Light measurement real scene simulation training method and system

Abstract

The invention provides an optical measurement live-action simulation training method and system, and the method comprises the steps: loading a historically recorded target image, reading absolute time, an azimuth angle, a pitch angle and a recorded frame frequency, and carrying out the image enhancement processing of the target image, and obtaining a training image; establishing a tracking window, wherein the tracking window is used for tracking a target center in the training image; constructing a simulation servo control model, continuously playing the training image in the tracking window according to a recorded frame frequency, and performing tracking target training by controlling the simulation servo control model; and after training is completed, target tracking scoring is carried out according to the target tracking condition, training scoring results are obtained, and the training scoring results comprise a guide item score, a data validity score, a tracking stability score and a comprehensive score. The method can achieve the precise simulation of the real motion track of the aerial target, is suitable for various target scenes, and can achieve the unified evaluation of the training result.

Description

Technical Field

[0001] The present invention relates to the field of optical measurement technology, and in particular to a light measurement real-scene simulation training method and system. Background Art

[0002] As the primary optical measurement device, the photoelectric theodolite plays an important role in capturing real-time images of aerial targets and measuring exterior ballistic parameters. In actual use, due to various factors such as weather complexity, the randomness of aerial target trajectories, and the low success rate of automatic target capture and tracking, the current stage mainly relies on single-rod operators to operate the single rod for stable tracking. Therefore, the operator's technical level directly affects the quality of target tracking. The daily training and development of single-rod operators mainly uses simulation exercises based on simulation training systems and practical training combined with tracking real targets. The following problems are commonly encountered in actual training: First, there's a significant gap between the operating experience of existing simulation training systems and actual tracking scenarios. Common simulation training systems' target trajectory, target display, and single-lever operation are all incapable of simulated tracking. The target trajectory is relatively simple, the display effect resembles animation, and the image sensor types are limited. Operators generally report significant differences from real-world tracking scenarios.

[0003] Second, the current target tracking scenarios for photoelectric theodolites are complex and varied. Existing simulation training systems fail to account for diverse factors, such as complex background environments and complex motion trajectories. Single-rod operators lack the platform and opportunity to improve their tracking capabilities.

[0004] Third, there is a lack of standardized evaluation and assessment methods. Due to factors such as differences in sensor types and actual operation between various types of photoelectric theodolites, as well as the significant differences in the difficulty of tracking measurements at different sites, single-rod operators are currently unable to conduct comparative evaluations of their tracking performance under the same conditions.

[0005] Therefore, there is an urgent need for a light measurement real-scene simulation training method that can accurately simulate the real motion trajectory of aerial targets, be applicable to various target scenarios, and evaluate training results. Summary of the Invention

[0006] Based on this, it is necessary to provide a light measurement real-scene simulation training method and system to address the above technical problems.

[0007] A method for photometry real-scene simulation training comprises the following steps: loading a target image recorded in a historical record, reading absolute time, azimuth, pitch angle, and recorded frame rate, and performing image enhancement processing on the target image to obtain a training image; establishing a tracking window, wherein the tracking window is used to track the center of a target in the training image; constructing a simulation servo control model, continuously playing the training image in the tracking window at the recorded frame rate, and performing target tracking training by controlling the simulation servo control model; after the training is completed, performing target tracking scoring according to the target tracking situation to obtain a training scoring result, wherein the training scoring result includes a guidance item score, a data validity score, a tracking stability score, and a comprehensive score.

[0008] In one embodiment, the recording frame rate is calculated by the absolute time difference between two consecutive frames of images.

[0009] In one embodiment, establishing a tracking window includes: establishing a tracking window so that the tracking window has the same resolution as the training image, and the window center of the tracking window coincides with the geometric center of the first frame training image, and the geometric center of the training image is the center of the tracking target; loading the training image, and adding a crosshair in the window center of the tracking window to display the reference values of the azimuth and pitch angles of the encoder corresponding to the window center, and establishing a simulated photoelectric theodolite tracking angle; and judging the offset of the target in the center of the field of view by identifying the position of the tracking target relative to the crosshairs.

[0010] In one embodiment, the construction of the simulation servo control model, continuously playing the training image in the tracking window at a recorded frame rate, and performing target tracking training by controlling the simulation servo control model include: using a serial port card, a data acquisition card, and a simulated single rod to construct the simulation servo control model, the simulated single rod having the same characteristics as the device single rod and capable of generating the same angular velocity output value as the device single rod; continuously playing the training image in the tracking window at the recorded frame rate to form a training video; when the simulation servo control model is used to perform target tracking training on the training video, the simulation servo control model controls the simulated single rod to emit an output signal, the simulation servo control model converts the output signal into a digital signal, and sends the signal to the serial port card via a serial port; the serial port card receives the digital signal, and the data acquisition card acquires the digital signal on the serial port card to generate azimuth and pitch angle data of the simulation encoder; obtaining an angle change based on the angular velocity output value generated by the simulated single rod, calculating a pixel distance of each frame of the image displacement based on the angle change, and performing overall image displacement based on the pixel distance; and performing edge padding on the displaced training image frame by frame in the tracking window.

