A visual axis dwell precision verification system and method for step gaze imaging

Abstract

The application discloses a boresight dwell accuracy verification system and method for step-by-step gaze imaging, and belongs to the technical field of photoelectric measurement. The verification method comprises the following steps: step 1, constructing a multi-source external optical verification link; step 2, controlling the measured step-by-step gaze imaging device to be in a static state, driving a fast reflecting mirror and a scanning mirror to input a known curve excitation, collecting spot centroid displacement by using a high-speed measurement camera, and establishing a static calibration coefficient of spot centroid pixel displacement and actual boresight deflection angle; step 3, synchronous interception of a dwell window time sequence; step 4, high-frequency transient angle time sequence solving; step 5, calculating the root mean square value and the peak-to-peak value of a two-dimensional time sequence of angle deviation in the dwell window time, judging whether a sub-micro-radian level dwell accuracy threshold is met or not, and switching a resolution target, performing MTF calculation based on an infrared camera to extract a target image, and realizing cross verification of a time domain trajectory and a frequency domain imaging quality.

Description

Technical Field

[0001] This invention belongs to the field of photoelectric measurement technology, specifically relating to a system and method for verifying the visual axis dwell accuracy of step-gaze imaging. Background Technology

[0002] To meet the demand for high-resolution, wide-swath measurements in long-distance, wide-area applications (such as natural disasters and precision agriculture), the "stamp-and-Stare" imaging technique has become an important development direction in the aerospace remote sensing field in recent years due to its combination of the high efficiency of push-broom and the wide swath of swing-broom. A typical step-and-stare imaging device relies on a large-inertia scanning mirror for wide-area stepping and uses a low-inertia fast mirror (FSM) for reverse high-frequency compensation during the short "stare" exposure to counteract image shift caused by continuous mirror movement and platform vibration.

[0003] As the resolution of optical systems continues to improve, extremely stringent requirements are placed on the line-of-sight (LOS) stability of step-gazing systems. For example, in some wide-area high-resolution imaging tasks, the system must achieve a line-of-sight dwell accuracy better than 1 μrad after image shift compensation within an extremely short dwell time (e.g., 5 ms).

[0004] However, existing methods for verifying and evaluating line-of-sight accuracy have the following significant shortcomings: (1) Internal sensor feedback method: It usually relies on the data of photoelectric encoder, eddy current sensor or gyroscope inside the scanning platform or FSM for calculation. However, due to the deformation of flexible hinge, mechanical structure resonance and assembly stress, the readings of internal sensors cannot truly represent the actual physical deflection of the final output optical axis, and there is a "blind zone".

[0005] (2) Indirect MTF evaluation method: For example, Sun Chongshang et al. used the slanted-edge method to measure the overall MTF in "Backscanning step and stareimaging system..." (Applied Optics, 2015). The MTF method is a macroscopic manifestation of the time integration effect and cannot reflect the transient time-domain characteristics and spectral distribution of the line-of-sight jitter within a very short dwell time (5ms), making it difficult to use for fine-grained error tracing.

[0006] (3) Traditional autocollimator method: Although commercial high-precision autocollimators have sub-micro-radian level measurement accuracy, their sampling bandwidth is usually low, which makes it impossible to effectively capture and separate the dramatic changes in the time domain between "large-range high dynamic step" and "small-angle high-frequency dwell".

[0007] In summary, there is an urgent need for a method and system for directly verifying the sub-micro-radian level visual axis dwell accuracy that is independent of the internal control loop of the step-gazing imaging device under test, has high-frequency dynamic measurement capabilities, and can perform precise time slicing for extremely short dwell windows. Summary of the Invention

[0008] To overcome the shortcomings of existing technologies, the present invention aims to provide a visual axis dwell accuracy verification system and method for step-gazing imaging, in order to solve the technical problem that existing testing methods cannot directly measure and verify sub-micro-radius level high-frequency transient visual axis jitter in a large dynamic two-mirror combined system within an extremely short dwell window (such as 5ms).

[0009] To achieve the above objectives, the present invention adopts the following specific technical solutions: On one hand, the present invention provides a visual axis dwell accuracy verification system for step-gaze imaging, wherein the step-gaze imaging device includes a main scanning mirror and a fast-reflecting mirror, and the verification system includes: The optical simulation and measurement module is used to construct the external standard optical verification environment of the tested step-gazing imaging device under laboratory conditions. It includes a large-aperture collimator, a precision optical platform, an autocollimator, a laser point source component, and a target component located on the focal plane of the collimator. The device under test fixture and environmental monitoring module are used for stress-free clamping of the stepping staring imaging device under test and for monitoring ground micro-vibration interference. The vision acquisition module is used to capture high dynamic optical signals, including a high-speed measuring camera located at the focal plane of the collimator and an infrared camera located at the focal plane of the stepping staring imaging device under test. The synchronous control and data processing hub is electrically connected to the controllers of the laser point source component, the high-speed measuring camera, and the measured stepping staring imaging device, respectively. It is used to uniformly issue clock beats, execute centroid extraction algorithms, and output dwell accuracy verification reports. It includes a programmable timing synchronizer and a data processing host computer.

[0010] Furthermore, the optical verification environment constructed by the optical simulation and measurement module includes at least the following: forming a collimated beam in front of the entrance pupil of the step-gazing imaging device under test using a large-aperture collimator to simulate the incident conditions of distant or infinitely distant targets; providing resolution targets, edge targets, or point targets through a target assembly located on the focal plane of the collimator to achieve image quality testing of the imaging link and verification of imaging quality within the dwell window; constructing a reverse or equivalent optical measurement link for measuring the visual axis deflection using a laser point source assembly in conjunction with a high-speed measurement camera to obtain the change in the position of the light spot during dwell; and performing angular reference measurement on the step-gazing imaging device under test or its reference reflective surface using an autocollimator to provide an external angular reference for the mapping relationship between pixel displacement and visual axis angle.

