Rotating mirror type binocular vision laser tracking broken light recovery system and method

Abstract

The invention discloses a rotating mirror type binocular vision laser tracking broken light recovery system and method. The system comprises a rotating mirror type laser tracker and a binocular vision module, two cameras in the binocular vision module are symmetrically installed on the two sides of a laser beam, and the visual axes of the cameras and the optical axis of the laser are reflected to a working area together through a rotating mirror of the tracker; the method comprises the following steps: pre-constructing a virtual imaging plane model in which a camera optical axis coincides with a laser optical axis; after light interruption occurs, modulating an infrared illumination light source of a visual system, obtaining and processing a frame difference image of a modulation wave crest and a modulation wave trough, obtaining a current position of a target mirror and a predicted light interruption recovery rotation angle, and performing error compensation on the current position and the predicted light interruption recovery rotation angle; and driving a corresponding motor to rotate according to the compensated light-cut recovery rotation angle, so that the laser beam is aligned with the target mirror again. According to the invention, the coaxial installation difficulty of the visual system of the rotating mirror type laser tracker is reduced, and the target mirror recognition rate and the light failure recovery servo angle accuracy in a large-range complex environment are improved.

Description

Technical Field

[0001] The invention relates to the field of laser tracking measurement technology and instruments, and in particular to a rotating mirror binocular vision laser tracking light interruption recovery system and method. Background Art

[0002] Laser trackers are typical instruments in the field of large-scale spatial coordinate precision measurement, and play an important role in industrial applications such as large-scale mechanical parts manufacturing measurement and large-scale equipment assembly monitoring. In laser tracking measurement, the laser beam needs to be aimed at the target mirror to ensure measurement efficiency. However, in actual operation, affected by various external factors, the laser beam will deviate from the target mirror and break the light. After the laser tracker breaks the light, it is mainly restored through manual guidance and machine vision guidance. Compared with manual guidance, machine vision guidance is more efficient. However, machine vision guidance requires the visual system axis and the laser optical axis to rotate synchronously. Due to the particularity of the structure of the rotating mirror laser tracker, the pitch rotation angle of the laser beam is twice the rotation angle of the pitch motor, which brings certain challenges to the recovery of the rotating mirror laser tracker.

[0003] The existing rotating mirror laser tracking light-break recovery method is to use a monocular camera vision system, install the monocular camera visual axis coaxially with the laser optical axis, and adjust the camera visual axis and the laser optical axis to be coaxial by mechanical structure. Since it is difficult to keep the two axes strictly coaxial, long-distance light-break recovery fails. Therefore, it is necessary to construct a nonlinear relationship model between the target mirror distance, the target mirror pixel deviation value and the tracking recovery servo angle to compensate for the calculation of the light-break recovery angle. However, this nonlinear model is relatively complex, and the distance parameter of the target mirror is required to calculate the recovery servo angle. When the distance parameter of the target mirror is unknown, light-break recovery cannot be performed. In addition, the current machine vision system mainly uses active infrared detection methods to identify the target mirror through infrared detection, and calculate the tracking recovery angle based on its positioning data. However, this method is easily interfered by other similar high-brightness light sources in the background, resulting in recognition failure, and as the measurement distance increases, the recognition accuracy of the target mirror will also drop significantly.

[0004] Therefore, how to improve the recognition accuracy of the target mirror in a large range of complex environments, and accurately calculate the servo angle for light failure recovery when the distance parameters of the target mirror are unknown, is a key issue facing the current laser tracking measurement technology. Summary of the invention

[0005] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a rotating mirror binocular vision laser tracking light failure recovery system and method, so as to avoid the influence of similar interfering light sources in complex environments on the recognition of target mirrors, solve the problem of accurate calculation of the light failure recovery angle when the distance parameters of the target mirror are unknown, and realize autonomous laser tracking light failure recovery in a large range of complex environments.

[0006] The specific technical solution adopted by the present invention to achieve the above-mentioned purpose is:

[0007] 1. A method for recovering from a light failure based on a rotating mirror binocular vision laser tracking light failure recovery system The method for recovering from a light failure comprises the following steps:

[0008] S1) when the laser tracker is interrupted, the infrared beams emitted by the two infrared illumination sources in the binocular vision module are modulated, and then a camera corresponding to each infrared illumination source is used to obtain a target mirror image at the modulation peak and modulation trough of the infrared beam, and a frame difference image of the two target mirror images is obtained. After image processing, the target mirror position on the camera imaging plane is obtained;