[0011] In one embodiment, obtaining the angle change according to the angular velocity output value generated by the simulated single-rod includes: calculating the direction angle change generated by the training image according to the angular velocity output value of the simulated single-rod when operating the simulated single-rod, using the formula: , Where, is the initial driving value of the simulated single pole in the horizontal direction, To simulate the horizontal angular velocity output value of a single rod, is the driving coefficient of the simulated single rod, which is a constant. To manipulate the azimuth angle change produced by the simulated single rod, When the operation range is small and does not reach the single-pole initial drive value, the azimuth angle does not change. To simulate the maximum output value that a single rod can produce in the horizontal direction, when When , the resulting azimuth angle change is only related to the maximum output value; similarly, the pitch angle change is obtained as: , Where, is the initial drive value of the simulated single rod in the vertical direction, To simulate the vertical angular velocity output value of a single rod, To manipulate the pitch angle change produced by the simulated single stick, When the control range is small and does not reach the single-stick initial drive value, the pitch angle will not change. To simulate the maximum output value that a single rod can produce in the vertical direction, when , the resulting pitch angle change is only related to the maximum output value.

[0012] In one embodiment, the pixel distance of each frame image displacement is calculated based on the angle change, and the overall image displacement is performed based on the pixel distance, further comprising: calculating the pixel distance of each frame image displacement based on the azimuth angle change and the pitch angle change; performing overall image displacement on continuously played training images based on the pixel distance to obtain the true motion trajectory of the target; wherein, starting from the second frame image, the overall image displacement processing is performed based on the encoder value of the previous frame image, so that the image displacement direction is consistent with the actual tracking target motion direction, and the formula for the pixel value of the training image horizontally shifted in the x-axis direction and the azimuth angle change is: , Where, is the azimuth value of the current image, is the azimuth value of the previous frame image, To manipulate the azimuth angle change produced by the simulated single pole, is the final azimuth change of the image, is the field of view angle in the x-axis direction, is the resolution of the current training image, is the pixel value of the image translated horizontally along the x-axis; similarly, the formula for the pixel value of the training image translated vertically along the y-axis and the change in pitch angle is: , Where, is the pitch angle value of the current image, is the pitch angle value of the previous frame image, To simulate the pitch angle change caused by a single-pole operation, is the final pitch angle change of the image, is the field of view angle in the y-axis direction, which is a fixed value. is the resolution of the current image, The pixel value by which the image is translated vertically along the y-axis.

[0013] In one embodiment, the edge filling of the shifted training image frame by frame in the tracking window includes: detecting the positional relationship between the training image after the overall displacement of the image and the tracking window, removing all pixels that are moved out of the tracking window to obtain a blank area; selecting pixel points of fixed width and height in the training image connecting the blank area as reference pixels; and copying the reference pixels in the blank area of the tracking window based on the horizontal and vertical directions to perform edge filling.

[0014] In one embodiment, the method further includes: during the training process, if the tracking target disappears in the field of view or deviates from the center of the field of view by more than a threshold, data guidance is performed, that is, the geometric center of the training image is moved to the center of the window according to the maximum angular velocity of the simulated single rod, and after the geometric center of the training image completely coincides with the center of the tracking window, manual tracking of the target is resumed.

[0015] In one embodiment, the guided project score is scored according to the proportion of image frames using data guidance in all training images, and the formula is: , Where, is the number of image frames guided by data, is the total number of training image frames, The final score of the guidance project; when the proportion of image frames using data guidance exceeds 0.1, it is considered that the guidance time is too long and is directly judged as 0 points; the effective score of the data is scored according to the situation of the target deviating from the field of view, and the formula is: , Where, The number of image frames with invalid data. is the total number of training image frames, The final score for data validity. When the percentage of image frames that deviate from the field of view exceeds 0.1, the tracking condition is considered poor and is scored as 0. When the azimuth angle change or the pitch angle change exceeds half of the field of view, the image frame is considered to have deviated from the field of view and the data is considered invalid. The tracking stability score is scored according to the offset of the tracking target from the image center. The formula is: , Where, is the pixel value of the image horizontally shifted in the x-axis direction, is the average value of the pixel values of each frame image horizontally translated in the x-axis direction, is the pixel value of the image translated vertically along the y-axis. is the average value of the pixel values of each frame image translated vertically in the y-axis direction, is the root mean square of the azimuth angle, reflecting the stability of the operation in the horizontal direction. is the root mean square of the pitch angle, reflecting the stability of the operation in the vertical direction. It can reflect the stability of the overall tracking of the operation. The final tracking stability score is obtained by calculating the comprehensive score based on the guided item score, data validity score and tracking stability score. The formula is: , Where, 、 and are the weights of the guided project score, data validity score and tracking stability score respectively.

[0016] A photometric real-scene simulation training system is used to implement the photometric real-scene simulation training method described above, comprising: a training image acquisition module for loading a target image recorded in a historical record, reading the absolute time, azimuth, pitch angle, and recorded frame rate, and performing image enhancement processing on the target image to obtain a training image; a tracking window establishment module for establishing a tracking window, wherein the tracking window is used to track the target center in the training image; a simulation model construction module for constructing a simulation servo control model, continuously playing the training image in the tracking window at the recorded frame rate, and performing target tracking training by controlling the simulation servo control model; and a target tracking scoring module for performing target tracking scoring according to the target tracking situation after training is completed to obtain a training scoring result, wherein the training scoring result includes a guidance item score, a data validity score, a tracking stability score, and a comprehensive score.