[0011] On the other hand, the present invention provides a method for verifying visual axis dwell accuracy for step-gaze imaging, which is based on the above-mentioned system for verifying visual axis dwell accuracy for step-gaze imaging, and includes the following steps: Step 1: Construct a multi-source external optical verification link: Set a collimator at the entrance pupil of the stepping staring imaging device under test, and switchably set a high-speed measuring camera or resolution target connected to an external synchronous controller at the focal plane of the collimator; set a laser point source at the focal plane of the stepping staring imaging device under test. Step 2: Establish the calibration mapping from pixel displacement to visual axis angle: Control the measured step staring imaging device to be static, drive the fast reflector and scanning mirror to input the known curve excitation, use the high-speed measuring camera to collect the displacement of the spot centroid, and establish the static calibration coefficient between the pixel displacement of the spot centroid and the actual visual axis deflection angle. Step 3, Dwelling Window Timing Synchronization: The external synchronization controller acquires the scan retrace synchronization signal of the stepping staring imaging device under test, and sends a trigger pulse with time delay compensation to the high-speed measuring camera, so that the high-speed measuring camera performs ultra-high frame rate continuous image acquisition only within the "dwelling exposure window" of the stepping staring imaging cycle. Step 4: High-frequency transient angle time series calculation: Extract the two-dimensional centroid coordinates of the laser spot in each frame of the image within the dwell exposure window, and combine them with the static calibration coefficients to calculate the two-dimensional time series of the line-of-sight angle deviation within the dwell window. Step 5: Multidimensional evaluation and cross-validation of dwell accuracy: Calculate the root mean square value and peak-to-peak value of the two-dimensional time series of angle deviation within the dwell window time to determine whether it meets the sub-microradian level dwell accuracy threshold; and switch the resolution target, perform MTF calculation based on the target image extracted by the infrared camera to achieve cross-validation of time domain trajectory and frequency domain imaging quality.

[0012] Furthermore, the specific content of step 2 is as follows: Step 2.1: When the tested step-gazing imaging device is in a static or quasi-static state, control the fast reflector and the main scanning mirror to operate individually or in a predetermined combination, and input multiple sets of known angle excitation signals. The excitation signals are step signals, ramp signals, sine signals, or piecewise linear scanning signals. The predetermined combination of actions includes: the main scanning mirror maintaining a zero position while the fast reflector performs single-axis or dual-axis scanning; the fast reflector maintaining a zero position while the main scanning mirror performs low-speed scanning; and the main scanning mirror and the fast reflector operating synchronously in phase, out of phase, or with staggered timing. Step 2.2: Under each set of known angle excitations, obtain the corresponding actual line-of-sight deflection angle reference value using an autocollimator or reference measuring device, and at the same time, acquire the spot image sequence at the corresponding moment using a high-speed measuring camera. Step 2.3: Preprocess the acquired spot images. The preprocessing includes background subtraction, dark field correction, threshold segmentation, region of interest extraction, and abnormal frame removal. Step 2.4: Perform centroid extraction and subpixel fitting on the preprocessed spot images to obtain the center coordinates of the spot in each frame; Step 2.5: Using the center coordinates of the spot in the zero-position reference frame as a reference, calculate the pixel displacement of the spot under each excitation condition. Step 2.6: Fit the pixel displacement of the spot with the corresponding actual visual axis deflection angle reference value to establish a calibration mapping model of pixel displacement to visual axis angle; the calibration mapping model adopts a one-dimensional scaling factor model, a two-dimensional linear decoupling model, a two-dimensional coupling model with cross terms, or a low-order polynomial model. Step 2.7: Perform residual verification on the fitted calibration mapping model, and use the residuals that meet the preset threshold as the calibration result used for subsequent dwell window angle calculation.

[0013] Furthermore, the calibration mapping model for pixel displacement to viewing axis angle is expressed as follows: the displacement of the spot in the two-dimensional image coordinate system is mapped to the deflection angle of the viewing axis in two orthogonal directions, while simultaneously compensating for pixel size error, imaging magnification error and coordinate axis non-orthogonality error.

[0014] Further, in step 3, the synchronization control and data processing center acquires synchronization reference signals characterizing the scanning cycle, retrace start point, dwell start point, and exposure interval from the tested step-gazing imaging device, as well as the transmission delay and trigger response delay between components; based on the timing parameters of the tested step-gazing imaging device within a single cycle, it generates a gated acquisition command corresponding to the dwell exposure window, the gated acquisition command including at least the acquisition start time, acquisition duration, acquisition frame rate, and trigger advance; the synchronization control and data processing center outputs a trigger pulse to the high-speed measurement camera, so that the high-speed measurement camera only starts continuous acquisition within the corresponding dwell exposure window.

[0015] Furthermore, the specific content of step 4 is as follows: Step 4.1: Perform background suppression, threshold segmentation, and region of interest cropping on the spot images of each frame continuously acquired within the dwell exposure window; Step 4.2: Perform centroid extraction on the spot points in each frame image based on the sub-pixel localization algorithm to obtain the two-dimensional center coordinates of the spot points in each frame; Step 4.3: Using a pre-set zero-position reference frame as a reference, calculate the two-dimensional pixel displacement of the center of the spot in each frame relative to the reference position. Step 4.4: Substitute the two-dimensional pixel displacement of each frame into the pixel displacement to view axis angle calibration mapping model established in Step 2, and calculate the two-dimensional view axis angle deviation of the corresponding frame. Step 4.5: Sort the two-dimensional view angle deviation of each frame according to a unified timestamp to form a two-dimensional view angle time series within the dwell exposure window; Step 4.6: When the sampling clock of the high-speed measuring camera is inconsistent with the control clock of the measured step-gazing imaging device, the two-dimensional viewing axis angle time series is interpolated, resampled, time-base corrected, or filtered to obtain a high-frequency transient angle time series that strictly corresponds to the dwell window.

[0016] Furthermore, the centroid extraction employs a gray-scale weighted centroid algorithm, specifically: the coordinates are weighted and averaged based on the gray values ​​of each pixel within the spot area to obtain the center position of the spot; a two-dimensional Gaussian function is used to fit the initial centroid value to obtain the sub-pixel level center coordinates of the spot.

[0017] Furthermore, in step 5, cross-validation includes: using the MTF result obtained under static conditions as a benchmark value, comparing the MTF result obtained under dynamic dwell conditions with the benchmark value, and combining the root mean square value or peak-to-peak value change of the two-dimensional line-of-sight angle time series within the dwell window to establish a correspondence between "angle perturbation and imaging degradation", thereby verifying whether the dwell accuracy result obtained by the spot trajectory calculation is consistent with the actual imaging quality change.

[0018] Furthermore, the specific content of step 5 is as follows: Step 5.1: Calculate the mean, root mean square value, peak-to-peak value, maximum absolute deviation value, and standard deviation for the two-dimensional time series of line-of-sight angle deviation within the dwell window obtained in Step 4.

