Precise displacement measurement method of laser level meter based on data analysis

Abstract

The application relates to the technical field of building engineering monitoring, in particular to a precise displacement measurement method of a laser level measuring instrument based on data analysis, which uses an improved tracking algorithm to track and analyze a coordinate data sequence obtained through a laser level measuring instrument, and updates a tracking model in the improved tracking algorithm based on a model update learning rate during the analysis process; wherein the improved tracking algorithm is analyzed based on tracking accuracy and tracking frame rate, and the maximum straight-line distance between a transmitting end and a receiving end of the laser level measuring instrument and the maximum tracking angular velocity when the receiving end rotates can be determined during use; based on the tracking analysis result, the measurement parameters of the laser level measuring instrument can be adaptively adjusted, and the next cycle of displacement measurement is carried out, so that the final precise displacement measurement result is output. During the measurement process, the application can respond to environmental changes in real time, dynamically compensate for measurement deviations, stably track the measured target, and improve the measurement efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of construction engineering monitoring, and in particular to a precise displacement measurement method of a laser level measuring instrument based on data analysis. BACKGROUND

[0002] During the operation of hydraulic structures such as reservoir dams and sluices, the vertical and horizontal displacements thereof change with time and need to be accurately monitored periodically. Traditional monitoring methods mainly rely on precise optical levels and establish a vertical displacement monitoring control network including reference points, working base points and displacement measurement points. Although these methods have certain reliability in terms of monitoring accuracy, they still rely on manual operation and cannot meet the modern monitoring requirements of automation, all-weather, real-time and unattended operation. In addition, the horizontal displacement monitoring usually uses the collimation line method or the tension line method, and the existing methods cannot simultaneously realize the integrated monitoring of horizontal and vertical displacements in the same instrument or device, resulting in low integration of the monitoring system, complex operation and limited efficiency.

[0003] Chinese Patent Publication No. CN118654577A discloses an arch dam displacement monitoring device and method based on laser collimation and ranging. In the related technical solution, the relative displacement measurement element is used to obtain the three-phase displacement between adjacent measurement points of multiple measurement points of the arch dam, and the displacement coordinate conversion element is used to convert the relative horizontal displacement and the relative vertical displacement in the target coordinate system. Then, the arch dam displacement monitoring element is used to calculate the absolute three-direction displacement of each measurement point to obtain the actual displacement of the arch dam, so as to simultaneously monitor the horizontal displacement and the vertical displacement, thereby solving the problem of simultaneously monitoring the horizontal and vertical displacements to a certain extent. However, this solution does not consider the influence of dynamic changes in the structure displacement scale in the actual monitoring process, lacks the ability to adaptively adjust the monitoring scale, and when the displacement change range is large or the structure response is complex, it is difficult to ensure the measurement accuracy and system tracking robustness under all working conditions, thereby limiting its popularization and application in modern engineering monitoring scenarios with unattended operation and high automation.

[0004] Therefore, there is an urgent need for a measurement method that can automatically identify the structure displacement change scale and adjust the monitoring strategy in real time. On the basis of ensuring the simultaneous acquisition of horizontal and vertical displacements, the method has multi-scale adaptive perception and dynamic response capability, thereby meeting the urgent needs of water conservancy engineering for high-precision intelligentization and all-weather monitoring. SUMMARY

[0005] To this end, the present application provides a precise displacement measurement method of a laser level measuring instrument based on data analysis, which overcomes the problem of low measurement efficiency caused by the inability to respond to environmental changes in real time and dynamically compensate for measurement deviations in the prior art.

[0006] To achieve the above object, the application provides a precise displacement measurement method of a laser leveling instrument based on data analysis, comprising:

[0007] The transmitting end in the laser leveling instrument is arranged on the measured target, and the image data of the measured target is acquired by the receiving end at a preset time interval to determine the original three-dimensional coordinate data, wherein the original three-dimensional coordinate data comprises a vertical displacement value and a horizontal displacement value;

[0008] The original three-dimensional coordinate data is preprocessed to eliminate gross errors and random noise, and a preprocessed coordinate data sequence is obtained;

[0009] The coordinate data sequence is subjected to multi-scale tracking analysis using an improved tracking algorithm, and the tracking model in the improved tracking algorithm is updated based on the model update learning rate during the analysis process;

[0010] Based on the multi-scale tracking analysis result, the measurement parameters of the laser leveling instrument are adaptively adjusted, and the measurement parameters include the straight-line distance between the transmitting end and the receiving end during use and the tracking angular velocity;

[0011] Based on the adjusted measurement parameters, the displacement measurement in the next cycle is carried out to output the final precise displacement measurement result;

[0012] When performing multi-scale tracking analysis, the algorithm performance of the improved tracking algorithm is analyzed based on the tracking accuracy and the tracking frame rate, and the maximum straight-line distance between the transmitting end and the receiving end during use of the laser leveling instrument and the maximum tracking angular velocity of rotation of the laser leveling instrument are determined based on the analysis result of the algorithm performance.

[0013] Further, the process of determining the vertical displacement value and the horizontal displacement value comprises:

[0014] A vertical displacement monitoring network is constructed, and the vertical displacement value of each measured target is determined based on the spot center height of the transmitting end and the spot center height of the receiving end in the laser leveling instrument;

[0015] A horizontal displacement monitoring network is constructed, and the horizontal displacement value of each measured target is determined based on the spot position change amount of the transmitting end on the receiving end.