[0009] S2) using the pre-constructed corrected virtual imaging plane model, according to the target mirror positions on the two camera imaging planes, obtaining the target mirror position on the virtual imaging plane and the predicted light-break recovery rotation angle; the predicted light-break recovery rotation angle includes a predicted azimuth rotation angle and a predicted pitch rotation angle;

[0010] In step S2, the target mirror positions on the two camera imaging planes are processed by the following formula to obtain the target mirror positions on the virtual imaging plane:

[0011] O X =(P LX +P RX ) / 2

[0012] O Y =(P LY +P RY ) / 2

[0013] In the formula, O X , O Y Respectively represent the X and Y coordinates of the target mirror position on the virtual imaging plane; P LX , P LY Respectively represent the X and Y coordinates of the target mirror position on the first camera imaging plane, P RX , P RY They respectively represent the X and Y coordinates of the target mirror position on the second camera imaging plane.

[0014] In step S2, according to the position of the target mirror on the imaging plane of the two cameras, the predicted light-break recovery rotation angle is obtained by the following formula:

[0015]

[0016] In the formula, θ X Represents the predicted azimuth rotation angle, θ Y represents the predicted pitch rotation angle, P LX, P LY They represent the X and Y coordinates of the target mirror position on the first camera imaging plane, respectively. RX , P RY They represent the X and Y coordinates of the target mirror position on the second camera imaging plane, respectively. X , C Y They represent the X and Y coordinates of the zero point position of the target mirror on the virtual imaging plane, p represents the distance coefficient from the calibrated virtual imaging plane coordinates to the actual measurement space, and f represents the focal length of the camera lens.

[0017] In step S2, in the corrected virtual imaging plane model, the camera visual axis coincides with the laser optical axis; the construction process of the corrected virtual imaging plane model is specifically: establishing a virtual imaging plane model in which the camera visual axis coincides with the laser optical axis; aiming the laser beam at the target target mirror, and calibrating the zero point position of the target target mirror on the virtual imaging plane in the virtual imaging plane model; and correcting the inter-axis angle between the virtual imaging plane coordinates and the two-dimensional turntable coordinates of the laser tracker.

[0018] S3) Output the target mirror position on the virtual imaging plane obtained in step S2 to the pre-constructed error model to obtain the corresponding prediction error, and compensate the predicted light-break recovery rotation angle obtained in step S2 by the prediction error to obtain the compensated light-break recovery rotation angle; the compensated light-break recovery rotation angle includes the compensated azimuth rotation angle and the compensated pitch rotation angle. The step S3 is specifically as follows: input the target mirror position on the virtual imaging plane obtained in step S2 into two linear relationship models in the error model respectively to obtain the corresponding azimuth rotation angle prediction error and pitch rotation angle prediction error; the two linear relationship models are linear relationship models of the target mirror position on the virtual imaging plane and the azimuth rotation angle prediction error and the pitch rotation angle prediction error respectively; then use the azimuth rotation angle prediction error and the pitch rotation angle prediction error to compensate the predicted azimuth rotation angle and the predicted pitch rotation angle respectively, to obtain the compensated azimuth rotation angle and the compensated pitch rotation angle, and combine them to obtain the compensated light-break recovery rotation angle.

[0019] The construction process of the error model is specifically as follows:

[0020] 1) Aim the laser beam at the target mirror and calibrate the zero point position of the target mirror on the virtual imaging plane;

[0021] 2) Control the azimuth motor to step according to a preset step length; after each step, use the angle grating to obtain the actual light-break recovery angle, obtain the target mirror positions on the two camera imaging planes according to step S1, and then obtain the target mirror position and the predicted azimuth rotation angle on the virtual imaging plane according to step S2; obtain the difference between the actual light-break recovery angle and the predicted azimuth rotation angle, and obtain the azimuth rotation angle prediction error corresponding to the current position of the target mirror on the virtual imaging plane;

[0022] 3) A fitting method is used to process the azimuth rotation angle prediction error corresponding to different positions of the target mirror on the virtual imaging plane, and a linear relationship model between the position of the target mirror on the virtual imaging plane and the azimuth rotation angle prediction error is obtained;

[0023] 4) Through the same process as steps 2 to 3, a linear relationship model of the target mirror position and the pitch rotation angle prediction error on the virtual imaging plane is obtained;

[0024] 5) The two linear relationship models obtained in step 3 and step 4 constitute the error model.

[0025] S4) Controlling the rotating mirror laser tracker according to the compensation light-off recovery rotation angle, so that the laser beam is realigned to the target mirror. The step S4 is specifically: driving the azimuth motor to rotate the same angle according to the compensation azimuth rotation angle, and driving the pitch motor to rotate half the angle according to the compensation pitch rotation angle, so that the laser beam can be realigned to the target mirror.

