moment in real time. This invention, through its node-based light curtain arrangement and parallel signal processing architecture, solves the technical problems of poor real-time performance, weak anti-interference capabilities, and the inability of single-point sensors to acquire three-dimensional information in high-speed camera systems. It enables high-precision and low-latency synchronous measurement of all parameters, including position, velocity, and attitude, of high-speed flying targets, providing crucial real-time trajectory data support for applications such as high-repetition-rate laser target shooting. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for real-time measurement of target trajectory based on a multi-node orthogonal light curtain, as described in this invention. Figure 2 This is a geometric diagram illustrating the spatial line intersection and time difference algorithm in this invention. Figure 3 This is a flowchart illustrating the recursive solution process and reconstruction steps in this invention. Detailed Implementation

[0018] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] Example 1 To better understand the method shown in this embodiment, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. (Reference) Figure 1 The technical solution of the present invention provides a method for real-time measurement of target trajectory based on multi-node orthogonal light curtain, comprising: S1. At least two measurement nodes are set up along the flight path of the target pellet, wherein each measurement node includes two sets of orthogonally arranged linear array photoelectric sensors and corresponding collimated light sources to form a two-dimensional detection light curtain covering the cross section of the flight path. A single node can only provide information that the target is located on a straight line in space, but cannot determine its specific depth position on that line. By introducing at least two nodes with known spatial positions, and using the two straight lines obtained when the target passes through different nodes, the precise position of the target in three-dimensional space can be uniquely determined through spatial geometric intersection methods. This is the geometric basis for realizing three-dimensional trajectory reconstruction.

[0020] The number of measurement nodes here is preferably three or more, arranged at intervals along the flight path in a non-collinear manner. The core component of each measurement node includes a... Directional linear array sensor, one The system consists of a linear array sensor and a shared collimated light source that illuminates both. The linear array sensor employs a high-speed response photodiode array with multiple photosensitive pixels arranged in a straight line. The collimated light source emits a parallel beam with a very small divergence angle, uniformly illuminating the entire effective detection area of ​​the sensor array. direction and Two sets of sensors are arranged orthogonally at 90 degrees in space, and the overlapping areas of their photosensitive surfaces together form a two-dimensional detection light curtain. When the target pellet passes through this light curtain, it will... Sensors and The sensors partially block light, generating corresponding electrical pulse signals. By measuring the specific pixel positions where the pulse signals appear, the two-dimensional coordinates of the target pellet on the light curtain plane can be determined. These coordinates include... coordinates and coordinate.

[0021] In this embodiment, after setting up at least two measurement nodes along the flight path of the target pellet and before performing target pellet measurement, the following steps are also included: 1. Use the calibrator to pass sequentially through the two-dimensional detection light curtain of each measurement node along the preset calibration trajectory; 2. Record the calibration signal generated when the calibration object passes through each two-dimensional detection light curtain, and calculate the spatial plane equation and relative position relationship of each two-dimensional detection light curtain in a unified coordinate system based on the geometric relationship between the calibration signal recorded at each node and the preset calibration trajectory.

[0022] The calibration object is a high-precision sphere with known geometry and motion characteristics, driven by a high-precision mechanical device. The preset calibration trajectory is a smooth curve, the equation of which is known in the world coordinate system. The calibration object is driven strictly along this known trajectory, sequentially passing through the light curtains of all measurement nodes. When the calibration object passes through the light curtain of a certain node, the light curtain of that node... Sensors and The sensor generates a light-blocking pulse similar to that used during target measurement; this signal is defined as the calibration signal. The two-dimensional coordinates and the time of passage corresponding to the calibration signal at each node are recorded. Since the true spatial position of the calibrated object can be calculated from the equation of the preset calibration trajectory at the moment it passes through each light curtain, multiple sets of correspondences between two-dimensional measurement points on the light curtain plane and known three-dimensional points in space are obtained. Based on these correspondences, the least squares fitting algorithm is used to solve the spatial plane equation of the two-dimensional detection light curtain at each node in its own sensor coordinate system. Furthermore, through coordinate transformation, the plane equations of all nodes are unified into a common world coordinate system, obtaining their relative positional relationships, i.e., translation and rotation parameters. This calibration process is a prerequisite for fusing the independent measurement data of each node into a unified three-dimensional space for calculation.