[0019] Step 5.2: Compare the time-domain evaluation index with the preset sub-microradian level dwell accuracy threshold to determine whether the tested step-gazing imaging device meets the visual axis dwell accuracy requirements under the current operating conditions.

[0020] The time-domain evaluation metrics include at least the root mean square (RMS), peak-to-peak (peak-to-peak), maximum absolute deviation, and standard deviation of the line-of-sight angle deviations in the x and y directions. Based on the pre-set performance parameters of the device under test, preferably, the RMS values ​​in the x and y directions within a 5 ms dwell exposure window should not exceed 1 μrad, and the peak-to-peak values ​​should not exceed 4 μrad.

[0021] Step 5.3: While maintaining the scanning conditions and synchronization sequence of the stepping staring imaging device under test, switch the high-speed measuring camera at the focal plane of the collimator to a resolution target, a beveled target, or a striped target, and use the infrared camera at the focal plane of the stepping staring imaging device under test to acquire the corresponding target image. The two focal planes are in different positions and are used for different working modes.

[0022] Step 5.4: Evaluate the imaging quality of the target image acquired by the infrared camera: When using a slanted target, perform edge spread function extraction, line spread function calculation and Fourier transform to obtain the modulation transfer function (MTF) curve under the corresponding working condition; when using a striped target or a resolution target, calculate the modulation index, contrast transfer value or equivalent MTF index at the target spatial frequency as the imaging domain evaluation result under the corresponding working condition.

[0023] Step 5.5: Correlate the angular stability evaluation results within the dwell window with the MTF results under the corresponding working conditions to determine the degree of influence of changes in the dwell accuracy of the line of sight on the image sharpness.

[0024] Step 5.6: When both the angle domain evaluation result and the imaging domain evaluation result meet the preset criteria, output the conclusion that the tested step-gazing imaging device has passed the sub-micro-radian level visual axis dwell accuracy verification.

[0025] The preset criteria include: the root mean square values ​​of the x and y directions within the 5ms dwell exposure window in the angle domain are not greater than 1 μrad and the peak-to-peak values ​​are not greater than 4 μrad; the decrease ratio of dynamic MTF50 to static MTF50 in the imaging domain is not greater than 10%, or the decrease ratio of MTF at the target working space frequency is not greater than 10%. The output concludes that the tested step-gazing imaging device has passed the sub-microradian level visual axis dwell accuracy verification. The conclusion includes at least three types of output: "Verification passed," "Verification failed," or "Further optimization and retesting required."

[0026] Compared with the prior art, the present invention has the following advantages: (1) Breaking through the testing bottleneck of extremely short dwell window: This invention uses a nanosecond-level programmable timing synchronization mechanism to closely align the sampling window of the external testing instrument with the extremely short "dwell time (e.g. 5ms)" of the step-gazing imaging device under test, filtering out the large-angle interference of the "scanning step period" and realizing the focused measurement of transient small errors.

[0027] (2) Realize physical truth traceability at the sub-micro-radian level: This invention abandons the approach of relying on internal sensors and uses a pure optical external circuit test of "large-diameter parallel light tube + point light source + sub-pixel centroid extraction" to truly restore the physical deflection of the outgoing optical axis. Combined with the sub-pixel algorithm, its physical verification resolution can reach the micro-radian level or even the sub-micro-radian level.

[0028] (3) Pioneering decoupled multi-layer verification mechanism: This invention provides a full-chain layered verification process from single fast mirror, single scanning mirror to two mirrors combined. Combining the static high precision of the autocollimator with the dynamic high bandwidth of the high-speed camera, and the final imaging effect of the MTF, a complete evidence chain of "time domain trajectory - static benchmark - final image quality verification" is formed.

[0029] (4) Compared with the existing verification schemes that indirectly evaluate step staring imaging devices based solely on imaging sharpness, resolution results or a single MTF index, this application constructs a joint verification link of the angle domain, image domain and imaging domain under a unified clock reference. This allows for the direct extraction of high-frequency time series of visual axis deviation within the dwell exposure window. Furthermore, through external angle reference calibration and cross-verification of imaging quality, the credibility, repeatability and engineering applicability of the dwell accuracy verification results at the sub-micro-radian level are improved. Attached Figure Description

[0030] Figure 1 This is a system block diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0032] Example 1

[0033] This embodiment provides a visual axis dwell accuracy verification system for step-by-step staring imaging, wherein the step-by-step staring imaging device includes a main scanning mirror and a fast-reflecting mirror. For example... Figure 1 As shown, the verification system includes an optical simulation and measurement module, a device under test fixture and environmental monitoring module, a vision acquisition module, and a synchronous control and data processing center.

[0034] (1) The optical simulation and measurement module is used to construct the external standard optical verification environment of the step-gazing imaging device under laboratory conditions. It includes a large-aperture collimator, a precision optical platform, an autocollimator, a laser point source component, and a target component located on the focal plane of the collimator.

[0035] The optical simulation and measurement module is used to construct an external standard optical verification environment for the tested step-gazing imaging device under laboratory conditions. Its functions include at least: 1) forming a collimated beam in front of the entrance pupil of the tested step-gazing imaging device using a large-aperture collimator to simulate the incident conditions of distant or infinitely distant targets; 2) providing resolution targets, edge targets, or point targets through a target assembly located on the focal plane of the collimator to achieve image quality testing of the imaging chain and verification of image quality within the dwell window; 3) cooperating with a high-speed measurement camera through a laser point source assembly. 4) Construct a reverse or equivalent optical measurement link for viewing axis deflection measurement to obtain the change in spot position during dwell time; 5) Perform angular reference measurement on the tested step staring imaging device or its reference reflective surface using an autocollimator to provide an external angular reference for the mapping relationship between pixel displacement and viewing axis angle; 6) Ensure the relative position stability and coaxiality between the collimator, autocollimator, target assembly and the tested step staring imaging device through a precision optical platform, thereby providing a stable and repeatable measurement basis for verifying sub-micro-radian dwell accuracy.

[0036] (2) The device under test fixture and environmental monitoring module are used to stress-free clamp the stepping staring imaging device under test and monitor the micro-vibration interference of the foundation.

[0037] (3) The visual acquisition module is used to capture high dynamic optical signals, including a high-speed measurement camera located at the focal plane of the collimator and an infrared camera located at the focal plane of the stepping staring imaging device under test. The high-speed measurement camera is used to acquire a high dynamic spot sequence within the dwell exposure window and output centroid displacement information in the image domain; the infrared camera is used to acquire target images and output edge spread function, line spread function and MTF results in the imaging domain. The former solves the problem of "how much deviation is actually within the exposure window", and the latter solves the problem of "whether this deviation has caused perceptible image degradation".