[0026] 2. A rotating mirror binocular vision laser tracking light-break recovery system applied to the above light-break recovery method

[0027] The light-off recovery system includes a rotating mirror laser tracker and a binocular vision module; the binocular vision module includes two recognition modules, each recognition module includes a camera and an infrared illumination light source, and the cameras in the two recognition modules are symmetrically arranged on both sides of the laser optical axis of the rotating mirror laser tracker.

[0028] The rotating mirror laser tracker comprises a laser ranging module, a pitch motor, an azimuth motor, a tracking rotating mirror and a target mirror; the laser ranging module is used to emit and receive a laser beam, an azimuth motor is arranged above the laser ranging module, a pitch motor is installed on the azimuth motor, and the azimuth motor and the pitch motor form a two-dimensional turntable; a tracking rotating mirror is installed on the pitch motor, the laser beam emitted by the laser ranging module is reflected by the tracking rotating mirror and then incident on the target mirror, and the laser beam returned from the target mirror returns to the laser ranging module; two binocular vision modules are symmetrically arranged on both sides of the laser beam emitted by the laser ranging module, and are both connected to the azimuth motor; the infrared illumination light source in each recognition module emits an infrared beam, which is reflected by the tracking rotating mirror and the target mirror in turn, and then returns to the camera in the recognition module along the original optical path and is received by the camera.

[0029] Furthermore, the light failure recovery system further comprises:

[0030] The data processing main control module has a first input end and a second input end which are respectively connected to the two cameras for communication, and is used to receive the peak target mirror image and the trough target mirror image collected by each camera, and obtain the compensation light-break recovery rotation angle after processing using the corrected virtual imaging plane model and the error model;

[0031] The motor drive control module has an input end that is communicatively connected to the first output end of the data processing main control module, and an output end that is electrically connected to the control ends of the pitch motor and the azimuth motor respectively; and is used to receive the compensation post-light failure recovery rotation angle from the data processing main control module, and control the two motors to rotate to corresponding angles according to the compensation post-light failure recovery rotation angle;

[0032] The host computer is used to recover the rotation angle after receiving compensation from the data processing main control module and then cutting off the light.

[0033] The beneficial effects of the present invention are:

[0034] 1. The present invention utilizes two cameras of a binocular vision module to be symmetrically installed on both sides of the laser optical axis, constructs a virtual imaging plane model in which the camera visual axis coincides with the laser optical axis, and calculates the light-break recovery servo angle according to the two-dimensional offset of the target mirror on the virtual imaging plane, thereby avoiding the problem of nonlinear change of the monocular vision light-break recovery servo angle with the distance from the target mirror;

[0035] 2. The present invention calibrates the zero point position of the target ball on the virtual imaging plane, and corrects the angle between the virtual imaging plane coordinates and the coordinate axis of the two-dimensional turntable of the laser tracker, thereby reducing the coaxial installation accuracy requirements of the visual system camera axis;

[0036] 3. The present invention constructs a target mirror position error compensation model on a virtual imaging plane with decoupled measurement distance, compensates for the light-break recovery servo angle error caused by the divergence angle of the illumination light source, and realizes the accurate calculation of the light-break recovery servo angle when the target mirror distance parameters are unknown;

[0037] 4. The present invention uses binocular vision module data for double confirmation, and uses the mean value of binocular vision module data to calculate the rotation angle, thereby improving the target ball recognition accuracy and the accuracy of the light-off recovery servo angle calculation;

[0038] 5. The present invention modulates the light intensity of the infrared illumination source to obtain the modulation trough background image and the modulation peak foreground image, and uses the foreground and background frame difference images to identify the target, thereby avoiding the influence of the background similar high-brightness interference light source on the target recognition.

[0039] In summary, the present invention reduces the coaxial installation accuracy requirements of the visual system of the rotating mirror laser tracker, avoids the influence of similar background high-brightness light sources on the recognition of the target mirror, improves the calculation accuracy of the light failure recovery servo angle, and realizes the autonomous and rapid light failure recovery of the rotating mirror laser tracker in a large range of complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of a rotating mirror binocular vision laser tracking light interruption recovery system involved in this embodiment example is shown.

[0041] Figure 2 The schematic diagram of the rotating mirror binocular vision laser tracking light interruption recovery method involved in this embodiment example is shown.