[0023] In this embodiment, the method for forming a two-dimensional detection light curtain covering the cross-section of the flight path includes the following: 1. At the measurement node, move the light-shielding plate along the preset detection area of ​​the two-dimensional detection light curtain and monitor the output signal of the linear array photoelectric sensor in real time; 2. Adjust at least one of the following parameters based on the changing characteristics of the output signal: the luminous intensity of the collimating light source, the beam angle, or the operating parameters of the linear array photoelectric sensor, until the two-dimensional detection light curtain has uniform and consistent shading response characteristics within the preset detection area.

[0024] The coverage of the flight path cross-section here means that the area of ​​the two-dimensional detection light curtain must be large enough to completely cover all the spatial areas that the target pellet may fly over, thus ensuring that the target pellet is not missed. The specific debugging process is as follows: A light-shielding plate smaller than the target pellet is used, and a two-dimensional electronically controlled translation stage is used to make it scan at a constant speed within the preset detection area of ​​the light curtain. This preset detection area is the effective area that the target pellet is expected to fly over. During this process, data is collected and recorded in real time. and The output signals, i.e., voltage values, of all channels of the two sensors. Under an ideal uniform light curtain, when the light-shielding plate scans, the signal amplitude drop of the blocked channels should be basically consistent, and the background noise levels between different channels should also be similar. If there are changes such as significant attenuation of signal amplitude at the edge of the area, inconsistent drop depth at different positions, or excessive noise fluctuations, it indicates that the light curtain is not uniform or the sensor response is inconsistent. At this time, dynamic adjustment is required. Specifically, the following parameters can be adjusted individually or in combination: adjust the luminous intensity of the collimating light source to make the overall illuminance moderate by changing the driving current; fine-tune the beam angle of the collimating light source or its relative position with the sensor to optimize the optical path so that the beam covers the sensor array as vertically and uniformly as possible; adjust the operating parameters of the linear array photoelectric sensor, such as its bias voltage, amplifier gain, or comparator trigger threshold. Through iterative adjustment and scanning tests, the final result is that when the light-shielding plate passes through any position in the effective area, the amplitude fluctuation of the pulse signal generated by the sensor is less than 5% of the preset threshold, and the signal-to-noise ratio is higher than the minimum requirement. At this time, the light curtain is considered to have uniform and consistent light-shielding response characteristics in this area.

[0025] S2. When the target pellet passes through the two-dimensional detection light curtain of each measurement node in sequence, the pulse signals generated by the two sets of linear array photoelectric sensors in each measurement node due to the target pellet blocking the light are collected. When a high-speed projectile passes through a two-dimensional light curtain formed by a measurement node, its opaque surface blocks a portion of the light originally illuminating the target. direction and The parallel beam of light on the linear array sensor causes a sudden drop in light intensity received by the blocked pixel. The photosensitive unit inside the sensor converts this change in light intensity into an electrical signal, thereby generating a distinctive pulse signal. The generation of this pulse signal is related to the moment when the target passes through the light curtain plane, and the specific sensor channel number where the pulse appears uniquely corresponds to the blocking position of the target on the light curtain plane.

[0026] In this embodiment, the pulse signals generated by the two sets of linear array photoelectric sensors in each measurement node due to the target ball blocking the light are collected, including the following: 1. Monitor the signal-to-noise ratio and pulse waveform of the output signal of each linear array photoelectric sensor during the acquisition process; 2. Dynamically adjust the gain of the signal amplifier for the corresponding sensor channel based on the monitored signal-to-noise ratio, and dynamically adjust the sampling clock frequency and trigger threshold of the high-speed signal processing unit based on the monitored pulse waveform width.