[0016] Further, the improved tracking algorithm is:

[0017] In the initialization stage of target tracking, the tracking quality index value of the first N frames of image data is acquired by the laser leveling instrument, and the average value is calculated as the reference quality index value, wherein N is less than or equal to 5;

[0018] In the subsequent tracking process, the tracking quality index value of the current frame is calculated in real time;

[0019] determining a model update learning rate corresponding to updating the tracking model based on the reference quality indicator value, the tracking quality indicator value of the current frame, and an adaptive learning rate;

[0020] The adaptive learning rate is calculated according to the reference quality indicator value, the tracking quality indicator value of the current frame, and a fixed learning rate.

[0021] Further, the process of determining the model update learning rate corresponding to updating the tracking model based on the reference quality indicator value, the tracking quality indicator value of the current frame, and the adaptive learning rate comprises:

[0022] If the tracking quality indicator value of the current frame is less than a first quality indicator value, it is determined to stop updating the tracking model;

[0023] If the tracking quality indicator value of the current frame is greater than a second quality indicator value, it is determined to update the tracking model using the fixed learning rate;

[0024] If the tracking quality indicator value of the current frame is greater than or equal to the first quality indicator value and less than or equal to the second quality indicator value, it is determined to update the tracking model using the adaptive learning rate;

[0025] The first quality indicator value is a product of a first ratio coefficient and the reference quality indicator value, and the second quality indicator value is a product of a second ratio coefficient and the reference quality indicator value.

[0026] Further, the process of calculating the tracking quality indicator value of the current frame comprises:

[0027] generating a target response map based on a correlation filter response in the improved tracking algorithm;

[0028] extracting a maximum value in the target response map and recording it as a response maximum value;

[0029] calculating a mean value and a standard deviation of the target response map in a sidelobe region excluding a predetermined region around the peak value;

[0030] calculating the tracking quality indicator value of the current frame based on the response maximum value, the mean value, and the standard deviation.

[0031] Further, the tracking accuracy rate is determined based on a ratio of the number of images successfully tracked to the total number of images;

[0032] The tracking frame rate is determined based on the number of images processed per second during program running.

[0033] The image with a tracking success is an image in which a pixel coordinate difference between a center point of a target frame tracked by the improved tracking algorithm and an actual center point is less than a preset pixel.

[0034] Further, a number of times that the tracking accuracy is greater than or equal to a preset tracking accuracy is counted and recorded as an accurate tracking number.

[0035] A number of times that the tracking frame rate is greater than or equal to a preset tracking frame rate is counted and recorded as a super frame number.

[0036] The maximum linear distance is determined based on the accurate tracking number, and the maximum tracking angular velocity is determined based on the super frame number.

[0037] Further, if the accurate tracking number is greater than or equal to a preset tracking number, a linear distance between the current transmitting end and the receiving end is determined as the maximum linear distance.

[0038] If the super frame number is less than a preset super frame number, an angular velocity of the current rotation of the receiving end is determined as the maximum tracking angular velocity.

[0039] Further, if the accurate tracking number is less than the preset tracking number, it is determined to adjust a position of the receiving end.

[0040] If the super frame number is greater than or equal to a preset super frame number, it is determined to adjust an angular velocity of the rotation of the receiving end.

[0041] Further, based on a comparison result of a tracking number difference value and a preset tracking number difference value, it is determined to shorten the linear distance between the transmitting end and the receiving end, and a shortening amplitude of the linear distance is in a positive correlation with the tracking number difference value.

[0042] Based on a comparison result of a super frame number difference value and a preset super frame number difference value, it is determined to reduce the angular velocity of the rotation of the receiving end, and a reduction amplitude of the angular velocity is in a positive correlation with the super frame number difference value.

[0043] The tracking number difference value is a difference value between the preset tracking number and the accurate tracking number, and the super frame number difference value is a difference value between the super frame number and the preset super frame number.

[0044] Compared with the prior art, the precision displacement measurement method of the data analysis based laser leveling instrument has the beneficial effects that: the original three-dimensional coordinate data of the measured target is collected by the laser leveling instrument and preprocessed, then the processed coordinate data sequence is subjected to multi-scale tracking analysis using the improved tracking algorithm, and the tracking model in the improved tracking algorithm is updated based on the model update learning rate during the analysis; during the multi-scale tracking analysis, the performance of the improved tracking algorithm is analyzed based on the tracking accuracy and the tracking frame rate, and the maximum straight-line distance between the transmitting end and the receiving end of the laser leveling instrument during use and the maximum tracking angular velocity during rotation of the receiving end can be accurately determined based on the performance analysis result; the measurement parameters of the laser leveling instrument can be adaptively adjusted based on the tracking analysis result, and then the next period of displacement measurement is performed based on the adjusted measurement parameters, so as to output the final precision displacement measurement result. In this way, the improved tracking algorithm can stably track the transmitting end, so that the measured target is always in the camera field of view, and the environmental changes can be responded in real time and the measurement deviation can be dynamically compensated, thereby improving the tracking accuracy of the laser leveling instrument for the target in the changing environment, and improving the measurement efficiency.