[0017] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: by loading the target image of the historical record, reading the absolute time, azimuth, pitch angle and recording frame, and performing image enhancement processing on the target image, a training image is obtained; a tracking window is established to track the target center in the training image to facilitate target tracking and determine the offset of the target in the field of view; a simulation servo control model is constructed, the training image is continuously played in the tracking window according to the recorded frame rate, and the target tracking training is performed by controlling the simulation servo control model to achieve simulation tracking training of the target; after the training is completed, the target tracking is scored according to the target tracking situation to obtain a training scoring result, which includes a guidance item score, a data validity score, a tracking stability score and a comprehensive score, so that the training results can be uniformly evaluated to facilitate subsequent operator skill improvement, and the real motion trajectory of the target is accurately simulated. The system is applicable to a variety of target scenarios and provides a platform and a real operating experience for daily training, assessment and evaluation of single-pole operators of optical measurement equipment. The system has strong simulation, wide applicability, high tracking efficiency, and effectively solves the training problem of single-pole operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 1 is a flow chart of a light measurement real scene simulation training method according to an embodiment; Figure 2 is a schematic diagram of edge filling in one embodiment; Figure 3 Schematic diagram of the structure of a light measurement real scene simulation training system in one embodiment; Figure 4 FIG. 1 is a schematic diagram of the architecture of a light measurement real-scene simulation training system in one embodiment. DETAILED DESCRIPTION

[0019] Before describing the specific embodiments of the present invention, the overall concept of the present invention is described as follows: The present invention is mainly developed based on the target tracking training process of a photoelectric theodolite. The existing simulated photoelectric theodolite single-pole target tracking training system cannot realistically simulate the actual motion trajectory of aerial targets. The image sensor types are few and the imaging realism is low. The training effect after the operator uses it is poor, and the operation level and psychological quality are difficult to be effectively improved.

[0020] Therefore, the present invention proposes a real-scene simulation training method for optical measurement, which loads a target image of a historical record, reads the absolute time, azimuth, pitch angle and recording frame, and performs image enhancement processing on the target image to obtain a training image; establishes a tracking window to track the center of the target in the training image to facilitate target tracking and determine the offset of the target in the field of view; constructs a simulation servo control model, continuously plays the training image in the tracking window according to the recorded frame rate, and controls the simulation servo control model to perform target tracking training to achieve simulation tracking training of the target; after the training is completed, the target tracking is scored according to the target tracking situation to obtain a training scoring result, which includes a guidance item score, a data validity score, a tracking stability score and a comprehensive score, so that the training results can be uniformly evaluated to facilitate subsequent operator skill improvement, achieves accurate simulation of the real motion trajectory of the target, is applicable to a variety of target scenarios, and provides a platform and real operating experience for daily training, assessment and evaluation of single-pole operators of optical measurement equipment. It has strong simulation, wide applicability, high tracking efficiency, and effectively solves the training problem of single-pole operators.

[0021] After introducing the overall concept of the present invention, in order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0022] In one embodiment, Figure 1 As shown, a light measurement real scene simulation training method is provided, comprising the following steps: Step S110 , loading the target image of the historical record, reading the absolute time, azimuth angle, pitch angle and recording frame rate, and performing image enhancement processing on the target image.

[0023] Specifically, because different types of photoelectric theodolites exist at different sites, the field of view, tracking difficulties, and specific operations for the same batch of mission targets vary depending on factors such as equipment layout, observation angle, and equipment parameters. Therefore, it is possible to fully utilize the vast amount of historical optical measurement images of various types and varying levels of tracking difficulty as material for simulated tracking training.

[0024] Load historical records of aerial target images, distinguish different sensors, such as medium-wave infrared, long-wave infrared, visible light, and images of different formats, display and read the absolute time, encoder azimuth angle, encoder pitch angle, recording frame rate and other information in each frame as important data basis for simulation training, The recorded frame rate is calculated by the absolute time difference between two consecutive frames of images.

[0025] Because some target images may have insufficient pixel brightness, image enhancement is required. Grayscale correction is applied to the loaded target images to enhance image contrast and generate training images. The processed training images, which distinguish sensor parameters and trajectory characteristics, are classified and saved for easy recall during subsequent simulation training.

[0026] Step S120: establishing a tracking window, where the tracking window is used to track the center of the target in the training image.

[0027] Specifically, a tracking window of the same resolution is established based on the obtained training image to track the center of the target in the training image, so as to realize tracking measurement of the target azimuth and pitch angle, thereby determining the position offset of the target.

[0028] Among them, step S120 includes: establishing a tracking window so that the tracking window has the same resolution as the training image, and the window center of the tracking window coincides with the geometric center of the first frame of the training image, and the geometric center of the training image is the center of the tracking target; loading the training image, and adding a crosshair in the window center of the tracking window to display the reference values of the azimuth and pitch angles of the encoder corresponding to the window center, and establishing a simulated photoelectric theodolite tracking angle; by identifying the position of the tracking target relative to the crosshairs, determining the offset of the target in the center of the field of view.

[0029] Specifically, a tracking window is established. The tracking window has the same resolution as the current training image, and the center of the window completely coincides with the geometric center of the first frame of the training image. Since optical recording images with high tracking quality and relatively stable tracking targets are usually used as simulation training materials, the tracking targets in the original image are all located around the geometric center of the image. Therefore, the geometric center of the training image is used as the center of the tracking target.