[0027] To address the impact of target velocity variations and ambient light disturbances on signal quality, this embodiment employs an adaptive signal acquisition strategy. The system monitors the output signal of each sensor channel using metrics including signal-to-noise ratio (SNR), evaluated by calculating the ratio of the root mean square of the baseline noise to the effective pulse amplitude. Another metric is the pulse waveform; the system captures and analyzes the pulse rise time, fall time, and pulse width in real time. If the SNR of a channel is detected to be below a preset threshold, indicating that the signal may be overwhelmed by noise, the system automatically increases the gain of the signal amplifier for that channel to enhance the signal amplitude and ensure reliable pulse identification. Conversely, if the signal amplitude is close to saturation, the gain is appropriately reduced to prevent distortion. When a narrowing pulse width is detected, indicating a potential increase in target velocity, the system dynamically increases the sampling clock frequency of the high-speed signal processing unit to ensure sufficient data points are acquired within a shorter obstruction period to accurately characterize the pulse shape and thus accurately determine its leading edge. In addition, the system dynamically fine-tunes the comparator's trigger threshold based on the real-time background noise level and pulse amplitude, keeping it at the optimal position that is slightly above the noise level and can respond quickly to valid pulses.

[0028] S3. Based on the acquired pulse signals, the high-speed signal processing unit processes all channel signals in parallel to extract the timestamp of the target passing through each two-dimensional detection light curtain in real time, and extract the two-dimensional position coordinates of the target on each two-dimensional detection light curtain plane. The high-speed signal processing unit here is implemented based on a field-programmable gate array (FPGA) architecture. This architecture provides an independent signal processing logic link for each sensor channel, enabling signals from all measurement nodes to be processed simultaneously and independently. Extracting the timestamp is to obtain the precise moment the target crosses each light curtain plane, which is the basis for subsequent velocity vector calculations and time-difference analysis. Extracting the two-dimensional position coordinates is to determine the specific position of the target on each light curtain plane, providing necessary directional information for spatial line intersection.

[0029] In this embodiment, the timestamps of the target pellet passing through each two-dimensional detection light curtain are extracted in real time, including the following: 1. Receive pulse signals generated by two sets of orthogonal linear array photoelectric sensors at each measurement node; 2. The moment when the rising edge of each group of pulse signals reaches a preset threshold is taken as the first candidate moment, and the width of each group of pulse signals is calculated. 3. Determine whether the time difference between two first candidate times corresponding to the same node is less than a preset threshold; 4. If the time difference is less than the preset threshold, calculate the timestamp corresponding to the current node based on the two first candidate times; 5. If the time difference is greater than or equal to the preset threshold, then generate the timestamp corresponding to the current node based on the width of the pulse signal and the measurement data of the adjacent nodes.

[0030] Specifically, the high-speed signal processing unit receives digitized pulse signals from all sensor channels. For each channel's signal, the processing unit detects in real time when its voltage value exceeds a specific voltage level, which is the preset threshold. This preset threshold is set based on the signal's average noise level plus several times the noise standard deviation to ensure reliable triggering and suppress noise-induced false triggering. The instant the signal voltage first exceeds this threshold is recorded as the rising edge moment, and this moment is used as the first candidate moment for that channel's signal. Simultaneously, the processing unit calculates the pulse width, which is the duration for which the signal voltage is above the preset threshold. For the same measurement node, signals from... Sensors and The system identifies two candidate time points for the sensors. Since both sensors measure the same target passing through the same physical plane, these two time points should theoretically be very close. Therefore, the system determines whether the absolute value of the time difference between these two time points is less than a preset threshold. This threshold can be estimated based on the light curtain thickness and the target's maximum velocity, for example, set to a few microseconds. If the time difference is less than this threshold, the two signals are considered valid and synchronized. In this case, the arithmetic mean of the two candidate time points is taken as the final timestamp for that node. If the time difference is greater than or equal to the threshold, it indicates that at least one signal may be interfered with or malfunctioning. In this case, the candidate time points cannot be used directly. The system then analyzes the pulse width and, combined with the reliable timestamps and velocity estimates already determined by upstream neighboring nodes, generates the timestamp corresponding to the current node through extrapolation interpolation. This timestamp is then marked as an estimated value in the data to ensure the continuity of the data stream and provide a reference for subsequent processing.

[0031] In this embodiment, the extraction of the two-dimensional position coordinates of the target pellet on each two-dimensional detection light curtain plane includes the following steps: 1. Determine the sensitive pixel range blocked by the target pellet based on the pulse signals from the two sets of sensor channels; 2. Based on the sensitive pixel interval query, the pre-stored lookup table is used to obtain the initial coordinates with sub-pixel precision; 3. Apply pre-stored correction parameters to the initial coordinates to obtain one-dimensional coordinates; 4. By synthesizing two orthogonal one-dimensional coordinates, the two-dimensional position coordinates of the target on the two-dimensional detection light curtain plane are obtained.