[0042] Figure 3 The schematic diagram of the calculation principle of the azimuth rotation angle and the pitch rotation angle of the rotating mirror binocular vision laser tracking light interruption recovery method involved in the example of this embodiment is shown.

[0043] Figure 4 The schematic diagram of the angle model between the virtual imaging plane coordinates and the laser tracker two-dimensional turntable coordinate axes involved in this embodiment example is shown.

[0044] In the figure: 1. laser ranging module, 2. pitch motor, 3. azimuth motor, 4. tracking mirror, 5. target mirror, 6. first camera, 7. second camera, 8. first infrared lighting source, 9. second infrared lighting source, 10. data processing main control module, 11. motor drive control module, 12. host computer. DETAILED DESCRIPTION

[0045] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0046] The purpose of the present invention is to provide a rotating mirror binocular vision laser tracking light-break recovery system and method, which utilizes a binocular vision system in which two cameras are symmetrically installed on both sides of the laser optical axis to construct a virtual imaging plane in which the camera visual axis coincides with the laser optical axis, and calculates the light-break recovery servo angle according to the two-dimensional offset of the target mirror on the virtual imaging plane, thereby avoiding the problem of nonlinear change of the monocular vision light-break recovery servo angle with the distance of the target mirror; constructs a target mirror position error compensation model on the virtual imaging plane with decoupled measurement distance, compensates for the light-break recovery servo angle error caused by the divergence angle of the illumination light source, and realizes the accurate calculation of the light-break recovery servo angle when the distance parameters of the target mirror are unknown; by modulating the light intensity of the infrared light source, the modulation wave valley background image and the modulation wave peak foreground image are obtained, and the target mirror is identified by using the foreground and background frame difference images, thereby avoiding the influence of the background similar high-brightness interference light source on the target mirror identification, thereby realizing fast and accurate autonomous laser tracking light-break recovery.

[0047] The first aspect of the present invention provides a system and method for recovering the broken light based on a rotating mirror binocular vision laser tracking. The method of the present invention reduces the coaxial installation accuracy requirements of the rotating mirror laser tracker vision system, avoids the influence of the background-similar high-brightness light source on the target mirror recognition, improves the calculation accuracy of the broken light recovery servo angle, and realizes the autonomous and rapid broken light recovery of the rotating mirror laser tracker in a large range of complex environments.

[0048] The method of the present invention comprises the following steps:

[0049] S1) When the rotating mirror laser tracker is interrupted, the infrared light beams emitted by the two infrared illumination sources in the binocular vision module are modulated, and then a camera corresponding to each infrared illumination source is used to obtain a target mirror image at the modulation peak and modulation trough of the infrared light beam, and the two target mirror images are subtracted to obtain a frame difference image with only the infrared light source illumination information, and the frame difference image is processed to obtain the target target mirror position on the camera imaging plane.

[0050] In the rotating mirror binocular vision laser tracking light interruption recovery system, in this step, the beam transmission path is specifically as follows: the first infrared illumination light source 8 emits the first infrared beam, the second infrared illumination light source 9 emits the second infrared beam, and after the target target mirror 5 receives the first infrared beam and the second infrared beam, the first infrared beam and the second infrared beam are reflected along the original path, and the first camera 6 and the second camera 7 respectively receive the first infrared beam and the second infrared beam reflected by the target target mirror 5.

[0051] In some embodiments, the first infrared illumination light source 8 and the second infrared illumination light source 9 are modulated simultaneously. After the laser tracker locks the target mirror, the two infrared light sources stop modulating simultaneously.

[0052] In some implementations, the process of performing image processing on the frame difference image is specifically: identifying the target mirror 5 and acquiring its position information.

[0053] In some embodiments, the infrared light beam is modulated by modulating the light intensity of the infrared light beam.

[0054] In some embodiments, the intensity of the infrared light beam is modulated by periodically changing the intensity of the infrared light beam (eg, a sine wave or a square wave).

[0055] In some embodiments, the intensity of the infrared light beam is periodically changed by using a PWM controller to adjust the driving current of the infrared illumination light source.

[0056] S2) Using the pre-constructed corrected virtual imaging plane model, the target target mirror position on the virtual imaging plane is obtained according to the target target mirror position on the two camera imaging planes, and then the predicted light-break recovery rotation angle is obtained according to the two-dimensional offset of the target target mirror position on the virtual imaging plane.