[0045] Further, in the process of multi-scale tracking analysis based on the coordinate data sequence, the real-time performance index of the tracking algorithm is introduced for multi-parameter fusion analysis, and a composite control model is constructed by fusing two types of key multi-source information of coordinate data and algorithm state; first, multi-scale dynamic analysis is performed based on the coordinate data sequence, and then secondary verification and correction are performed in combination with the tracking accuracy and the tracking frame rate; the laser leveling instrument can accurately respond to the real-time fluctuations of the measurement environment and the target state, so that high-precision parameter pre-adaptation is completed before displacement measurement is performed, thereby fundamentally improving the precision of the measurement result and the environmental robustness of the system.

[0046] Further, when it is determined that the measurement distance needs to be optimized, the shortening range of the straight-line distance can be accurately determined based on the comparison result of the tracking number difference and the preset tracking number difference; when it is determined that the rotational angular velocity needs to be speed-limited adjusted, the reduction range of the tracking angular velocity can be accurately determined based on the comparison result of the super-frame number difference and the preset super-frame number difference, so as to realize differentiated intelligent adjustment of the pose parameters (distance and angular velocity) of the measuring instrument, thereby improving the final accuracy of displacement measurement and the dynamic reliability of the measurement process. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 The flowchart of the precision displacement measurement method of the data analysis based laser leveling instrument of the embodiment of the present application Figure One ;

[0048] Figure 2Flowchart of the precise displacement measurement method of the laser leveling instrument based on data analysis according to the embodiment of the present application Figure Two ;

[0049] Figure 3 Schematic diagram of the comparison result of the tracking algorithm according to the embodiment of the present application

[0050] Figure 4 Logical determination diagram for adjusting the maximum straight-line distance based on the accurate tracking number according to the embodiment of the present application

[0051] Figure 5 Logical determination diagram for adjusting the maximum angular velocity based on the superframe number according to the embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the objects and advantages of the present application clearer, the present application will be further described below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0053] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0054] It should be noted that, in the description of the present application, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer" and the like are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0055] In addition, it should also be noted that, in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the internal communication of two elements. Those skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0056] In actual monitoring scenarios, the measured target such as a dam structure produces displacement of different amplitudes and directions due to environmental load changes, which is reflected in the monitoring data as differences in "scale", thereby causing the distance between the receiving end in the measuring instrument and the transmitting end installed on the measured target to always change in the actual application scenario of target tracking, leading to changes in the imaging size of the target, which is manifested as changes in the scale of the target on the image. If the monitoring system uses a fixed measurement scale or mode for data acquisition and processing, when the displacement amplitude is large, key deformation details are easily lost, affecting the structural integrity of the judgment; and when the displacement change is very subtle, environmental interference may be introduced due to insufficient resolution, causing misjudgment or missed judgment, ultimately leading to "drift" of the monitoring data and a decrease in reliability. Therefore, when monitoring the displacement of reservoir dams, sluices and other hydraulic structures, it is particularly important to construct a displacement monitoring mechanism with multi-scale perception and self-adaptive capability.

[0057] Referring to Figure 1 Fig. 1 is a flowchart of a precise displacement measurement method of a laser leveling instrument based on data analysis according to an embodiment of the present application. Figure One The flowchart includes at least the following steps:

[0058] S1: The transmitting end in the laser leveling instrument is arranged on the measured target, and the receiving end acquires image data of the measured target at a predetermined time interval to determine original three-dimensional coordinate data, wherein the original three-dimensional coordinate data includes a vertical displacement value and a horizontal displacement value;

[0059] S2: The original three-dimensional coordinate data is preprocessed to eliminate gross errors and random noise, and a preprocessed coordinate data sequence is obtained;

[0060] S3: The coordinate data sequence is subjected to multi-scale tracking analysis using an improved tracking algorithm, and a tracking model in the improved tracking algorithm is updated based on a model update learning rate during the analysis process;

[0061] S4: Based on the multi-scale tracking analysis result, the measurement parameters of the laser leveling instrument are adaptively adjusted, including the straight-line distance between the transmitting end and the receiving end during use and the tracking angular velocity.

[0062] S5: Based on the adjusted measurement parameters, the next period of displacement measurement is performed to output the final precise displacement measurement result.

[0063] Referring to Figure 2 Fig. 2 is a flowchart of a precise displacement measurement method of a laser leveling instrument based on data analysis according to another embodiment of the present application. Figure Two The flowchart includes at least the following steps:

[0064] S31: When performing multi-scale tracking analysis, the algorithm performance of the tracking algorithm is analyzed based on tracking accuracy and tracking frame rate;

[0065] S32: Based on the analysis result of the algorithm performance, the maximum straight-line distance between the transmitting end and the receiving end during use is determined, as well as the maximum tracking angular velocity when the receiving end rotates.

[0066] Specifically, the laser leveling instrument automatically scans all the measured targets of the measured targets every 30 minutes according to the initial setting, obtains the original three-dimensional coordinate numbers and sends the data to the central processing server. The central processing server is preprocessed to obtain a preprocessed coordinate data sequence, wherein the process of eliminating gross errors and random noise includes, for example, identifying and removing coordinate abnormal jump points caused by temporary bird obstruction, instantaneous strong refraction, etc., and using Kalman filtering or wavelet denoising technology to smooth the data and retain the true displacement trend.