[0032] Specifically, when the target pellet generates a pulse signal by blocking light, one or more adjacent pixels simultaneously exceed a threshold. The system identifies the pixel numbers of all pixels whose signals exceed the threshold, with the pixel having the largest signal amplitude as the center, and several adjacent pixels forming the sensitive pixel interval. Since the light spot has a certain size and may cover multiple pixels, calculating coordinates solely based on the center pixel number will result in quantization errors. Therefore, during the system calibration phase, a mapping relationship between the signal amplitude distribution pattern of each pixel within the sensitive pixel interval and its actual physical location is established through scanning, and this relationship is stored as a lookup table. During processing, based on the real-time acquired signal amplitude distribution of each pixel within the interval, this lookup table is consulted to interpolate and obtain initial coordinates with sub-pixel accuracy higher than that of a single pixel size. Sensor installation, optical lenses, etc., may introduce slight nonlinear distortions. During the calibration phase, the deviation between the actual coordinates and ideal coordinates at different positions on the light curtain plane is measured and fitted into a set of correction parameters. In real-time processing, substituting the initial coordinates obtained in the previous step into the correction function eliminates system distortions and yields accurate one-dimensional coordinates. To each direction and The orientation sensor repeats the above steps to obtain two precise one-dimensional coordinates, denoted as follows: coordinates and Coordinates. Combining these two coordinates yields the final two-dimensional position coordinates of the target pellet on the two-dimensional detection light curtain plane of the measurement node.

[0033] S4. Based on the known fixed coordinates of each measurement node in three-dimensional space and the corresponding timestamps and two-dimensional position coordinates extracted from each node, the three-dimensional spatial position, velocity vector and flight attitude of the target at each moment are reconstructed in real time through spatial line intersection and time difference algorithms.

[0034] refer to Figure 2 In the diagram, P1, P2, and P3 are the fixed spatial coordinates of three measurement nodes, and L1, L2, and L3 are the spatial straight lines determined by the measurement data of each node. "" represents the position of the measurement node, and "-" represents the trajectory of the target. Given the spatial position of each measurement node in a unified world coordinate system and the precise orientation of its light curtain plane, the measured two-dimensional position coordinates are as follows when the target passes through the light curtain of a certain node. This means that the target is located on a specific spatial straight line that passes through the plane of the node's light curtain. Points, perpendicular to the light curtain plane. For example, for node P1, its measurement determines a spatial straight line L1. Similarly, the measurements for nodes P2 and P3 determine straight lines L2 and L3, respectively. Ideally, at the corresponding moment when the target passes through each light curtain, its spatial position should simultaneously lie on these three straight lines, that is, the three straight lines should intersect at a single point. Due to measurement errors, the three straight lines usually do not intersect perfectly. Therefore, the task of the spatial straight line intersection algorithm is to find the most probable sequence of three-dimensional points, considering the errors, such that each point is close to the measurement line at the corresponding moment. The temporal difference algorithm then uses precise timestamp information to perform temporal difference operations on the above three-dimensional point sequence, thereby calculating the velocity vector of the target's flight. By closely combining spatial geometric intersection with time series analysis, it is possible to reconstruct a continuous, smooth three-dimensional motion trajectory and all its derived parameters with high precision from discrete, noisy two-dimensional measurement data. Reference Figure 3 By using spatial linear intersection and time difference algorithms, the three-dimensional spatial position, velocity vector, and flight attitude of the target at each moment are reconstructed in real time, including the following: In this embodiment, the spatial line intersection and time difference algorithm adopts a recursive solution process based on time series constraints: 1. Based on the two-dimensional position coordinates extracted from a measurement node and the spatial equation of the light curtain plane, determine a spatial straight line characterizing the possible position of the target pellet; 2. Introduce the spatial straight lines of subsequent measurement nodes in sequence, and verify their spatiotemporal consistency with the currently constructed target motion state estimation; 3. When new node data is introduced, the optimal estimate of the target's three-dimensional position and velocity is updated by utilizing the velocity continuity constraint of the target's motion; 4. The recursive solution process is executed after each acquisition of data from a new measurement node to achieve real-time incremental reconstruction of the trajectory.