[0057] The two-dimensional offset of the target mirror position on the virtual imaging plane refers to the X-direction and Y-direction offset of the target mirror position relative to the zero point position of the target mirror on the virtual imaging plane. The target mirror position on the two camera imaging planes is processed by the following formula to obtain the target mirror position on the virtual imaging plane:

[0058] O X =(P LX +P RX ) / 2

[0059] O Y =(P LY +P RY ) / 2

[0060] In the formula, O X , O Y Respectively represent the X and Y coordinates of the target mirror position on the virtual imaging plane; P LX , P LY Respectively represent the X and Y coordinates of the target mirror position on the first camera imaging plane, P RX , P RY They respectively represent the X and Y coordinates of the target mirror position on the second camera imaging plane.

[0061] The predicted rotation angle after light failure includes the predicted azimuth rotation angle and the predicted pitch rotation angle. The specific acquisition process is as follows: according to the position of the target mirror on the imaging plane of the two cameras, the predicted rotation angle after light failure is obtained by the following formula:

[0062]

[0063] In the formula, θ X Represents the predicted azimuth rotation angle, θ Y represents the predicted pitch rotation angle, P LX , P LY They represent the X and Y coordinates of the target mirror position on the first camera imaging plane, respectively. RX , P RY They represent the X and Y coordinates of the target mirror position on the second camera imaging plane, respectively. X , C Y They represent the X and Y coordinates of the zero point position of the target mirror on the virtual imaging plane, respectively; p represents the distance coefficient from the calibrated virtual imaging plane coordinates to the actual measurement space; and f represents the focal length of the camera lens (the focal lengths of the lenses of the two cameras are the same).

[0064] In step S2, in the virtual imaging plane model and the modified virtual imaging plane model, the camera visual axis coincides with the laser optical axis. The construction process of the modified virtual imaging plane model is specifically as follows:

[0065] D1) According to the binocular vision module of the rotating mirror binocular vision laser tracking light-break recovery system, two cameras are symmetrically installed on both sides of the laser optical axis, and a virtual imaging plane model is established in which the camera visual axis coincides with the laser optical axis;

[0066] D2) Aim the laser beam at the target mirror 5, calibrate the zero point position of the target mirror on the virtual imaging plane in the virtual imaging plane model; and correct the angle between the virtual imaging plane coordinates and the two-dimensional turntable coordinates of the laser tracker.

[0067] Step D2 is specifically as follows:

[0068] D21) Calibrate the zero point position of the target mirror: aim the laser beam at the target mirror 5, use the binocular vision module to collect the target mirror image, obtain the target mirror position on the two camera imaging planes according to the target mirror image, and input it into the following formula, and use the obtained target mirror position on the virtual imaging plane as the zero point position of the target mirror:

[0069] O X =(P LX +P RX ) / 2

[0070] O Y =(P LY +P RY ) / 2

[0071] In the formula, O X , O Y Respectively represent the X and Y coordinates of the target mirror position on the virtual imaging plane; P LX , P LYRespectively represent the X and Y coordinates of the target mirror position on the first camera imaging plane, P RX , P RY They respectively represent the X and Y coordinates of the target mirror position on the second camera imaging plane.

[0072] D22) Correct the inter-axis angle between the virtual imaging plane coordinates and the laser tracker two-dimensional turntable coordinates: Control the azimuth motor 3 in the two-dimensional turntable to rotate around the horizontal axis, and obtain the change in the target mirror position along the Y direction on the virtual imaging plane under different target mirror position X direction coordinates, and then correct the inter-axis angle in the X direction according to the change, so that when the azimuth motor 3 rotates, the target mirror position coordinate in the Y direction remains unchanged. In the same way, control the pitch motor 2 in the two-dimensional turntable to rotate around the vertical axis, and obtain the change in the target mirror position in the virtual imaging plane model along the X direction under different target mirror position Y direction coordinates, and then correct the inter-axis angle in the Y direction, so that when the pitch motor 2 rotates, the target mirror position coordinate in the X direction remains unchanged.

[0073] Among them, the change in the target target mirror position along the Y direction on the virtual imaging plane is obtained through the following process: after each rotation, according to the same method as step D21, the current position of the target target mirror on the virtual imaging plane is obtained according to the target mirror image collected by the binocular vision module, and then combined with the zero point position of the target target mirror, the change in the target target mirror position along the Y direction on the virtual imaging plane under the X-direction coordinate of the current target target mirror position is obtained.

[0074] S3) Output the target mirror position on the virtual imaging plane obtained in step S2 to the pre-built error model to obtain the corresponding prediction error, and compensate the predicted light-break recovery rotation angle by the prediction error to obtain the compensated light-break recovery rotation angle; this step is used to compensate for the light-break recovery rotation angle prediction error caused by the divergence angle of the illumination light. The prediction error includes the azimuth rotation angle prediction error and the pitch rotation angle prediction error. The compensated light-break recovery rotation angle includes the compensated azimuth rotation angle and the compensated pitch rotation angle.