[0030] Load the training image and add a crosshair to the geometric center of the tracking window. The crosshair is a core auxiliary tool for photoelectric theodolites to achieve high-precision measurement. It is used to display the standard values of the encoder azimuth and pitch values corresponding to the current geometric center of the tracking window. By identifying the tracking target's position relative to the crosshairs, the target's offset in the field of view can be determined. During the simulation training process, the crosshair center is always kept at the center of the tracking window, and this center is used as the reference value for the current encoder output azimuth and pitch angles to establish a simulated photoelectric theodolite tracking angle with authentic and reliable data.

[0031] Step S130 , constructing a simulation servo control model, continuously playing the training image in the tracking window according to the recorded frame rate, and performing target tracking training by controlling the simulation servo control model.

[0032] Specifically, a corresponding simulation servo control model is constructed based on the actual photoelectric theodolite, and the training image is continuously played in the tracking window at the recorded frame rate to form a video identical to the actual tracking process. By controlling the simulation servo control model to perform target tracking training, simulation training of the actual photoelectric theodolite is achieved.

[0033] Step S130 includes: constructing a simulation servo control model using a serial port card, a data acquisition card, and a simulation unit, wherein the simulation rod has the same characteristics as the device rod and can generate the same angle output value as the device rod; continuously playing the training image in the tracking window at a recorded frame rate to form a training video; when the training video is trained to track a target using the simulation servo control model, controlling the simulation rod to emit an output signal, converting the output signal into a digital signal, and sending the signal to the serial port card via a serial port; the serial port card receives the digital signal, and using the data acquisition card to acquire the digital signal on the serial port card to generate azimuth and pitch angle data of the simulation encoder; obtaining an angle change based on the angular velocity output value generated by the simulation rod, calculating a pixel distance of displacement of each frame of the image based on the angle change, and performing overall image translation based on the pixel distance; and performing edge padding on the translated training image frame by frame in the tracking window.

[0034] Specifically, the servo control system for a photoelectric theodolite is primarily composed of a servo controller, a torque motor, a power drive module, and an under-the-hood operating rod. It is primarily used to control the dual axes of the tracking frame to achieve stable target tracking. Therefore, when constructing a simulation servo control model for a photoelectric theodolite, a serial port card, a data acquisition card, and a simulated rod can be used. The simulated rod shares the same characteristics as the operating rod and produces the same angular velocity output values. The simulated rod's control direction can be set to reverse, meaning that upward control of the simulated rod causes the target in the training image to move downward. This control direction is altered by controlling the positive and negative relationship of the simulated rod's angular velocity output values.

[0035] Among them, the hardware requirements for simulating a single pole are: a mechanical rotation angle of 360° with no stop position; an azimuth working angular velocity of 0.01-35; a pitch working angular velocity of 0.01-20; a maximum angular velocity of ≥ 60; an azimuth working angular acceleration of 0-25; a pitch working angular acceleration of 0-12; a maximum angular acceleration of ≥ 60; and the hardware requirements for the data acquisition card are: sampling accuracy: 16 bits; and a sampling rate of 1MHz.

[0036] The simulation servo control model converts the output signal of the simulated single rod into a digital signal in real time. After processing the digital signal, it is sent to the serial port card through the serial port. The serial port card receives the data signal of the simulated single rod and collects the angular velocity data output by the simulated single rod through the data acquisition card. The data is finally used to generate the azimuth and pitch angle data of the simulation encoder.

[0037] The obtained training image is played continuously in the tracking window at the recorded frame rate to form a video identical to the target tracking video of the actual photoelectric theodolite. The operator can perform target tracking training by manipulating the simulated single rod to control the size of the angular velocity output value.

[0038] When using the simulation servo control model to conduct target tracking training in the training video, the operator sends an output signal by controlling the simulated single rod. The simulation servo control model converts the output signal into a digital signal and sends it to the serial port card through the serial port. The digital signal is received by the serial port card and the digital signal on the serial port card is collected by the data acquisition card to finally generate the azimuth and pitch angle data of the simulation encoder.

[0039] When an electro-optical theodolite tracks an aerial target, the target will move normally within the image display field of view. Since the target position cannot be moved within the loaded image, the resulting image must be translated to represent the target's true trajectory. After training, the angular velocity output of the simulated single-rod output is used to determine the angular change in the training image, including changes in azimuth and pitch. This angular change is used to calculate the pixel distance of each frame's displacement. This pixel distance is then used to translate the entire image to ensure that the resulting target trajectory matches the true target's trajectory.

[0040] After image translation, some pixels of the image will move out of the tracking window, and blank areas will appear in the tracking window. Therefore, it is necessary to fill the edges of the translated training image frame by frame to ensure that the tracking background is uniform.