[0067] The coordinate data sequence is subjected to multi-scale tracking analysis using an improved tracking algorithm. Five scales are exemplarily set, the scale pool technology is used to track the target at five scales, only one correlation filter is trained to solve the target position and target scale, the algorithm extracts candidate samples at multiple scales, uses the trained correlation filter to calculate the candidate region response value and target position of each scale, and determines the scale and position of the target; wherein the scale pool technology is a technology used when the scale adaptive multiple feature (SAMF) algorithm is used for multi-scale tracking to solve the target center position and the best target scale, and the selected scale coefficient is as follows:

[0068] ;

[0069] According to the size of the target in the last frame, it is multiplied by the scale coefficient to obtain five scale candidate regions. In order to facilitate feature extraction and correlation filter training, and at the same time to ensure the calculation speed of the algorithm under multi-scale tracking, the five candidate regions are compressed to the same template size after normalization processing, and the formula is as follows:

[0070] ;

[0071] wherein, W and H are the width and height of each candidate region, respectively, T is the template size, is the normalization coefficient, W and H are the width and height of each candidate region after normalization, respectively.

[0072] The candidate regions of the five scales are all generated by a cyclic shift operation to generate a candidate sample set, the correlation response values of the candidate samples of the five scales are calculated according to the trained correlation filter, the candidate region with the maximum response value is determined as the size of the target current frame, and the center position of the target is determined according to the peak position of the response graph; the center position of the target is repeatedly obtained to realize multi-scale tracking of the target.

[0073] Specifically, after completing the multi-scale tracking analysis, the measurement parameters of the laser leveling instrument can be adaptively adjusted, including the straight-line distance between the transmitting end and the receiving end of the laser leveling instrument during use, the tracking angular velocity; in other embodiments, the measurement parameters also include laser power, sampling frequency and data processing process, and after adjusting the laser power, sampling frequency and data processing process, the next period of displacement measurement is performed to output the final accurate displacement measurement result.

[0074] The transmitting end is installed on the corresponding measured target of the measured target, and the receiving end is adjustably arranged at another stable position capable of receiving the signal sent by the transmitting end; wherein the transmitting end moves with the measured target, and the receiving end includes a camera and a two-axis turntable.

[0075] If it is confirmed through analysis that the displacement time history curves of the multiple measured targets show high-frequency, low-amplitude vibration characteristics, and the noise significantly increases; at this time, the tracking angular velocity can be increased to follow the rapid jumping of the laser light spot on the receiving end caused by vibration, the laser power instrument can be increased to ensure that a clear light spot signal can still be obtained in rapid vibration, the sampling frequency can be increased to capture high-frequency vibration details, and the algorithm in the data processing process can be temporarily switched to a more powerful filtering algorithm to more reliably extract the real vibration signal from the noisy original data.

[0076] If it is confirmed through analysis that the displacement sequence is stable and the standard deviation is minimal; at this time, the tracking angular velocity can be restored to the regular setting, the sampling frequency can be reduced and the laser power can be adjusted, and the algorithm in the data processing process can be switched back to the standard Kalman filter with smaller calculation amount; by adaptively reducing the measurement parameters, the system greatly reduces the power consumption and data storage amount on the premise of ensuring continuous monitoring.

[0077] In the embodiment, the vertical displacement values are also obtained by constructing a vertical displacement monitoring network, wherein the measuring instrument accurately calculates the elevation change of each measured target relative to the instrument reference surface by measuring the geometric relationship between the center height of the transmitting laser of the transmitting end and the center height of the receiving end of the received light, thereby determining the vertical displacement values of the measured targets; and the horizontal displacement values are obtained by constructing a horizontal displacement monitoring network, wherein the measuring instrument accurately identifies the light spot center formed by the laser on the receiving end, and directly calculates the horizontal displacement values of the measured targets perpendicular to the line of sight by monitoring the change amount of the light spot position.

[0078] Wherein, the laser level measurement instrument based on high-precision turntable is adopted, on the basis of the vertical displacement measurement of the original surface deformation observation point, the high-precision turntable is used as a reference, and the horizontal displacement and vertical displacement of the surface deformation observation point are realized at the same time, so as to meet the requirements of automatic monitoring of the regulations and specifications. The measurement of horizontal displacement adopts high-precision turntable as a reference, wherein the horizontal displacement measurement of single-point monitoring position adopts the receiving end collector in the laser level measurement instrument for measurement. In the aspect of vertical displacement measurement, the transmitting end and the receiving end camera collector of the laser level measurement instrument are used to obtain the vertical displacement change, and the reference line is used as a horizontal reference to transfer the vertical displacement.

[0079] Further, when the dam or the large gate is the measurement target object, the vertical displacement measurement of the laser level measurement instrument includes the following steps:

[0080] 1. Known reference elevation: obtain the elevations of the level base points A and B of the dam or the large gate, which are Ha and Hb respectively.

[0081] 2. Measuring the elevation of the work base point and the measured target: after the installation of the base is completed, the elevations of the level base points A and B are measured to the vertical displacement monitoring work base points 0# and 4#, and the monitoring points 1#, 2# and 3# using the Leica first-order level, to determine the elevations of the bottom level marks of the cement piers, which are H0, H4, H1, H2 and H3 respectively, with the unit of m; wherein 1# and 4# are the base points, and the elevations corresponding thereto are generally constant, which can be periodically re-measured from the level base points A and B.

[0082] 3. Measuring the height of the base platform: when the measurement instrument is first installed, the height of the bottom level mark of each installation base to the top platform at 0#, 1#, 2#, 3# and 4# is measured using the Leica first-order level, which is h0, h1, h2, h3 and h4 respectively, with the unit of m.