[0075] In step S3, the compensation process is specifically as follows: the error model includes a linear relationship model between the target target mirror position on the virtual imaging plane and the azimuth rotation angle prediction error, and a linear relationship model between the target target mirror position on the virtual imaging plane and the pitch rotation angle prediction error; the target target mirror position on the virtual imaging plane obtained in step S2 is respectively input into the two linear relationship models to obtain the corresponding azimuth rotation angle prediction error and pitch rotation angle prediction error; the azimuth rotation angle prediction error and the pitch rotation angle prediction error are used to compensate the predicted azimuth rotation angle and the predicted pitch rotation angle respectively to obtain the compensated azimuth rotation angle and the compensated pitch rotation angle, which are combined to obtain the compensated light-off recovery rotation angle.

[0076] The specific process of constructing the error model is as follows:

[0077] 1) Aim the laser beam at the target mirror 5, and calibrate the zero point position of the target mirror on the virtual imaging plane;

[0078] 2) Control the azimuth motor 3 to step according to a preset step length; after each step, use the angle grating to obtain the actual light-break recovery angle, obtain the target mirror positions on the two camera imaging planes through step S1, and then obtain the target mirror position and the predicted azimuth rotation angle on the virtual imaging plane through step S2; obtain the difference between the actual light-break recovery angle and the predicted azimuth rotation angle, and obtain the azimuth rotation angle prediction error corresponding to the current position of the target mirror on the virtual imaging plane;

[0079] 3) A fitting method is used to process the azimuth rotation angle prediction error corresponding to different positions of the target mirror on the virtual imaging plane, and a linear relationship model between the position of the target mirror on the virtual imaging plane and the azimuth rotation angle prediction error is obtained;

[0080] 4) Through the same process as steps 2 to 3, a linear relationship model of the target mirror position and the pitch rotation angle prediction error on the virtual imaging plane is obtained;

[0081] 5) The two linear relationship models obtained in step 3 and step 4 constitute the error model.

[0082] S4) Control the rotating mirror laser tracker according to the compensated light-off recovery rotation angle, so that the laser beam is realigned to the target mirror 5. Step S4 is specifically: according to the compensated azimuth rotation angle in the compensated light-off recovery rotation angle, drive the azimuth motor 3 in the two-dimensional turntable to rotate the same angle, according to the compensated pitch rotation angle in the compensated light-off recovery rotation angle, drive the pitch motor 2 in the two-dimensional turntable to rotate half the angle, and finally realign the laser beam to the target mirror 5.

[0083] The second aspect of the present invention provides a rotating mirror binocular vision laser tracking light-break recovery system applied to the above-mentioned light-break recovery method. It includes a rotating mirror laser tracker and a binocular vision module; the binocular vision module includes two recognition modules, each recognition module includes a camera and an infrared illumination light source, and the cameras in the two recognition modules are symmetrically arranged on both sides of the laser optical axis of the rotating mirror laser tracker. The camera visual axis and the laser optical axis are reflected to the working area through the rotating mirror of the tracker.

[0084] The rotating mirror laser tracker comprises a laser ranging module 1, a pitch motor 2, an azimuth motor 3, a tracking rotating mirror 4 and a target mirror 5; the laser ranging module 1 is used to emit and receive laser beams, an azimuth motor 3 is arranged above the laser ranging module 1, a pitch motor 2 is installed on the azimuth motor 3 and can drive the pitch motor 2 to rotate around a horizontal axis, and the azimuth motor 3 and the pitch motor 2 form a two-dimensional turntable; the pitch motor 2 is installed on the tracking rotating mirror 4 and can drive the tracking rotating mirror 4 to rotate synchronously in pitch, that is, to rotate around a vertical axis, and the laser beam emitted by the laser ranging module 1 is transmitted to the target mirror 5. The light beam is reflected by the tracking mirror 4 and incident on the target mirror 5. The laser beam returned from the target mirror 5 returns to the laser ranging module 1 along the original optical path; the infrared illumination light source in each recognition module emits an infrared light beam, which is reflected by the tracking mirror 4 and the target mirror 5 in turn, and then returns to the acquisition end of the camera in the recognition module to which it belongs along the original optical path and is received by the camera; the two binocular vision modules are symmetrically arranged on both sides of the laser beam emitted by the laser ranging module 1, and the two binocular vision modules are connected to the output end of the azimuth motor 3 and rotate with the azimuth motor 3.