[0041] The step of obtaining the angle variation according to the angular velocity output value generated by the simulated single pole specifically includes: when manipulating the simulated single pole, calculating the direction angle variation generated by the training image according to the angular velocity output value of the simulated single pole, using the formula: , Where, is the initial driving value of the simulated single pole in the horizontal direction, To simulate the horizontal angular velocity output value of a single rod, is the driving coefficient of the simulated single rod, which is a constant. To manipulate the azimuth angle change produced by the simulated single rod, When the operation range is small and does not reach the single-pole initial drive value, the azimuth angle does not change. To simulate the maximum output value that a single rod can produce in the horizontal direction, when When , the resulting azimuth angle change is only related to the maximum output value; similarly, the pitch angle change is obtained as:

[0042] Where, is the initial drive value of the simulated single rod in the vertical direction, To simulate the vertical angular velocity output value of a single rod, To manipulate the pitch angle change produced by the simulated single stick, When the control range is small and does not reach the single-stick initial drive value, the pitch angle will not change. To simulate the maximum output value that a single rod can produce in the vertical direction, when , the resulting pitch angle change is only related to the maximum output value.

[0043] Specifically, the operator controls the size of the angular velocity output value by manipulating the simulated single rod, and obtains the angle change based on the angular velocity output value generated by the simulated single rod. The angle change includes the azimuth angle change and the pitch angle change, so as to calculate the pixel distance of the displacement of each frame of the image based on the angle change.

[0044] The pixel distance of each frame image displacement is calculated based on the angle change, and the steps of performing overall image displacement based on the pixel distance specifically include: calculating the pixel distance of each frame image displacement based on the azimuth angle change and the pitch angle change; performing overall image displacement on the continuously played training images based on the pixel distance to obtain the true motion trajectory of the target; starting from the second frame image, overall displacement processing is performed based on the encoder value of the previous frame image to make the image displacement direction consistent with the actual tracking target motion direction. The formula for the pixel value of the training image horizontally shifted in the x-axis direction and the azimuth angle change is: , Where, is the azimuth value of the current image, is the azimuth value of the previous frame image, To manipulate the azimuth angle change produced by the simulated single pole, is the final azimuth change of the image, The field of view angle in the x-axis direction is related to the device type and sensor type and is a fixed value. is the resolution of the current training image, is the pixel value of the image translated horizontally along the x-axis; similarly, the formula for the pixel value of the training image translated vertically along the y-axis and the change in pitch angle is: , Where, is the pitch angle value of the current image, is the pitch angle value of the previous frame image, To simulate the pitch angle change caused by a single-pole operation, is the final pitch angle change of the image, is the field of view angle in the y-axis direction, which is a fixed value. is the resolution of the current image, The pixel value by which the image is translated vertically along the y-axis.

[0045] Specifically, when performing overall image displacement, the azimuth angle change, pitch angle change and the previous frame image can be combined to perform overall image displacement to ensure that the image displacement direction is consistent with the actual tracking target movement direction, so as to obtain the target's true displacement trajectory and improve the authenticity and reliability of simulation training.

[0046] Among them, the step of filling the edges of the translated training image frame by frame in the tracking window specifically includes: detecting the positional relationship between the training image after the overall displacement of the image and the tracking window, removing all pixels that move out of the tracking window to obtain a blank area; selecting pixel points of fixed width and height in the training image connecting the blank area as reference pixels; based on the horizontal and vertical directions, copying the reference pixels in the blank area of the tracking window for edge filling.

[0047] Specifically, after the image is translated by a number of pixels in the horizontal direction of the x-axis and the vertical direction of the y-axis, some pixels will move out of the tracking window, and some blank areas will appear in the tracking window. To ensure that each frame of the image can be displayed normally in the tracking window, it is necessary to remove all pixels that move out of the tracking window, and select pixels of fixed width and height in the image connecting the blank areas as reference pixels, distinguishing between horizontal and vertical directions, and copy all the reference pixels in the blank areas of the tracking window to complete the filling of the blank areas of the tracking window image, so that the spliced image can be fully displayed in the tracking window. The operation method is as follows Figure 2 As shown in the figure, the edge filling of each frame of the displayed image is completed according to the above method. After multiple tests, the stitched images can achieve the effect of tracking a uniform background, improving the authenticity of the simulation training.

[0048] This also includes: during the training process, if the tracking target disappears in the field of view or deviates from the center of the field of view by more than a threshold, data guidance is performed, that is, the geometric center of the training image is moved to the center of the window according to the maximum angular velocity of the simulated single pole. After the geometric center of the training image completely coincides with the center of the tracking window, manual tracking of the target is restored.

[0049] Specifically, the switching data guidance function is one of the important functions of the photoelectric theodolite. When tracking a target, if the target disappears from the field of view or deviates too much from the center of the field of view (that is, exceeds the preset threshold range), the operator can use the switching data guidance function. The servo system will use the values of the current target's theoretical azimuth and pitch angles in the external guidance information to track in real time, reintroduce the tracking target into the field of view, and then switch to manual tracking to ensure that the tracking target is not lost for a long time. The system has a switching virtual data guidance function. That is, when the tracking target deviates significantly from the image center or is completely lost, the operator uses this function. The system automatically moves the current image geometric center to the center of the tracking window, that is, the crosshair position, according to the simulated single-rod maximum angular velocity output value. When the two are completely overlapped, it indicates that the target has been reintroduced into the field of view, and then returns to manual tracking of the target.

[0050] Step S140: After the training is completed, the training tracking score is performed according to the target tracking situation to obtain a training score result.