[0083] 4. Calibrating the center height of the laser: before the first installation, the center height zi of the transmitting end of each measurement instrument and the zero center height bi of the matching receiving end should be calibrated on the horizontal platform to ensure that they are equal in height, with an error of less than 0.02 mm, i.e. z0≈b0, z1≈b1, z2≈b2, z3≈b3, z4≈b4, with the unit of mm.

[0084] 5. Image acquisition and data processing: the receiving end image collector needs to collect multiple images, first of all, abnormal images are eliminated, and then the Laplace criterion is applied to eliminate gross error data.

[0085] 6. System zero calibration: after the first installation is completed, the zero calibration of the elevation measurement system is carried out.

[0086] 7. Operation monitoring and temperature compensation: During system operation, the vertical displacement values at the top of each mounting base are monitored, and the measurement results are compensated according to the ambient temperature changes.

[0087] Further, the horizontal displacement measurement of the laser leveling instrument includes the following steps:

[0088] 1. Initial positioning: the initial position of the receiving end collector a is x0; when the receiving end collector b moves from position 1 to position 2, the spot position of the transmitting end 1 changes from x10 to x11, and the change is Δx1; if the transmitting end 1 does not move, the displacement Δd1 of the receiving end collector b is Δx1.

[0089] 2. Considering the displacement of the transmitting end: if the transmitting end 1 also moves Δx0 (which can be measured by target 0), the total displacement of the receiving end collector b relative to target 0 is Δx0+Δx1.

[0090] 3. Multi-transmitting-end joint solution: if the spot position of the transmitting end 2 changes from x20 to x21, the change is Δx2; the front and rear viewing angles are θ2 (approximately equal to 180°); if both transmitting ends 1 and 2 do not move, the measurement value of the receiving end collector c is Δx2; if only the transmitting end 1 moves, the displacement of the receiving end collector b is Δd1=Δx1, and at this time the receiving end collector c remains stationary; if the transmitting end 1 does not move, the total displacement of the receiving end collector b relative to target 0 is Δx0+Δx1+Δx2.

[0091] 4. Realize horizontal displacement transmission of monitoring stations: by analogy, it can be transmitted down without limit, that is, the horizontal displacement transmission of all monitoring stations can be realized.

[0092] Specifically, the improved tracking algorithm is: in the initialization stage of target tracking, the tracking quality index values of the first N frames of image data are obtained by the laser leveling instrument, and the average value is calculated as the reference quality index value, wherein N is less than or equal to 5, and the target in the initial five frames of images generally does not occur scenes such as occlusion, motion blur, rapid deformation, etc.; in the subsequent tracking process, the tracking quality index value of the current frame is calculated in real time; based on the reference quality index value, the tracking quality index value of the current frame and the adaptive learning rate, the model update learning rate corresponding to the updating of the tracking model is determined; wherein the adaptive learning rate is calculated according to the reference quality index value, the tracking quality index value of the current frame and the fixed learning rate.

[0093] The formula for calculating the model update learning rate is as follows:

[0094] ;

[0095] wherein, a fixed learning rate used in a Kernelized Correlation Filter (KCF) algorithm, an adaptive learning rate, a tracking quality indicator value of a current frame, a reference quality indicator value.

[0096] In the embodiment, the process of determining the model update learning rate corresponding to the update of the tracking model based on the reference quality indicator value, the tracking quality indicator value of the current frame and the adaptive learning rate includes: if the tracking quality indicator value of the current frame is less than a first quality indicator value, it is determined that the tracking fails at this time, and in this case, the filter template is no longer updated, and it is determined to stop updating the tracking model; if the tracking quality indicator value of the current frame is greater than a second quality indicator value, it is determined that the tracking effect is better at this time, and it is determined to update the tracking model using the fixed learning rate; if the tracking quality indicator value of the current frame is greater than or equal to the first quality indicator value and less than or equal to the second quality indicator value, it is determined that the tracking effect is not ideal at this time, and it is determined to update the tracking model using the adaptive learning rate to reduce the noise introduced by the filter; wherein the first quality indicator value is a product of a first ratio coefficient k1 and the reference quality indicator value, and the second quality indicator value is a product of a second ratio coefficient k2 and the reference quality indicator value.

[0097] In the embodiment, k1=0.4 and k2=0.8 can be exemplarily set. Using the adaptive filter update strategy as described above, the robustness of the laser leveling instrument in the measurement and tracking process can be improved.

[0098] In the embodiment, the process of calculating the tracking quality indicator value of the current frame includes: extracting a maximum value in a correlation filter target response map and recording it as a response maximum value; calculating the mean value and the standard deviation of the target response map in a sidelobe region excluding a predetermined region around the peak value; and calculating the tracking quality indicator value of the current frame based on the response maximum value, the mean value and the standard deviation. The improved tracking algorithm in the embodiment uses the correlation filtering technology, and the target response map is a peak value response map calculated by the improved tracking algorithm based on the correlation filtering technology.

[0099] The formula for calculating the tracking quality indicator value of the current frame is as follows:

[0100] ;

[0101] wherein, the response maximum value is, the mean value is, and the standard deviation is.

[0102] ​When the target appears to be blocked, motion blurred, rapidly deformed, etc., the tracking quality index value of the current frame of the response graph will decrease significantly, and the tracking quality index value of the current frame represents the target tracking scene change.