[0085] In addition, the light failure recovery system also includes:

[0086] The data processing main control module 10, the first input end and the second input end are respectively connected to the two cameras for communication, and is used to receive the peak target mirror image and the trough target mirror image collected by each camera, and obtain the compensation after the light failure recovery rotation angle after processing using the corrected virtual imaging plane model and the error model;

[0087] The motor drive control module 11 has an input end that is communicatively connected to the first output end of the data processing main control module 10, and an output end that is electrically connected to the control ends of the pitch motor 2 and the azimuth motor 3 respectively; it is used to receive the compensation post-light failure recovery rotation angle from the data processing main control module 10, and control the two motors to rotate to corresponding angles according to the compensation post-light failure recovery rotation angle;

[0088] The host computer 12 is used to receive the compensation light-off recovery rotation angle from the data processing main control module 10 .

[0089] Furthermore, the host computer 12 can be used to perform operations such as storage, visualization, and data processing on the compensation light failure recovery rotation angle.

[0090] The specific embodiments of the present invention are as follows:

[0091] like Figure 1 As shown, the rotating mirror binocular vision laser tracking light interruption recovery system used in this embodiment includes a laser ranging module 1, a pitch motor 2, an azimuth motor 3, a tracking rotating mirror 4, a target mirror 5, a first camera 6, a second camera 7, a first infrared lighting source 8, a second infrared lighting source 9, a data processing main control module 10, a motor drive control module 11, and a host computer 12.

[0092] The laser ranging module 1, the pitch motor 2, the azimuth motor 3, the tracking mirror 4, and the target mirror 5 constitute a rotating mirror laser tracker. The laser ranging module 1 is fixed and does not rotate with the two-dimensional turntable composed of the pitch motor 2 and the azimuth motor 3. The tracking mirror 4 is fixed on the pitch motor 2 and rotates with the pitch motor 2. The laser beam emitted by the laser ranging module 1 is reflected by the tracking mirror 4 and then incident on the target mirror 5. The laser beam returned from the target mirror 5 returns to the laser ranging module 1 to achieve target tracking and spatial coordinate measurement.

[0093] The first camera 6, the second camera 7, the first infrared illumination light source 8, and the second infrared illumination light source 9 are fixed on the azimuth motor 3 and rotate with the azimuth motor 3. The first camera 6 and the second camera 7 are symmetrically mounted on both sides of the laser beam emitted by the laser ranging module 1, and the first infrared illumination light source 8 and the second infrared illumination light source 9 are respectively mounted around the first camera 6 and the second camera 7, so that the first infrared beam and the second infrared beam reflected back by the target mirror 5 reflect a uniform target mirror spot profile.

[0094] The target mirror images captured by the first camera 6 and the second camera 7 are processed by the data processing main control module 10 to obtain the compensated azimuth rotation angle θ X ' and the pitch rotation angle θ after compensation Y ', and sent to the motor drive control module 11. The motor drive control module 11 is used to control the azimuth motor 3 to rotate θ X 'angle, control the pitch motor 2 to rotate θ Y ' / 2 angle, so that the laser tracker laser beam is re-aimed at the target mirror 5, and the light failure recovery is completed. The data processing main control module 10 exchanges data with the host computer 12.

[0095] like Figure 2 As shown, the rotating mirror binocular vision light-break recovery model used in this embodiment includes the target mirror 5, the laser optical axis M, the first camera visual axis M1, the second camera visual axis M2, the X direction X, the Y direction Y, the target mirror zero point position C on the virtual imaging plane, the target mirror current position O on the virtual imaging plane, and the target mirror current position O on the first camera imaging plane. 1 , the current position O of the target mirror on the second camera imaging plane 2 、First camera imaging center P1 , the second camera imaging center P 2 , the midpoint R of the line connecting the imaging centers of the first camera and the second camera.