[0051] Specifically, after training is completed using the above method, a target tracking score can be generated based on the target tracking performance during training. This training score can be used to effectively evaluate the operator's tracking quality and help operators improve their target tracking skills. The training score includes a guidance item score, a data validity score, a tracking stability score, and a comprehensive score.

[0052] Step S140 includes: scoring the guided project score according to the proportion of the number of image frames guided by the data in all training images, and the formula is: , Where, is the number of image frames guided by data, is the total number of training image frames, The final score of the guidance project; when the proportion of image frames using data guidance exceeds 0.1, it is considered that the guidance time is too long and is directly judged as 0 points; the data validity score is scored according to the situation of the target deviating from the field of view, and the formula is: , Where, The number of image frames with invalid data. is the total number of training image frames, The final score for data validity. When the percentage of image frames that deviate from the field of view exceeds 0.1, the tracking condition is considered poor and is scored 0 points. When the azimuth angle change or the pitch angle change exceeds half of the field of view, the image frame is considered to have deviated from the field of view and the data is considered invalid. The tracking stability score is scored based on the offset of the tracking target from the image center. The formula is: , Where, is the pixel value of the image horizontally shifted in the x-axis direction, is the average value of the pixel values of each frame image horizontally translated in the x-axis direction, is the pixel value of the image translated vertically along the y-axis. is the average value of the pixel values of each frame image translated vertically in the y-axis direction, is the root mean square of the azimuth angle, reflecting the stability of the operator in the horizontal direction. is the root mean square of the pitch angle, reflecting the stability of the operator in the vertical direction. It can reflect the stability of the operator's overall tracking. The final score for tracking stability; the comprehensive score is calculated based on the guided project score, data validity score and tracking stability score, and the formula is:

[0053] Where, 、 and are the weights of the guided project score, data validity score and tracking stability score respectively.

[0054] Specifically, the guidance project score is scored based on the proportion of image frames using guidance in all training images, which can truly reflect the degree to which the operator relies on the data guidance function. The comprehensive score is obtained by multiplying the above three values by the weights and then summing them. When setting the weights, considering that most single-pole operators have relatively few cases of tracking targets deviating from the field of view and using data guidance, but there are large differences in stabilizing the tracking target, it is necessary to give a higher weight to the tracking stability score. Therefore, the guidance project score, data validity score, and tracking stability score can be weighted 0.2, 0.2, and 0.6 respectively, so that the final training score is more in line with the actual situation and ensure the reliability of the simulation training.

[0055] In this embodiment, a training image is obtained by loading a target image from a historical record, reading the absolute time, azimuth, pitch angle, and recorded frame, and performing image enhancement processing on the target image; a tracking window is established to track the target center in the training image to facilitate target tracking and determine the target offset in the field of view; a simulation servo control model is constructed, the training image is continuously played in the tracking window at the recorded frame rate, and target tracking training is performed by controlling the simulation servo control model to achieve simulation tracking training of the target; after the training is completed, the target tracking is scored according to the target tracking situation to obtain a training scoring result, which includes a guidance item score, a data validity score, a tracking stability score, and a comprehensive score, so that the training results can be uniformly evaluated to facilitate subsequent operator skill improvement, achieve accurate simulation of the target's real motion trajectory, and be applicable to a variety of target scenarios. It provides a platform and a real operating experience for daily training, assessment, and evaluation of single-pole operators of optical measurement equipment, has strong simulation, wide applicability, high tracking efficiency, and effectively solves the training problem of single-pole operators.

[0056] like Figure 3 As shown, a photometric real-scene simulation training system 20 is provided, which is used to implement the photometric real-scene simulation training method described above, including: a training image acquisition module 31, a tracking window establishment module 32, a simulation model construction module 33 and a target tracking scoring module 34, wherein: The training image acquisition module 31 is used to load the target image recorded in the historical records, read the absolute time, azimuth, pitch angle and recording frame rate, and perform image enhancement processing on the target image to obtain a training image; A tracking window establishing module 32 is used to establish a tracking window, and the tracking window is used to track the center of the target in the training image; The simulation model construction module 33 is used to construct a simulation servo control model, continuously play the training image in the tracking window according to the recorded frame rate, and perform target tracking training by controlling the simulation servo control model; The target tracking scoring module 34 is used to perform target tracking scoring according to the target tracking situation after the training is completed, and obtain the training scoring result. The training scoring result includes the guidance project score, data validity score, tracking stability score and comprehensive score.

[0057] In one embodiment, the tracking window establishment module 32 is specifically used to: establish a tracking window so that the tracking window has the same resolution as the training image, and the window center of the tracking window coincides with the geometric center of the first frame of the training image, and the geometric center of the training image is the center of the tracking target; load the training image, and add a crosshair in the window center of the tracking window to display the reference values of the azimuth and pitch angles of the encoder corresponding to the window center, and establish a simulated photoelectric theodolite tracking angle; and determine the offset of the target in the center of the field of view by identifying the position of the tracking target relative to the crosshairs.