[0103] In the embodiment, a standard dataset and a dataset made according to an actual scene are used for verification, wherein the standard dataset uses an OTB100 dataset, 20 groups of image sequences are selected in the dataset, and a total of 5000 images are used for testing. An algorithm platform is based on Visual Studio 2015, OpenCV 3.4.3 and OpenCV_Contrib_3.4.3. A computer processor is Intel(R) Core(TM) i5-7400 CPU @3.00GHz, a Win7 system, and a machine band RAM is 8.00GB. The improved tracking algorithm in the embodiment is based on an improved Kernelized Correlation Filter (KCF), and the improved tracking algorithm is compared with a Kernelized Correlation Filter, a Circulant Structure with Kernels (CSK) and a Background-Aware Correlation Filter (BACF) to verify the performance of the algorithm.

[0104] Specifically, the tracking accuracy is determined based on a ratio of the number of images successfully tracked to the total number of images; and the tracking frame rate is determined based on the number of images processed per second during program execution. The image successfully tracked is an image in which a pixel coordinate difference between a center point of a target frame tracked by the improved tracking algorithm and an actual center point is less than a preset pixel. The preset pixel is exemplarily set to 22 pixels.

[0105] Referring to Figure 3 As shown in the figure, the pixel deviation of the tracking frame center point is a pixel coordinate difference between a center point of a target frame tracked by the algorithm and an actual center point, and the algorithm corresponding to Ours is the improved tracking algorithm mentioned in the embodiment. When the pixel deviation of the tracking frame center point is less than 20 pixels, the KCF tracking accuracy on the OTB100 dataset is 86.36%, the CSK tracking accuracy is 69.06%, the BACF tracking accuracy is 59.88%, and the improved tracking algorithm in the embodiment has a tracking accuracy of 90.02%. Therefore, the performance of the improved tracking algorithm in the embodiment is better than that of the comparative algorithm.

[0106] It should be noted that when the tracking accuracy is verified, when the target tracking frame is completely separated from the target real position, the tracking result has no evaluation significance, so in the embodiment, only the scene in which the target is not completely lost is counted.

[0107] Specifically, the industrial camera model for image data acquisition in the receiving end is FLIR GS3-U3-41C6C-C, the resolution is 2048x2048, the pixel unit size is 5.5 μm, the lens focal length is 50 mm, and the straight-line distance between the receiving end and the transmitting end is set to 2 m. According to the experimental process, the data set is made and the corresponding calculation is carried out, and the tracking algorithm comparison results obtained are shown in Table 1.

[0108] Table 1 Tracking algorithm comparison results

[0109] ;

[0110] As shown in Table 1, the accuracy of the improved tracking algorithm in the test results of the embodiment is 91.15%, and the frame rate is 19.5 frames per second, which has a high tracking accuracy while ensuring real-time. The accuracy result of the improved tracking algorithm obtained in the embodiment after the test is roughly consistent with the accuracy result calculated by the standard data set.

[0111] Referring to FIGS. 1 and 2, Figure 4 and Figure 5 , Figure 4 is a logic determination diagram for determining adjustment of the maximum straight-line distance based on the number of accurate tracking times in the embodiment of the present application, Figure 5 is a logic determination diagram for determining adjustment of the maximum angular velocity based on the number of superframes in the embodiment of the present application.

[0112] Specifically, the number of times that the tracking accuracy is greater than or equal to the preset tracking accuracy is counted and recorded as the number of accurate tracking times E; the number of times that the tracking frame rate is greater than or equal to the preset tracking frame rate is counted and recorded as the number of superframes B; the maximum straight-line distance is determined based on the number of accurate tracking times E, and the maximum angular velocity is determined based on the number of superframes B. By adjusting the maximum straight-line distance and the maximum angular velocity, the improved tracking algorithm can more stably track the measured target.

[0113] In the embodiment, a preset tracking number E0 corresponding to the number of accurate tracking times E is set, and the comparison result of the number of accurate tracking times E and the preset tracking number E0 is used to determine whether the straight-line distance between the transmitting end and the receiving end needs to be adjusted; a preset superframe number B0 corresponding to the number of superframes B is set, and the comparison result of the number of superframes B and the preset superframe number B0 is used to determine whether the tracking angular velocity of the receiving end needs to be adjusted.

[0114] In one specific embodiment, in order to ensure that the current measurement instrument position (distance) can continuously and stably provide high-quality tracking data, the preset tracking number E0=5 times can be set, for example. If E is greater than or equal to E0, it indicates that the receiving end at the current position can effectively ensure the tracking accuracy of the improved tracking algorithm, and thus the straight-line distance between the current receiving end and the transmitting end can be directly determined as the maximum straight-line distance. If E is less than E0, it indicates that the tracking accuracy at the current position is greater than or equal to the preset tracking accuracy, and the number of times is insufficient, that is, the receiving end at the current position cannot ensure the tracking accuracy of the improved tracking algorithm when performing image acquisition, and thus the position of the receiving end needs to be adjusted.