[0038] (4) The synchronization control and data processing center is electrically connected to the controllers of the laser point source component, the high-speed measuring camera, and the measured step-gazing imaging device, respectively. It is used to uniformly issue clock beats, execute centroid extraction algorithms, and output dwell accuracy verification reports. It includes a programmable timing synchronizer and a data processing host computer. Specifically, the synchronization control and data processing center receives scanning cycle signals, retrace synchronization signals, dwell start and end signals, or exposure trigger signals from the device under test, and uniformly generates external trigger beats so that the high-speed measuring camera only collects data within the dwell exposure window. Afterward, the synchronization control and data processing center performs background subtraction, centroid extraction, subpixel fitting, pixel-angle mapping, time alignment, and result judgment on the high-speed camera image, and performs correlation analysis with the MTF results output by the infrared camera. The synchronization control and data processing center is not a simple data recorder, but rather integrates the three tasks of "synchronous acquisition - unified processing - cross-validation" into the same system.

[0039] It is understood that the innovation of this application does not lie solely in the use of any single device among collimators, autocollimators, high-speed measuring cameras, or synchronous controllers, but rather in the construction of a multi-domain joint verification system around the specific working stage of the step-gazing imaging device's "dwelling exposure window." This system comprises an optical simulation and measurement module, a visual acquisition module, and a synchronous control and data processing center. Under a unified clock reference, this system can simultaneously acquire angle domain, image domain, and imaging domain data for the same measured step-gazing imaging process and complete cross-verification of the visual axis dwell accuracy through the same solution model.

[0040] Angular domain data: the amount of line-of-sight deflection supported by an external angular reference and a pixel-angle calibration model.

[0041] Image domain data: Two-dimensional trajectory of the centroid of the light spot within the dwell window acquired by a high-speed measurement camera.

[0042] Imaging domain data: Target images and MTF results within the dwell window obtained by the infrared camera.

[0043] These three types of data are not mined and calculated separately, but are acquired synchronously within the same cycle, the same time series, and the same dwell window. Then, through a unified mapping and judgment model, a cross-validation conclusion is output on whether the line-of-sight dwell accuracy meets the requirements.

[0044] In one specific embodiment, a laser point source component is positioned at the focal plane of the step-gazing imaging device under test, providing a point source signal to the device. The point source signal, after being reflected by the imaging optical path of the device and the main scanning mirror and fast-reflecting mirror, exits from the front end of the device and enters a large-aperture collimator. A high-speed measuring camera is positioned at the focal plane of the collimator to receive the point spot image converged by the collimator. The high-speed measuring camera and the target component are switchably positioned at the focal plane of the collimator for use in both the line-of-sight angle measurement mode and the imaging mode. The system switches between quality verification modes; the autocollimator establishes an optical correspondence with the device under test or its reference reflector to output angular reference quantities; the synchronization control and data processing center is electrically connected to the laser point source component, the high-speed measuring camera, and the controller of the stepping staring imaging device under test, respectively, to receive the scanning retrace synchronization signal, issue a unified trigger clock, and complete data time alignment; the device under test fixture and environmental monitoring module are used to fix the stepping staring imaging device under test without stress and synchronously record the foundation micro-vibration, temperature, or external disturbance information to eliminate measurement deviations caused by factors other than those of the device under test itself.

[0045] This system utilizes the same external verification platform and the same time series reference to integrate "spot displacement measurement - angle mapping calculation - dwell window interception - MTF cross-verification" into an integrated verification scheme, thereby solving the problem that existing technologies can only perform single imaging evaluation and are difficult to directly obtain the sub-microradian level line-of-sight angle time series within the dwell window.

[0046] Example 2

[0047] This embodiment provides a method for verifying visual axis dwell accuracy in step-gaze imaging, implemented based on the visual axis dwell accuracy verification system for step-gaze imaging described in Embodiment 1. Figure 2 As shown, the verification method of the present invention includes the following steps: (a) Step 1: Construct a multi-source external optical verification link: Set a collimator at the entrance pupil of the stepping staring imaging device under test, and switchably set a high-speed measuring camera or resolution target connected to an external synchronous controller at the focal plane of the collimator; set a laser point source at the focal plane of the stepping staring imaging device under test.

[0048] (ii) Step 2: Establish the calibration mapping of pixel displacement to visual axis angle: Control the measured step staring imaging device to be static, drive the fast reflector and scanning mirror to input the known curve excitation, use the high-speed measuring camera to collect the centroid displacement of the spot, and establish the static calibration coefficient of the pixel displacement of the centroid of the spot and the actual visual axis deflection angle.

[0049] The specific steps for this procedure are as follows: Step 2.1: With the tested step-gazing imaging device in a static or quasi-static state, control the fast-reflecting mirror and the main scanning mirror to operate individually or in a predetermined combination, and input multiple sets of known angle excitation signals. The excitation signals are step signals, ramp signals, sinusoidal signals, or piecewise linear scanning signals. The predetermined combination of actions includes: the main scanning mirror maintaining a zero position while the fast-reflecting mirror performs single-axis or dual-axis scanning; the fast-reflecting mirror maintaining a zero position while the main scanning mirror performs low-speed scanning; and the main scanning mirror and the fast-reflecting mirror operating synchronously in phase, out of phase, or with staggered timing.

[0050] Step 2.2: Under each set of known angle excitations, obtain the corresponding actual line-of-sight deflection angle reference value using an autocollimator or reference measuring device, and simultaneously acquire the spot image sequence at the corresponding moment using a high-speed measuring camera.

[0051] Step 2.3: Preprocess the acquired spot images. The preprocessing includes background subtraction, dark field correction, threshold segmentation, region of interest extraction, and abnormal frame removal.

[0052] Step 2.4: Perform centroid extraction and subpixel fitting on the preprocessed spot images to obtain the center coordinates of the spot in each frame.

[0053] Step 2.5: Using the center coordinates of the spot in the zero-position reference frame as a reference, calculate the pixel displacement of the spot under each excitation condition.

[0054] Step 2.6: Fit the pixel displacement of the spot with the corresponding actual visual axis deflection angle reference value to establish a calibration mapping model of pixel displacement to visual axis angle; wherein, the calibration mapping model adopts a one-dimensional scaling factor model, a two-dimensional linear decoupling model, a two-dimensional coupling model with cross terms, or a low-order polynomial model.