[0035] Specifically, the spatial equation of the light curtain plane was obtained during the system calibration phase, and its form is as follows: This defines the orientation and position of the plane in the world coordinate system. Combined with real-time extracted two-dimensional coordinates... Based on the plane normal vector, the equation of a straight line in space can be determined, and this line passes through the point... (In the local coordinate system of the sensor) and launched along the plane normal direction, after coordinate transformation to the world coordinate system, this represents the straight line where the target ball might be located at that moment. After obtaining data from the first node, the system initializes a coarse estimate of the target's motion state. When data from the second and subsequent nodes arrive, the system performs a spatiotemporal consistency check, which examines the newly determined spatial straight line. The system assesses whether the target's position at the new timestamp is spatially close to the predicted position based on the current state, and whether the time interval is reasonable. Here, the velocity continuity constraint means that the target's velocity cannot undergo instantaneous changes; its variation is limited by physical laws. When introducing new data, the system combines the new straight-line measurement information, timestamp information, and the previous state estimate (including position and velocity). Under the premise of satisfying motion continuity, the algorithm iteratively calculates the new optimal estimate that best matches all existing information—that is, the most likely three-dimensional position and velocity at the current moment. Each time the measurement data of a node is processed, the above recursive calculation is triggered, updating the trajectory estimate. This makes trajectory output almost synchronous with data acquisition, achieving true real-time reconstruction without waiting for all node data to be collected before batch processing.

[0036] In this embodiment, the three-dimensional spatial position of the target pellet at each moment is reconstructed as follows: 1. In the recursive solution process, the three-dimensional spatial position of the target is expressed as a function that changes with time; 2. An optimization problem is constructed and solved with the objective of minimizing the weighted sum of squares of the vertical distances from the spatial lines to the function at the corresponding timestamps of each node. The weights in the weighted sum of squares are dynamically allocated based on the confidence level of the timestamps extracted from each node and the residuals of the two-dimensional position coordinates.

[0037] Specifically, the function is to measure its three-dimensional position coordinates within the time window. Express it using a polynomial. Assume we have... Valid measurement data for each node, for each data point Includes: timestamp and the spatial straight line determined by it. Define the objective function. .in, It is a straight line in space To the target of function prediction Location at any moment vertical distance, These are the weights assigned to that data point. The goal is to minimize... Weight The dynamic allocation rule is as follows: if the confidence level of the timestamp extracted by a node is high, it is assigned a higher weight; if the residual of the two-dimensional coordinate solution of that node is small, it is also assigned a higher weight. Conversely, for data points with low confidence or large residuals, their weight is reduced to minimize their negative impact on the overall solution result. By solving this weighted nonlinear least squares problem, the optimal parameters describing the function of position changing with time can be obtained, and thus the three-dimensional spatial position at any time can be calculated.

[0038] In this embodiment, the reconstructed velocity vector of the target pellet is specifically as follows: 1. After obtaining the sequence of the three-dimensional spatial position of the target pellet as a function of time, the instantaneous velocity sequence is calculated by performing time differentiation on the position sequence. 2. An adaptive filtering algorithm based on a kinematic model is used to smooth the instantaneous velocity sequence in order to suppress measurement noise and meet physical constraints, and output the smoothed velocity vector.

[0039] Specifically, a series of discrete time points are obtained through the above optimization. Corresponding three-dimensional position coordinates Then, a position sequence is obtained. Differentiating the position sequence over time involves calculating the ratio of the change in position between adjacent time points to the time interval. This yields a series of discrete instantaneous velocity sequences. The adaptive filtering algorithm based on kinematic models incorporates a reasonable kinematic model to constrain the rationality of velocity changes. It can estimate the noise level in real time based on the residuals of the measurement data and adaptively adjust the smoothing intensity of the filter. While filtering out high-frequency noise, the algorithm can effectively preserve the true trend of motion changes, and finally output a velocity vector curve that changes continuously and smoothly over time, with each value containing the magnitude and direction of the velocity.