[0058] In one embodiment, the simulation model construction module 33 is specifically configured to: construct a simulation servo control model using a serial port card, a data acquisition card, and a simulated single rod, wherein the simulated single rod has the same characteristics as the device single rod and can generate the same angular velocity output value as the device single rod; continuously play the training image in the tracking window at a recorded frame rate to form a training video; when the training video is trained to track a target using the simulation servo control model, the simulated single rod is controlled to send an output signal, the simulated servo control model converts the output signal into a digital signal, and sends the signal to the serial port card via a serial port; the serial port card receives the digital signal, and the data acquisition card is used to acquire the digital signal on the serial port card to generate azimuth and pitch angle data of the simulation encoder; obtain an angle change based on the angular velocity output value generated by the simulated single rod, calculate the pixel distance of the displacement of each frame of the image based on the angle change, and perform overall image translation based on the pixel distance; and perform edge padding on the translated training image frame by frame in the tracking window.

[0059] In addition, the final training status and assessment results will be saved to the data management module, which is mainly a MySQL database used to store training results, query historical results and display results through charts.

[0060] When programming the system, the design and development of the optical measurement real-scene simulation training system can be completed on the Qt programming platform. In order to reproduce the real optical tracking scene in the training system, the C\S (Client\Server) architecture is adopted. First, the target image of the historical record is loaded, and the absolute time, azimuth, pitch angle, recording frame rate and other information are read in the image. The image enhancement processing is performed and a tracking window is established. The target image is played continuously according to the recorded frame rate. Then a simulation servo control model is established. The operator conducts simulation tracking training by manipulating a simulated single rod. The simulation servo control model generates an encoder value and compares it with the encoder change value in the original image. The pixel distance of each frame image displacement is calculated based on the change in the value. In order to make the displaced image display complete, the edge filling of the image is completed frame by frame in the tracking window. After all the images are played, the system will automatically generate four scoring values to evaluate the quality of this tracking. The system architecture is as follows Figure 4As shown, the system can be used for daily target tracking training of single-pole operators of photoelectric theodolite to improve their tracking level.

[0061] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0062] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by the computing device, which can then be stored in a computer storage medium (ROM / RAM, magnetic disk, optical disk) and executed by the computing device. In some cases, the steps shown or described herein can be performed in a different order than that shown, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, the present invention is not limited to any particular combination of hardware and software.

[0063] The above content is a further detailed description of the present invention in conjunction with specific embodiments, and the specific implementation of the present invention cannot be considered to be limited to these descriptions. For those skilled in the art of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can be made, which should be considered to fall within the scope of protection of the present invention.

Claims

1. A photometric real-scene simulation training method, characterized in that: The following steps are involved: Loading a target image recorded in the historical records, reading the absolute time, azimuth, pitch angle, and recording frame rate, and performing image enhancement processing on the target image to obtain a training image; Establishing a tracking window, wherein the tracking window is used to track the center of the target in the training image; Constructing a simulation servo control model, continuously playing the training image in the tracking window according to the recorded frame rate, and performing target tracking training by controlling the simulation servo control model; After the training is completed, the target tracking score is performed according to the target tracking situation to obtain the training score result, which includes the guidance project score, data validity score, tracking stability score and comprehensive score.

2. The optical measurement real scene simulation training method according to claim 1, characterized in that: The recording frame rate is calculated by the absolute time difference between two consecutive frames of images.

3. The optical measurement real scene simulation training method according to claim 1, characterized in that: The establishing of the tracking window includes: Establishing a tracking window such that the tracking window has the same resolution as the training image, and the center of the tracking window coincides with the geometric center of the first frame of the training image, where the geometric center of the training image is the center of the tracking target; Loading the training image, and adding a crosshair in the center of the tracking window to display the reference values of the azimuth and pitch angles of the encoder corresponding to the center of the window, and establishing a simulated photoelectric theodolite tracking angle; By identifying the position of the tracking target relative to the crosshairs, the target's offset from the center of the field of view can be determined.

4. The optical measurement real scene simulation training method according to claim 1, characterized in that: The constructing of the simulation servo control model, continuously playing the training image in the tracking window according to the recorded frame rate, and performing target tracking training by controlling the simulation servo control model, includes: A simulation servo control model is constructed using a serial port card, a data acquisition card, and a simulated single rod. The simulated single rod has the same characteristics as the device single rod and can generate the same angular velocity output value as the device single rod. Continuously playing the training image in the tracking window at a recorded frame rate to form a training video; When the training video is subjected to target tracking training by the simulation servo control model, the simulation single rod is controlled to send an output signal, the simulation servo control model converts the output signal into a digital signal, and sends the signal to the serial port card via a serial port, the digital signal is received by the serial port card, and the digital signal on the serial port card is collected by the data acquisition card to generate azimuth and pitch angle data of the simulation encoder; According to the angular velocity output value generated by simulating a single rod, an angle change is obtained, a pixel distance of displacement of each frame of the image is calculated according to the angle change, and the entire image is displaced according to the pixel distance; The edge of the shifted training image is filled frame by frame in the tracking window.

5. The optical measurement real scene simulation training method according to claim 4, characterized in that: The method of obtaining the angle variation according to the angular velocity output value generated by simulating a single rod includes: When manipulating the simulated single pole, the angular change generated by the training image is calculated based on the angular velocity output value of the simulated single pole. The formula is: ， Where, is the initial driving value of the simulated single pole in the horizontal direction, To simulate the horizontal angular velocity output value of a single rod, is the driving coefficient of the simulated single rod, which is a constant. To manipulate the azimuth angle change produced by the simulated single rod, When the operation range is small and does not reach the single-pole initial drive value, the azimuth angle does not change. To simulate the maximum output value that a single rod can produce in the horizontal direction, when When , the resulting azimuth angle change is only related to the maximum output value; Similarly, the pitch angle change is obtained as: ， Where, is the initial drive value of the simulated single rod in the vertical direction, To simulate the vertical angular velocity output value of a single rod, To manipulate the pitch angle change produced by the simulated single stick, When the control range is small and does not reach the single-stick initial drive value, the pitch angle will not change. To simulate the maximum output value that a single rod can produce in the vertical direction, when , the resulting pitch angle change is only related to the maximum output value.