[0055] The calibration mapping model for pixel displacement to viewing axis angle is expressed as follows: the displacement of the point spot in the two-dimensional image coordinate system is mapped to the deflection angle of the viewing axis in two orthogonal directions, while simultaneously compensating for pixel size error, imaging magnification error and coordinate axis non-orthogonality error.

[0056] Step 2.7: Perform residual verification on the fitted calibration mapping model, and use the residuals that meet the preset threshold as the calibration result used for subsequent dwell window angle calculation.

[0057] (III) Step 3, Dwelling window timing synchronization capture: The external synchronization controller acquires the scanning retrace synchronization signal of the stepping staring imaging device under test, and sends a trigger pulse with time delay compensation to the high-speed measuring camera, so that the high-speed measuring camera can only perform ultra-high frame rate continuous image acquisition within the "dwelling exposure window" of the stepping staring imaging cycle.

[0058] The synchronization control and data processing center acquires synchronization reference signals characterizing the scan cycle, retrace start point, dwell start point, dwell end point, and exposure time from the tested step-gazing imaging device, and measures the inherent transmission delays in each link, including controller output delay, cable transmission delay, camera trigger response delay, and acquisition buffer delay. Based on the timing parameters of the tested step-gazing imaging device within a single cycle, a gated acquisition command corresponding to the dwell exposure window is generated. The gated acquisition command includes at least the acquisition start time, acquisition duration, acquisition frame rate, and trigger advance. The synchronization control and data processing center outputs a trigger pulse to the high-speed measurement camera, causing the high-speed measurement camera to start continuous acquisition only within the corresponding dwell exposure window, thereby avoiding the mixing of irrelevant motions introduced by the acceleration, deceleration, and retrace phases into the dwell accuracy calculation.

[0059] (iv) Step 4, High-frequency transient angle time series calculation: Extract the two-dimensional centroid coordinates of the laser spot of each frame image within the dwell exposure window, and combine them with the static calibration coefficient to calculate the two-dimensional time series of the line-of-sight angle deviation within the dwell window.

[0060] The specific steps for this operation are as follows: Step 4.1: Perform background suppression, threshold segmentation, and region of interest cropping on the spot images of each frame continuously acquired within the dwell exposure window.

[0061] Step 4.2: Perform centroid extraction on the spot points in each frame image based on the sub-pixel localization algorithm to obtain the two-dimensional center coordinates of the spot points in each frame. The centroid extraction adopts the gray-level weighted centroid algorithm, which is as follows: the coordinates are weighted and averaged according to the gray values ​​of each pixel in the spot area to obtain the center position of the spot; based on the initial centroid value, a two-dimensional Gaussian function is used for fitting to obtain the sub-pixel level center coordinates of the spot points.

[0062] Step 4.3: Using the pre-set zero-position reference frame as a reference, calculate the two-dimensional pixel displacement of the center of the spot in each frame relative to the reference position.

[0063] Step 4.4: Substitute the two-dimensional pixel displacement of each frame into the pixel displacement to view axis angle calibration mapping model established in Step 2, and calculate the two-dimensional view axis angle deviation of the corresponding frame.

[0064] Step 4.5: Sort the two-dimensional view angle deviation of each frame according to a unified timestamp to form a two-dimensional view angle time series within the dwell exposure window.

[0065] Step 4.6: When the sampling clock of the high-speed measuring camera is inconsistent with the control clock of the measured step-gazing imaging device, the two-dimensional viewing axis angle time series is interpolated, resampled, time-base corrected, or filtered to obtain a high-frequency transient angle time series that strictly corresponds to the dwell window.

[0066] (V) Step 5, Multidimensional evaluation and cross-validation of dwell accuracy: Calculate the root mean square value and peak-to-peak value of the two-dimensional time series of angle deviation within the dwell window time, and determine whether it meets the dwell accuracy threshold of sub-microradian level; and switch the resolution target, extract the target image based on the infrared camera to perform MTF calculation, and realize the cross-validation of time domain trajectory and frequency domain imaging quality.

[0067] The specific details of this step are as follows: Step 5.1: Calculate the mean, root mean square value, peak-to-peak value, maximum absolute deviation value and standard deviation for the two-dimensional time series of line-of-sight angle deviation within the dwell window obtained in Step 4, and use these as the evaluation results for the angle domain.

[0068] Step 5.2: Compare the temporal evaluation metrics with the preset sub-microradian level dwell accuracy threshold to determine whether the tested step-gazing imaging device meets the visual axis dwell accuracy requirements under the current operating conditions. The temporal evaluation metrics include at least the root mean square value, peak-to-peak value, maximum absolute deviation value, and standard deviation of the visual axis angle deviation in the x and y directions. Based on the preset task indicators of the tested device, preferably, the root mean square values ​​in the x and y directions should not exceed 1 μrad, and the peak-to-peak values ​​should not exceed 4 μrad within a 5 ms dwell exposure window.

[0069] Step 5.3: While maintaining the scanning conditions and synchronization sequence of the stepping staring imaging device under test, switch the high-speed measuring camera at the focal plane of the collimator to a resolution target, a beveled target, or a striped target, and use the infrared camera at the focal plane of the stepping staring imaging device under test to acquire the corresponding target image. The two focal planes are in different positions and are used for different working modes.

[0070] Step 5.4: Evaluate the imaging quality of the target image acquired by the infrared camera: When using a slanted target, perform edge spread function extraction, line spread function calculation and Fourier transform to obtain the modulation transfer function (MTF) curve under the corresponding working condition; when using a striped target or a resolution target, calculate the modulation index, contrast transfer value or equivalent MTF index at the target spatial frequency as the imaging domain evaluation result under the corresponding working condition.

[0071] Step 5.5: Correlate the angular stability evaluation results within the dwell window with the MTF results under the corresponding working conditions to determine the degree of influence of changes in the dwell accuracy of the line of sight on the image sharpness.

[0072] Step 5.6: When both the angle domain evaluation result and the imaging domain evaluation result meet the preset criteria, output the conclusion that the tested step-gazing imaging device has passed the sub-micro-radian level visual axis dwell accuracy verification.

[0073] The preset criteria include: the root mean square values ​​of the x and y directions within the 5ms dwell exposure window in the angle domain are not greater than 1 μrad and the peak-to-peak values ​​are not greater than 4 μrad; the decrease ratio of dynamic MTF50 to static MTF50 in the imaging domain is not greater than 10%, or the decrease ratio of MTF at the target working space frequency is not greater than 10%. The output concludes that the tested step-gazing imaging device has passed the sub-microradian level visual axis dwell accuracy verification. The conclusion includes at least three types of output: "Verification passed," "Verification failed," or "Further optimization and retesting required."