[0115] In one specific embodiment, in order to ensure that the current rotational angular velocity setting does not cause the system to be overburdened and can meet the real-time requirement, the preset superframe number B0=3 times can be set, for example. If B is less than B0, it indicates that the corresponding tracking angular velocity of the receiving end when performing displacement measurement is in line with the standard, and the receiving end can be rotated according to the corresponding tracking angular velocity to make the number of times when the tracking frame rate is greater than or equal to the preset tracking frame rate within an acceptable range, and thus the current tracking angular velocity of the receiving end can be determined as the maximum angular velocity. If B is greater than or equal to B0, it indicates that the tracking frame rate of the receiving end when tracking according to the current angular velocity is greater than or equal to the preset tracking frame rate, and the number of times is excessive, and thus the tracking angular velocity of the receiving end needs to be adjusted.

[0116] In this embodiment, the tracking number difference G is the difference between the preset tracking number E0and the accurate tracking number E. The greater the tracking number difference G, the smaller the corresponding accurate tracking number E, which indicates that even if the receiving end at the current position uses the improved tracking algorithm to cooperate with the transmitting end to perform accurate displacement measurement on the measured target, the number of times when the corresponding tracking accuracy is greater than or equal to the preset tracking accuracy is still insufficient. At this time, the straight-line distance between the receiving end and the transmitting end can be shortened to increase the accurate tracking number B, and thus the shortening amplitude of the straight-line distance and the tracking number difference G are in a positive correlation.

[0117] The superframe number difference Q is the difference between the superframe number B and the preset superframe number B0. The greater the superframe number difference Q, the greater the corresponding superframe number B, which indicates that the current rotational speed of the receiving end is too fast, causing the number of times when the tracking frame rate is greater than or equal to the preset tracking frame rate to be excessive. At this time, the angular velocity of the receiving end can be reduced to reduce the superframe number B, and thus the reduction amplitude of the angular velocity and the superframe number difference Q are in a positive correlation.

[0118] In one embodiment, a preset tracking number difference value G0 corresponding to the tracking number difference value G is set. In order to more accurately determine the reduction range of the straight line distance, the preset tracking number difference value G0 can be divided into a first preset tracking number difference value G1 and a second preset tracking number difference value G2. For example, G1 = 1 time and G2 = 3 times are set. The comparison process based on G, G1 and G2 is as follows:

[0119] If G is less than or equal to G1, a first distance adjustment instruction is determined to be generated. Based on the instruction, the receiving end is moved to approach the measured target. The straight line distance between the original receiving end and the transmitting end is shortened by 15%. For example, if the original straight line distance is 3 meters, the maximum straight line distance after the shortening is 2.55 meters. If G is greater than G1 and less than or equal to G2, a second distance adjustment instruction is determined to be generated. Based on the instruction, the receiving end is moved to approach the measured target. The straight line distance between the original receiving end and the transmitting end is shortened by 18%. If G is greater than G2, a third distance adjustment instruction is determined to be generated. Based on the instruction, the receiving end is moved to approach the measured target. The straight line distance between the original receiving end and the transmitting end is shortened by 25%.

[0120] In one embodiment, a preset superframe number difference value Q0 corresponding to the superframe number difference value Q is set. In order to more accurately determine the reduction range of the rotation angular velocity, the preset superframe number difference value Q0 can be divided into a first preset superframe number difference value Q1 and a second preset superframe number difference value Q2. For example, Q1 = 1 time and Q2 = 2 times are set. The comparison process based on Q, Q1 and Q2 is as follows:

[0121] If Q is less than or equal to Q1, a first angular velocity adjustment instruction is determined to be generated. Based on the instruction, the receiving end is controlled to rotate. The tracking angular velocity of the two-axis turntable in the original receiving end is reduced by 5%. For example, if the original tracking angular velocity is 25 degrees per second, the maximum tracking angular velocity after the reduction is 23.75 degrees per second. If Q is greater than Q1 and less than or equal to Q2, a second angular velocity adjustment instruction is determined to be generated. Based on the instruction, the receiving end is controlled to rotate. The tracking angular velocity of the two-axis turntable in the original receiving end is reduced by 8%. If Q is greater than Q2, a third angular velocity adjustment instruction is determined to be generated. Based on the instruction, the receiving end is controlled to rotate. The tracking angular velocity of the two-axis turntable in the original receiving end is reduced by 12%.

[0122] It can be understood that the shortening range of the straight line distance and the reduction range of the tracking angular velocity can also be set to other required values, for example, when G is greater than G2, it can also be determined that the original straight line distance is shortened by 23%, and when Q is greater than Q2, it can also be determined that the original rotational angular velocity is reduced by 14%. It should be noted that the straight line distance after being shortened and the tracking angular velocity after being reduced will not have a negative impact on the improved tracking algorithm. The shortened straight line distance here provides a reference relative installation position for the receiving end, so as to effectively perform accurate measurement. The receiving end in the laser leveling instrument can be moved through the linear movement platform; or the support structure of the receiving end can be directly driven by the stepping motor or the servo motor to realize high-precision position control.

[0123] The technologies not described in the above embodiments are applicable to the prior art.

[0124] It can be understood that in the embodiments of the present application, any one of the preset parameters or critical parameters is not specifically limited, and the above values are not limited thereto. The person skilled in the art can adjust the preset parameters or critical parameters according to the actual needs or the analysis of historical data or the use of the equipment.

[0125] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but the person skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. The person skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will fall within the protection scope of the present application.