[0040] In this embodiment, the flight attitude of the target pellet is reconstructed as follows: 1. Based on the surface geometric characteristics of the target pellet, establish a geometric relationship model between the surface marker points and their projections on at least two non-parallel light curtain planes; 2. Obtain the two-dimensional position coordinates of different parts of the target surface passing through the light curtain, measured at similar times by different measurement nodes; 3. Substitute the two-dimensional position coordinates into the geometric relationship model to perform inverse kinematics and calculate the roll, pitch and yaw angles of the target.

[0041] Specifically, the flight attitude of a target pellet refers to its rotational state around its center of mass, described by roll, pitch, and yaw angles. Reconstructing the attitude requires utilizing the surface geometry of the target pellet. For example, assuming the pellet is spherical but has visible asymmetric markings on its surface; or, for a non-spherical pellet, its shape itself is a feature. First, a three-dimensional model of these feature points is established in the target pellet coordinate system. Then, a geometric model is established showing the projection of these feature points onto detection light curtains at different spatial orientations when the target pellet flies in a specific attitude. This model is essentially a set of coordinate transformation equations that rotate (defined by roll, pitch, and yaw angles) and translate (defined by the target pellet's center of mass position) the coordinates of the feature points in the target pellet coordinate system, and then project them onto the respective light curtain planes to obtain the expected two-dimensional projected coordinates. In actual measurements, when the target pellet flies past multiple nodes, different parts of its surface (corresponding to feature points) may occlude the light curtains of different nodes at different moments. The system acquires these temporally close two-dimensional position coordinates and identifies which feature point they belong to. Finally, these measured two-dimensional coordinate sets, along with the known trajectory of the target's center of mass (obtained from position reconstruction), are substituted into the aforementioned geometric model for inverse kinematics. By minimizing the difference between the model's predicted projection and the measured projection, the most likely roll, pitch, and yaw angles of the target at the current moment can be calculated.

[0042] The above steps, through recursive calculation and dynamic weighted optimization, effectively improve the real-time performance, accuracy, and robustness to abnormal measurement data of trajectory reconstruction while satisfying spatiotemporal physical constraints; and realize the integrated calculation of all-dimensional motion parameters of high-speed flying target position, velocity, and attitude, providing a complete data foundation for accurate prediction and control.

[0043] It is understood that those skilled in the art can combine various implementation methods in the above embodiments under the guidance of the above examples to obtain technical solutions with multiple implementation methods.

[0044] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for real-time measurement of target trajectory based on multi-node orthogonal light curtains, characterized in that, include: At least two measurement nodes are set up along the flight path of the target pellet. Each measurement node includes two sets of orthogonally arranged linear array photoelectric sensors and corresponding collimated light sources to form a two-dimensional detection light curtain covering the cross-section of the flight path. As the target pellet passes through the two-dimensional detection light curtain at each measurement node in sequence, pulse signals generated by the two sets of linear array photoelectric sensors at each measurement node due to the target pellet blocking the light are collected. Based on the collected pulse signals, all channel signals are processed in parallel using a high-speed signal processing unit to extract the timestamp of the target passing through each of the two-dimensional detection light curtains in real time, and to extract the two-dimensional position coordinates of the target on each of the two-dimensional detection light curtain planes. Based on the known fixed coordinates of each measurement node in three-dimensional space and the corresponding timestamps and two-dimensional position coordinates extracted from each node, the three-dimensional spatial position, velocity vector and flight attitude of the target at each moment are reconstructed in real time through spatial line intersection and time difference algorithms.

2. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 1, characterized in that, After deploying at least two measurement nodes along the flight path of the target pellet and before performing target pellet measurements, the method further includes: The calibrator is used to pass through the two-dimensional probe light curtain of each measurement node sequentially along a preset calibration trajectory; The calibration signal generated when the calibration object passes through each of the two-dimensional detection light curtains is recorded, and the spatial plane equation and relative position relationship of each of the two-dimensional detection light curtains in a unified coordinate system are calculated based on the geometric relationship between the calibration signal recorded at each node and the preset calibration trajectory.

3. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 2, characterized in that, The formation of a two-dimensional detection light curtain covering the cross-section of the flight path includes: At the measurement node, a light-shielding plate is moved along the preset detection area of ​​the two-dimensional detection light curtain, and the output signal of the linear array photoelectric sensor is monitored in real time. Adjust at least one of the following based on the changing characteristics of the output signal: the luminous intensity of the collimated light source, the beam angle, or the operating parameters of the linear array photoelectric sensor, until the two-dimensional detection light curtain has uniform and consistent shading response characteristics within the preset detection area.

4. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 1, characterized in that, The acquisition of pulse signals generated by the two sets of linear array photoelectric sensors in each measurement node due to target blockage includes: During the acquisition process, the signal-to-noise ratio and pulse waveform of the output signal of each linear array photoelectric sensor are monitored; The signal amplifier gain of the corresponding sensor channel is dynamically adjusted based on the monitored signal-to-noise ratio, and the sampling clock frequency and trigger threshold of the high-speed signal processing unit are dynamically adjusted based on the monitored pulse waveform width.

5. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 1, characterized in that, The real-time extraction of the timestamp of the target pellet passing through each of the two-dimensional detection light curtains includes: It receives pulse signals generated by two sets of orthogonal linear array photoelectric sensors at each measurement node; The moment when the rising edge of each group of pulse signals reaches a preset threshold is taken as the first candidate moment, and the width of each group of pulse signals is calculated. Determine whether the time difference between two first candidate times corresponding to the same node is less than a preset threshold. If the time difference is less than a preset threshold, then the timestamp corresponding to the current node is calculated based on the two first candidate times. If the time difference is greater than or equal to a preset threshold, a timestamp corresponding to the current node is generated based on the width of the pulse signal and the measurement data of adjacent nodes.

6. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 5, characterized in that, The extraction of the two-dimensional position coordinates of the target pellet on each of the two-dimensional detection light curtain planes includes: The sensitive pixel range blocked by the target pellet is determined based on the pulse signals from the two sets of sensor channels. Based on the query of the pre-stored lookup table in the sensitive pixel interval, the initial coordinates with sub-pixel precision are obtained; Apply pre-stored correction parameters to the initial coordinates to obtain one-dimensional coordinates; By synthesizing two orthogonal one-dimensional coordinates, the two-dimensional position coordinates of the target on the two-dimensional detection light curtain plane are obtained.

7. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 1, characterized in that, The spatial line intersection and time difference algorithm adopts a recursive solution process based on time series constraints: Based on the two-dimensional position coordinates extracted from a measurement node and the spatial equation of the light curtain plane, a spatial straight line characterizing the possible position of the target pellet is determined. The spatial straight lines of subsequent measurement nodes are introduced in sequence, and spatiotemporal consistency is verified with the currently constructed target motion state estimation. When new node data is introduced, the optimal estimate of the target's three-dimensional position and velocity is updated by utilizing the velocity continuity constraint of the target's motion. The recursive solution process is executed after each acquisition of data from a new measurement node to achieve real-time incremental reconstruction of the trajectory.

8. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 7, characterized in that, The reconstruction of the target's three-dimensional spatial position at each moment specifically involves: In the recursive solution process, the three-dimensional spatial position of the target is expressed as a function that changes with time; An optimization problem is constructed and solved with the objective of minimizing the weighted sum of squared vertical distances from the spatial lines of each node to the function at the corresponding timestamp; the weights in the weighted sum of squares are dynamically allocated based on the confidence level of the timestamps extracted from each node and the residuals of the two-dimensional position coordinates.

9. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtains according to claim 8, characterized in that, The reconstructing of the velocity vector of the target pellet is specifically as follows: After obtaining the sequence of the three-dimensional spatial position of the target pellet changing with time, the position sequence is differentiated over time to calculate the instantaneous velocity sequence. An adaptive filtering algorithm based on a kinematic model is used to smooth the instantaneous velocity sequence in order to suppress measurement noise and meet physical constraints, and output the smoothed velocity vector.

10. The method for real-time measurement of target trajectory based on multi-node orthogonal light curtain according to claim 7, characterized in that, The reconstructing of the target's flight attitude specifically involves: Based on the surface geometric features of the target pellet, a geometric relationship model is established between the surface marker points and their projections on at least two non-parallel light curtain planes; Obtain the two-dimensional position coordinates of different parts of the target surface passing through the light curtain, measured at similar times by different measurement nodes; The two-dimensional position coordinates are substituted into the geometric relationship model for inverse solution to calculate the roll, pitch and yaw angles of the target.

Images