[0074] Cross-validation includes: using the MTF result obtained under static conditions as a benchmark, comparing the MTF result obtained under dynamic dwell conditions with the benchmark value, and combining the root mean square value or peak-to-peak value change of the two-dimensional line-of-sight angle time series within the dwell window to establish a "angle perturbation-imaging degradation" correspondence, thereby verifying whether the dwell accuracy result obtained from the spot trajectory calculation is consistent with the actual image quality change. This application does not use MTF to replace angle measurement, but uses MTF to cross-validate the angle calculation result; nor does it use an external angle benchmark to replace the imaging result, but allows the external angle benchmark, spot trajectory, and imaging quality to mutually verify each other.

[0075] Example 3

[0076] Implementation of multi-domain joint verification for "stamp-style" backscanning step-gaze scenarios The test device is a step-gaze imaging apparatus equipped with a main scanning mirror and a fast-reflecting mirror. This apparatus is used for "stamp-style" wide-area scanning imaging. The main scanning mirror handles the large-stroke sweeping, while the fast-reflecting mirror compensates for high-frequency image shift within the dwell exposure window. The design conditions of the apparatus under test are: the main scanning mirror sweeps approximately 10°, completing 15 images in a single scan cycle, with a total cycle time of approximately 0.7 s. The required line-of-sight dwell accuracy within a single exposure window is better than 1 μrad@5 ms.

[0077] In this embodiment, a collimator is arranged at the front end of the device under test (DUT), and a switchable high-speed measurement camera and target assembly are set at the focal plane of the collimator. A laser point source assembly is set at the focal plane of the DUT's step-gazing imaging device. An autocollimator is also arranged to the side of the DUT to establish an external angular reference. A synchronous control and data processing system is connected to the controller of the DUT's step-gazing imaging device, the high-speed measurement camera, and the laser point source, respectively, for unified clocking, unified triggering, and unified data processing. During verification, the angle trajectory measurement mode is first entered, and the high-speed measurement camera is arranged at the focal plane of the collimator to record the spot trajectory within the dwell window. Then, the imaging quality verification mode is entered, and the high-speed measurement camera is switched to the target assembly. Simultaneously, an infrared camera at the focal plane of the DUT's step-gazing imaging device acquires an image of the oblique target and calculates the MTF. By using the same scanning conditions, the same fast-reaction compensation parameters, and the same dwell window definition for both acquisitions, corresponding analysis of the angle domain and imaging domain results under the same conditions is achieved.

[0078] In this embodiment, the single-cycle motion of the main scanning mirror can be divided into an acceleration phase, a staring phase, a deceleration phase, and a reset phase. The synchronization controller only sends a gating signal to the high-speed measurement camera during the staring phase, ensuring that the high-speed measurement camera only records data corresponding to the dwell exposure window and does not collect data during the acceleration and reset phases. Taking a single effective exposure window of 5ms as an example, to ensure that the time series within the window has sufficient discrete sampling points, the sampling frame rate of the high-speed measurement camera can be set to 20kfps. At this time, approximately 100 frames of spot images can be obtained within a 5ms exposure window; if set to 10kfps, approximately 50 frames of images can also be obtained within an exposure window. This allows for the direct acquisition of the high-frequency transient line-of-sight deviation time series within the dwell window, rather than just obtaining a static result at a certain moment.

[0079] During the specific calibration phase, the tested step-gazing imaging device is placed in a stationary or quasi-stationary state. Known angular excitations are sequentially input to the main scanning mirror and the fast-reflecting mirror, such as step or ramp excitations of ±5μrad, ±10μrad, ±20μrad, and ±50μrad. For each set of excitations, an angle reference value is output by the autocollimator, and the high-speed measuring camera simultaneously acquires point spot images. After background subtraction and subpixel centroid extraction, the displacement of the point spot in the two-dimensional image coordinate system is obtained. For example, in a certain calibration set, a displacement of 0.12 pixels in the x-direction corresponds to a 1μrad line-of-sight deflection, and a displacement of 0.11 pixels in the y-direction corresponds to a 1μrad line-of-sight deflection. Further fitting yields a two-dimensional mapping matrix containing cross terms, used for subsequent angle conversion within the dwell window.

[0080] During the dwell condition verification phase, the synchronous controller triggers high-speed measurement camera sampling within the corresponding 5ms exposure window based on the dwell start signal output by the tested step-gazing imaging device. A continuous image sequence is obtained within a certain dwell window. After centroid extraction and pixel-angle mapping, a two-dimensional visual axis deviation time series is obtained. Taking a typical result from a prototype stage as an example, without fast mirror compensation, the root mean square value of the visual axis deviation in the x-direction within this dwell window is 3.8 μrad, with a peak-to-peak value of 11.6 μrad; the root mean square value in the y-direction is 2.9 μrad, with a peak-to-peak value of 8.7 μrad. After enabling fast mirror compensation, the root mean square value in the x-direction decreases to 0.62 μrad, and the peak-to-peak value decreases to 1.85 μrad; the root mean square value in the y-direction decreases to 0.57 μrad, and the peak-to-peak value decreases to 1.63 μrad, indicating that the system can directly characterize the improvement trend of visual axis stability within the dwell window before and after compensation. Therefore, under this working condition, the multi-domain joint verification system of this application can directly provide the two-dimensional line-of-sight angle time series within the dwell exposure window, and determine whether the device under test has reached the stage dwell accuracy target.

[0081] The experimental data such as 0.12 pixels / μrad here are just example values. The actual values ​​are determined by factors such as collimator parameters, camera pixel size, and optical path magnification.

[0082] Subsequently, while maintaining the same scanning parameters, fast-response compensation parameters, and dwell window settings, the high-speed measurement camera at the collimator focal plane was switched to an oblique target. Images were then acquired and the MTF calculated by the infrared camera at the focal plane of the step-gazing imaging device under test. In the prototype stage, the focus is initially on comparing the trends of the dynamic MTF curves before and after compensation, as well as their approximation to the static reference state. When the angular domain results show a significant convergence of the line-of-sight deviation within the dwell window, and the imaging domain results show a reduction in the degradation of the dynamic MTF relative to the static reference, it indicates that the verification scheme does not rely on a single link to reach a conclusion. Instead, it cross-validates the actual line-of-sight stability level within the dwell exposure window through consistency across the angular, image, and imaging domains.