Claims

1. A precise displacement measurement method for a laser leveling instrument based on data analysis, characterized in that, include: The transmitter of the laser leveling instrument is set on the target to be measured, and the receiver collects image data of the target to be measured at preset time intervals to determine the original three-dimensional coordinate data, which includes vertical displacement value and horizontal displacement value. The original three-dimensional coordinate data is preprocessed to eliminate gross errors and random noise, resulting in a preprocessed coordinate data sequence. The coordinate data sequence is subjected to multi-scale tracking analysis using an improved tracking algorithm. During the analysis, the tracking model in the improved tracking algorithm is updated based on the model update learning rate. Based on the results of multi-scale tracking analysis, the measurement parameters of the laser leveling instrument are adaptively adjusted. The measurement parameters include the straight-line distance and tracking angular velocity between the transmitting end and the receiving end during use. Based on the adjusted measurement parameters, the displacement measurement is performed in the next cycle to output the final accurate displacement measurement result; When performing multi-scale tracking analysis, the algorithm performance of the improved tracking algorithm is analyzed based on tracking accuracy and tracking frame rate. Based on the analysis results of the algorithm performance, the maximum straight-line distance between the transmitter and the receiver during use, as well as the maximum tracking angular velocity when the receiver rotates, are determined. The tracking accuracy is determined based on the ratio of the number of successfully tracked images to the total number of images. The tracking frame rate is determined based on the number of images processed per second during program runtime. Among them, the image that is successfully tracked is the image in which the difference between the pixel coordinates of the center point of the target box tracked by the improved tracking algorithm and the actual center point is less than a preset number of pixels; The number of times the tracking accuracy is greater than or equal to the preset tracking accuracy is counted and recorded as the number of accurate tracking times; The number of times the tracking frame rate is greater than or equal to the preset tracking frame rate is counted and recorded as the number of superframes; The maximum straight-line distance is determined based on the number of accurate tracking attempts, and the maximum tracking angular velocity is determined based on the number of superframe attempts. If the number of accurate tracking counts is greater than or equal to the preset number of tracking counts, the current straight-line distance between the transmitter and the receiver is determined to be the maximum straight-line distance. If the number of superframes is less than the preset number of superframes, the current angular velocity of the receiver is determined to be the maximum tracking angular velocity; If the number of accurate tracking attempts is less than the preset number of tracking attempts, the position of the receiving end is adjusted. If the number of superframes is greater than or equal to the preset number of superframes, determine to adjust the angular velocity of the receiver when it rotates. The straight-line distance between the transmitter and the receiver is shortened based on the comparison between the difference in the number of tracking attempts and the preset difference in the number of tracking attempts. The amount of shortening of the straight-line distance is positively correlated with the difference in the number of tracking attempts. The angular velocity of the receiver is reduced based on the comparison between the difference in the number of superframes and the preset difference in the number of superframes. The reduction in angular velocity is positively correlated with the difference in the number of superframes. Wherein, the difference in the number of tracking times is the difference between the preset number of tracking times and the accurate number of tracking times, and the difference in the number of superframes is the difference between the number of superframes and the preset number of superframes.

2. The precise displacement measurement method of the laser leveling instrument based on data analysis according to claim 1, characterized in that, The process of determining the vertical displacement value and the horizontal displacement value includes: A vertical displacement monitoring network is constructed, and the vertical displacement value of each target is determined based on the center height of the laser spot at the transmitting end and the center height of the laser spot at the receiving end of the laser leveling instrument. A horizontal displacement monitoring network is constructed, and the horizontal displacement value of each target is determined based on the change in the position of the light spot formed by the transmitting end on the receiving end.

3. The precise displacement measurement method of the laser leveling instrument based on data analysis according to claim 1, characterized in that, The improved tracking algorithm is as follows: During the initialization phase of target tracking, the tracking quality index values ​​of the first N frames of image data are obtained through the laser leveling instrument, and the average value is calculated as the benchmark quality index value, where N is less than or equal to 5. During subsequent tracking, the tracking quality index value of the current frame is calculated in real time; The model update learning rate is determined based on the baseline quality index value, the tracking quality index value of the current frame, and the adaptive learning rate when the tracking model is updated. The adaptive learning rate is calculated based on the baseline quality index value, the tracking quality index value of the current frame, and the fixed learning rate.

4. The precise displacement measurement method of the laser leveling instrument based on data analysis according to claim 3, characterized in that, The process of determining the model update learning rate for updating the tracking model based on the baseline quality metric value, the tracking quality metric value of the current frame, and the adaptive learning rate includes: If the tracking quality index value of the current frame is less than the first quality index value, it is determined to stop updating the tracking model; If the tracking quality index value of the current frame is greater than the second quality index value, it is determined that the tracking model will be updated using the fixed learning rate; If the tracking quality index value of the current frame is greater than or equal to the first quality index value and less than or equal to the second quality index value, it is determined that the adaptive learning rate will be used to update the tracking model. Wherein, the first quality index value is the product of the first ratio coefficient and the benchmark quality index value, and the second quality index value is the product of the second ratio coefficient and the benchmark quality index value.

5. The precise displacement measurement method of the laser leveling instrument based on data analysis according to claim 3, characterized in that, The process of calculating the tracking quality metric value of the current frame includes: Based on the correlation filter response in the improved tracking algorithm, a target response map is generated; Extract the maximum value from the target response graph and record it as the maximum response value; Within the sidelobe region excluding the predetermined region around the peak, the mean and standard deviation of the target response map are calculated; The tracking quality index value of the current frame is calculated based on the maximum response value, the mean value, and the standard deviation.

Images