6. A photometric real-scene simulation training method according to claim 5, characterized in that: The step of calculating the pixel distance of each frame image displacement according to the angle variation and performing overall image displacement according to the pixel distance further includes: Calculating the pixel distance of each frame image displacement according to the direction angle change and the pitch angle change; Performing overall image displacement on the continuously played training images according to the pixel distance to obtain a true motion trajectory of the target; Among them, starting from the second frame image, the overall image displacement processing is performed based on the encoder value of the previous frame image, so that the image displacement direction is consistent with the actual tracking target movement direction. The formula for the pixel value and azimuth angle change of the training image in the horizontal direction of the x-axis is: ， Where, is the azimuth value of the current image, is the azimuth value of the previous frame image, To manipulate the azimuth angle change produced by the simulated single pole, is the final azimuth change of the image, is the field of view angle in the x-axis direction, is the resolution of the current training image, is the pixel value of the image horizontally translated in the x-axis direction; Similarly, the formula for the pixel value and pitch angle change of the training image translated vertically along the y-axis is: ， Where, is the pitch angle value of the current image, is the pitch angle value of the previous frame image, To simulate the pitch angle change caused by a single-pole operation, is the final pitch angle change of the image, is the field of view angle in the y-axis direction, which is a fixed value. is the resolution of the current image, The pixel value by which the image is translated vertically along the y-axis.

7. A photometric real-scene simulation training method according to claim 6, characterized in that: The edge filling of the shifted training image frame by frame in the tracking window includes: Detecting the positional relationship between the training image after the overall displacement of the image and the tracking window, and removing all pixels that are moved out of the tracking window to obtain a blank area; Select pixels of fixed width and height in the training image connecting the blank areas as reference pixels; Based on the horizontal direction and the vertical direction, the reference pixels are copied to the blank area of the tracking window to perform edge filling.

8. The optical measurement real scene simulation training method according to claim 5, characterized in that: Also includes: During the training process, if the tracking target disappears from the field of view or deviates from the center of the field of view by more than a threshold, data guidance is performed, that is, the geometric center of the training image is moved to the center of the window according to the maximum angular velocity of the simulated single rod. After the geometric center of the training image completely coincides with the center of the tracking window, manual tracking of the target is resumed.

9. The optical measurement real scene simulation training method according to claim 8, characterized in that: After the training is completed, target tracking scoring is performed according to the target tracking situation to obtain a training scoring result, including: The guided project score is scored according to the proportion of image frames guided by data in all training images, and the formula is: ， Where, is the number of image frames guided by data, is the total number of training image frames, This is the final score for the guidance project. If the percentage of image frames using data guidance exceeds 0.1, it is considered that the guidance time is too long and is directly judged as 0 points. The effective score of the data is scored according to the target's deviation from the field of view, and the formula is: ， Where, The number of image frames with invalid data. is the total number of training image frames, This is the final score for data validity; if the percentage of image frames that deviate from the field of view exceeds 0.1, the tracking condition is considered poor and is scored 0 points; When the azimuth angle change or the pitch angle change exceeds half of the field of view, the frame image is deemed to have deviated from the field of view and the data is considered invalid. The tracking stability score is scored according to the offset of the tracking target from the image center, and the formula is: ， Where, is the pixel value of the image horizontally shifted in the x-axis direction, is the average value of the pixel values of each frame image horizontally translated in the x-axis direction, is the pixel value of the image translated vertically along the y-axis. is the average value of the pixel values of each frame image translated vertically in the y-axis direction, is the root mean square of the azimuth angle, reflecting the stability of the operation in the horizontal direction. is the root mean square of the pitch angle, reflecting the stability of the operation in the vertical direction. It can reflect the stability of the overall tracking of the operation. To keep track of the stable final score; The comprehensive score is calculated based on the guided project score, data validity score and tracking stability score, and the formula is: ， Where, 、 and are the weights of the guided project score, data validity score and tracking stability score respectively.

10. A photometric real-scene simulation training system, characterized in that: A method for implementing a photometric real-scene simulation training method according to any one of claims 1 to 9, comprising: A training image acquisition module is used to load a target image recorded in the historical records, read the absolute time, azimuth, pitch angle and recording frame rate, and perform image enhancement processing on the target image to obtain a training image; A tracking window establishing module, configured to establish a tracking window, wherein the tracking window is used to track the center of the target in the training image; A simulation model construction module is used to construct a simulation servo control model, continuously play the training image in the tracking window according to the recorded frame rate, and perform target tracking training by controlling the simulation servo control model; The target tracking scoring module is used to perform target tracking scoring according to the target tracking situation after training is completed, and obtain the training scoring results. The training scoring results include the guidance project score, data validity score, tracking stability score and comprehensive score.

Images