[0083] Example 4

[0084] Implementation examples for the prototype stage In this embodiment, it is not required to achieve all the specifications of the final engineering prototype at once. Instead, the core innovation of this application is first verified—namely, synchronously acquiring and uniformly calculating multi-domain data within the same dwell window under a unified clock reference. Specifically, a preliminary verification link is constructed using a 1m collimator, a small scanning mirror, a common laser, and a high-speed camera. Taking the single-cycle scanning process of the small scanning mirror as an example, it is still divided into acceleration, staring, deceleration, and reset segments. The external synchronization controller only triggers the high-speed measurement camera to sample during the staring segment. Through this method, the light spot trajectory caused by the speed fluctuation of the scanning mirror can be obtained first, and pixel-angle calibration can be completed. At this stage, the system components can be gradually built using existing experimental conditions. It is not required to have all components in a complete engineering state at once. Instead, the focus is on verifying whether the links for unified clock triggering, dwell window interception, light spot trajectory acquisition, and data calculation are connected and effective.

[0085] In the prototype stage, the feedback results from the scanning mirror encoder can be compared with the inversion results of the high-speed camera's spot trajectory. For example, under the same driving conditions, the speed fluctuation obtained from the encoder feedback is about 5%, while the speed fluctuation obtained from the high-speed camera inversion is about 5% to 6%. The trends and magnitudes of the two changes are basically consistent, indicating that the high-speed measurement camera link can realistically reflect the actual dynamic process of the scanning mirror. Based on this, dwell window gating logic can be added to the same link to limit the sampling time of the high-speed measurement camera, so that it only collects data within a predetermined staring segment, thereby verifying the synchronous measurement mechanism described in this application that "targets the dwell exposure window rather than the entire scanning process". Subsequently, a fast reflector compensation stage can be gradually added to compare the changes in spot trajectory, angle time series, and corresponding statistical indicators before and after compensation, realizing the transition from "speed fluctuation observation" to "angle time series calculation within the dwell window".

[0086] Furthermore, a step-by-step verification approach of "angle domain first, then imaging domain" can be adopted at the prototype stage. First, a laser point source and a high-speed camera are used to complete the acquisition of the light spot trajectory, centroid extraction, and pixel-angle mapping to confirm that the angle domain link is working properly. Then, depending on the completeness of the device, simplified or standard targets are gradually introduced to verify whether a correspondence can be established between the imaging domain link and the angle domain link. In this way, even before a complete engineering prototype is formed, the basic logic and data closed-loop relationship of the multi-domain joint verification system described in this application can be verified.

[0087] The significance of this embodiment lies in demonstrating that even in the prototype stage, the verification link of this application can complete the synchronous capture of the dwell window, spot trajectory measurement, pixel-angle calibration, time series calculation, and multi-cycle consistency analysis under limited experimental conditions. This upgrades the observation of simple motion phenomena to the synchronous and quantitative verification of the accuracy of the visual axis within the dwell exposure window, and lays the foundation for subsequent expansion to a complete two-mirror joint system and cross-verification of the imaging domain.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A system for verifying visual axis dwell accuracy in step-gazing imaging, characterized in that, The step-by-step staring imaging device includes a main scanning mirror and a fast-reflecting mirror, and the verification system includes: The optical simulation and measurement module is used to construct the external standard optical verification environment of the tested step-gazing imaging device under laboratory conditions. It includes a large-aperture collimator, a precision optical platform, an autocollimator, a laser point source component, and a target component located on the focal plane of the collimator. The device under test fixture and environmental monitoring module are used for stress-free clamping of the stepping staring imaging device under test and for monitoring ground micro-vibration interference. The vision acquisition module is used to capture high dynamic optical signals, including a high-speed measuring camera located at the focal plane of the collimator and an infrared camera located at the focal plane of the stepping staring imaging device itself. The synchronous control and data processing hub is electrically connected to the controllers of the laser point source component, the high-speed measuring camera, and the measured stepping staring imaging device, respectively. It is used to uniformly issue clock beats, execute centroid extraction algorithms, and output dwell accuracy verification reports. It includes a programmable timing synchronizer and a data processing host computer.

2. The visual axis dwell accuracy verification system for step-gazing imaging according to claim 1, characterized in that, The optical verification environment constructed by the optical simulation and measurement module includes at least the following: forming a collimated beam in front of the entrance pupil of the step-gazing imaging device under test using a large-aperture collimator to simulate the incident conditions of distant or infinitely distant targets; providing resolution targets, edge targets, or point targets through a target assembly located on the focal plane of the collimator to achieve image quality testing of the imaging link and verification of imaging quality within the dwell window; constructing a reverse or equivalent optical measurement link for measuring the visual axis deflection using a laser point source assembly in conjunction with a high-speed measurement camera to obtain the change in the position of the light spot during dwell; and performing angular reference measurement on the step-gazing imaging device under test or its reference reflective surface using an autocollimator to provide an external angular reference for the mapping relationship between pixel displacement and visual axis angle.

3. A method for verifying visual axis dwell accuracy in step-gazing imaging, implemented based on the visual axis dwell accuracy verification system for step-gazing imaging as described in claim 1, characterized in that, Includes the following steps: Step 1: Construct a multi-source external optical verification link: Set a collimator at the entrance pupil of the stepping staring imaging device under test, and switchably set a high-speed measuring camera or resolution target connected to an external synchronous controller at the focal plane of the collimator; set a laser point source at the focal plane of the stepping staring imaging device under test. Step 2: Establish the calibration mapping from pixel displacement to visual axis angle: Control the measured step staring imaging device to be static, drive the fast reflector and scanning mirror to input the known curve excitation, use the high-speed measuring camera to collect the displacement of the spot centroid, and establish the static calibration coefficient between the pixel displacement of the spot centroid and the actual visual axis deflection angle. Step 3, Dwelling Window Timing Synchronization: The external synchronization controller acquires the scan retrace synchronization signal of the stepping staring imaging device under test, and sends a trigger pulse with time delay compensation to the high-speed measuring camera, so that the high-speed measuring camera performs ultra-high frame rate continuous image acquisition only within the "dwelling exposure window" of the stepping staring imaging cycle. Step 4: High-frequency transient angle time series calculation: Extract the two-dimensional centroid coordinates of the laser spot in each frame of the image within the dwell exposure window, and combine them with the static calibration coefficients to calculate the two-dimensional time series of the line-of-sight angle deviation within the dwell window. Step 5: Multidimensional evaluation and cross-validation of dwell accuracy: Calculate the root mean square value and peak-to-peak value of the two-dimensional time series of angle deviation within the dwell window time to determine whether it meets the sub-microradian level dwell accuracy threshold; and switch the resolution target, perform MTF calculation based on the target image extracted by the infrared camera to achieve cross-validation of time domain trajectory and frequency domain imaging quality.

4. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 3, characterized in that, The specific details of step 2 are as follows: Step 2.1: When the stepping staring imaging device under test is in a static or quasi-static condition, control the fast reflector and the main scanning mirror to operate individually or in a predetermined combination, and input multiple sets of known angle excitation signals, wherein the excitation signals are step signals, ramp signals, sine signals or piecewise linear scanning signals. Step 2.2: Under each set of known angle excitations, obtain the corresponding actual line-of-sight deflection angle reference value using an autocollimator or reference measuring device, and at the same time, acquire the spot image sequence at the corresponding moment using a high-speed measuring camera. Step 2.3: Preprocess the acquired spot images. The preprocessing includes background subtraction, dark field correction, threshold segmentation, region of interest extraction, and abnormal frame removal. Step 2.4: Perform centroid extraction and subpixel fitting on the preprocessed spot images to obtain the center coordinates of the spot in each frame; Step 2.5: Using the center coordinates of the spot in the zero-position reference frame as a reference, calculate the pixel displacement of the spot under each excitation condition. Step 2.6: Fit the pixel displacement of the spot with the corresponding actual visual axis deflection angle reference value to establish a calibration mapping model of pixel displacement to visual axis angle; the calibration mapping model adopts a one-dimensional scaling factor model, a two-dimensional linear decoupling model, a two-dimensional coupling model with cross terms, or a low-order polynomial model. Step 2.7: Perform residual verification on the fitted calibration mapping model, and use the residuals that meet the preset threshold as the calibration result used for subsequent dwell window angle calculation.

5. The method for verifying visual axis dwell accuracy for step-gazing imaging according to claim 4, characterized in that, The calibration mapping model for pixel displacement to viewing axis angle is expressed as follows: the displacement of the point spot in the two-dimensional image coordinate system is mapped to the deflection angle of the viewing axis in two orthogonal directions, while simultaneously compensating for pixel size error, imaging magnification error and coordinate axis non-orthogonality error.

6. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 3, characterized in that, In step 3, the synchronization control and data processing center obtains from the tested step-gazing imaging device a synchronization reference signal characterizing the scanning cycle, retrace start point, dwell start point and exposure range, as well as the transmission delay and trigger response delay between components. Based on the timing parameters of the stepping staring imaging device under test within a single cycle, a gated acquisition command corresponding to the dwell exposure window is generated. The gated acquisition command includes at least the acquisition start time, acquisition duration, acquisition frame rate, and trigger advance. The synchronous control and data processing center outputs a trigger pulse to the high-speed measurement camera, so that the high-speed measurement camera starts continuous acquisition only within the corresponding dwell exposure window.

7. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 3, characterized in that, The specific details of step 4 are as follows: Step 4.1: Perform background suppression, threshold segmentation, and region of interest cropping on the spot images of each frame continuously acquired within the dwell exposure window; Step 4.2: Perform centroid extraction on the spot points in each frame image based on the sub-pixel localization algorithm to obtain the two-dimensional center coordinates of the spot points in each frame; Step 4.3: Using a pre-set zero-position reference frame as a reference, calculate the two-dimensional pixel displacement of the center of the spot in each frame relative to the reference position. Step 4.4: Substitute the two-dimensional pixel displacement of each frame into the pixel displacement to view axis angle calibration mapping model established in Step 2, and calculate the two-dimensional view axis angle deviation of the corresponding frame. Step 4.5: Sort the two-dimensional view angle deviation of each frame according to a unified timestamp to form a two-dimensional view angle time series within the dwell exposure window; Step 4.6: When the sampling clock of the high-speed measuring camera is inconsistent with the control clock of the measured step-gazing imaging device, the two-dimensional viewing axis angle time series is interpolated, resampled, time-base corrected, or filtered to obtain a high-frequency transient angle time series that strictly corresponds to the dwell window.

8. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 7, characterized in that, Centroid extraction employs a gray-scale weighted centroid algorithm, which involves: weighting the coordinates based on the gray values ​​of each pixel within the spot area to obtain the center position of the spot; and then fitting the initial centroid value using a two-dimensional Gaussian function to obtain the sub-pixel level center coordinates of the spot.

9. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 3, characterized in that, In step 5, cross-validation includes: using the MTF result obtained under static conditions as a benchmark value, comparing the MTF result obtained under dynamic dwell conditions with the benchmark value, and combining the root mean square value or peak-to-peak value change of the two-dimensional line-of-sight angle time series within the dwell window to establish a "angle perturbation-imaging degradation" correspondence, thereby verifying whether the dwell accuracy result obtained by the spot trajectory calculation is consistent with the actual imaging quality change.

10. The method for verifying visual axis dwell accuracy for step-gaze imaging according to claim 9, characterized in that, The specific details of step 5 are as follows: Step 5.1: Calculate the mean, root mean square value, peak-to-peak value, maximum absolute deviation value, and standard deviation for the two-dimensional time series of line-of-sight angle deviation within the dwell window obtained in Step 4. Step 5.2: Compare the time-domain evaluation index with the preset sub-microradian level dwell accuracy threshold to determine whether the tested step staring imaging device meets the visual axis dwell accuracy requirements under the current working conditions. Step 5.3: While keeping the scanning conditions and synchronization timing of the stepping staring imaging device under test unchanged, switch the high-speed measuring camera at the focal plane of the collimator to a resolution target, a beveled target, or a stripe target, and use the infrared camera at the focal plane of the stepping staring imaging device under test to acquire the corresponding target image. Step 5.4: Evaluate the imaging quality of the target image acquired by the infrared camera; Step 5.5: Perform correlation analysis between the angular stability evaluation results within the dwell window and the MTF results under the corresponding working conditions to determine the degree of influence of the change in the dwell accuracy of the line of sight on the image sharpness. Step 5.6: When both the angle domain evaluation result and the imaging domain evaluation result meet the preset criteria, output the conclusion that the tested step-gazing imaging device has passed the sub-micro-radian level visual axis dwell accuracy verification.

Images