[0096] Adjust the first camera visual axis M1 and the second camera visual axis M2 to be parallel to the laser optical axis M and symmetrical with respect to the laser optical axis M so that the imaging center P of the first camera 1 and the second camera imaging center P 2 The midpoint R of the connecting line is located on the optical axis M of the laser tracker. When the laser beam M is aimed at the target mirror 5, the zero point position C of the target mirror on the virtual imaging plane is calibrated. When the laser beam M deviates from the target mirror 5, the current position O of the target mirror on the first camera imaging plane can be obtained on the first camera 6. 1 , the current position O of the target mirror on the second camera imaging plane can be obtained on the second camera 7 2 The current position O of the target mirror on the virtual imaging plane can be obtained by the current position O of the target mirror on the imaging plane of the first camera. 1 and the current position O of the target mirror on the imaging plane of the second camera 2 Based on the geometric relationship, the specific formula is:

[0097] O X =(P LX +P RX ) / 2

[0098] O Y =(P LY +P RY ) / 2

[0099] Among them, O X and O Y are the X-direction and Y-direction coordinates of the target mirror position O on the virtual imaging plane. LX and P LY are the current position O of the target mirror on the imaging plane of the first camera respectively. 1 The X and Y coordinates of RX and P RY are the current position O of the target mirror on the imaging plane of the second camera respectively. 2 The X and Y coordinates of the

[0100] like Figure 3 As shown, in the rotating mirror binocular vision laser tracking light-off recovery method used in this embodiment, the azimuth rotation angle and pitch rotation angle prediction model includes the target mirror 5, the laser optical axis M, the zero point position C of the target mirror on the virtual imaging plane, the current position O of the target mirror on the virtual imaging plane, the midpoint R of the line connecting the imaging centers of the first camera and the second camera, and the azimuth rotation angle θ X , pitch rotation angle θ Y, X-direction target mirror offset d X , Y direction target mirror offset d Y .

[0101] According to the current position O of the target mirror on the virtual imaging plane and the zero point position C of the target mirror on the virtual imaging plane, the target mirror offset d in the X direction is obtained respectively. X and the target mirror offset d in the Y direction Y , and then calculate the predicted azimuth rotation angle θ X and the predicted pitch rotation angle θ Y , the specific formula is:

[0102] The predicted azimuth rotation angle is:

[0103] The predicted pitch rotation angle is:

[0104] Among them, C X and C Y They represent the X-direction and Y-direction coordinates of the zero point position C of the target mirror on the virtual imaging plane, respectively; p is the distance coefficient from the calibrated virtual imaging plane coordinate to the actual measurement space; and f is the focal length of the first camera and the second camera.

[0105] like Figure 4 As shown, in this embodiment, the angle correction model between the virtual imaging plane coordinates and the laser tracker two-dimensional turntable coordinate axis includes the target mirror 5, the virtual imaging plane X direction D XR , virtual imaging plane Y direction D YR , the target mirror position on the virtual imaging plane is D YR The deviation D from the zero position of the target mirror in the direction and the target mirror position O after angle compensation C 、Axis angle a E , the midpoint R of the line connecting the imaging centers of the first camera and the second camera.

[0106] In order to correct the axis angle a E The light spot position deviation caused by the control of the azimuth motor 3 makes the imaging point of the target mirror 5 move from point C along the X direction until the imaging point is located at the edge of the image. X Coordinates and zero point C X Coordinates in the Y direction of the virtual imaging plane D YR The deviation on O is D; D varies with O X Linear change of coordinates:

[0107] D=K D ×(O X -C X )

[0108] Among them, KD is the coefficient. Then when moving in the X direction, the deviation △Y of the imaging point in the Y direction can be calculated as:

[0109] ΔY=cosa E ×K D ×(O X -C X )

[0110] Angle between axes a E is a small enough angle, then cos a E can be considered as 1. Then the Y coordinate after compensation is O YC :

[0111] O YC =O Y +ΔY

[0112] Similarly, the pitch motor 2 is controlled to rotate to compensate for the deviation of the target mirror position in the X direction to obtain O XC .

[0113] The principle of compensating the prediction error model of the light-break recovery rotation angle caused by the divergence angle of the illumination light in this embodiment is specifically as follows: when the irradiation angle of the infrared light beam to the target mirror 5 is zero, the target mirror position O on the virtual imaging plane is different from the actual imaging point O of the target mirror. R When the irradiation angle of the infrared beam to the target mirror 5 is not zero, the target mirror position O on the virtual imaging plane is aligned with the actual imaging point O of the target mirror. R There is a deviation between I , and the deviation d I It is positively correlated with the irradiation angle I and has nothing to do with the distance from the target mirror 5. Deviation d I This results in an error between the predicted rotation angle of the light-off recovery and the actual angle of the light-off recovery θ. The error is linearly related to the target mirror position O on the virtual imaging plane, so an error model is established for compensation.

[0114] In this embodiment, the specific implementation steps of the rotating mirror binocular vision laser tracking light interruption recovery method are as follows:

[0115] 1) Establish a virtual imaging plane model in which the camera visual axis coincides with the laser optical axis: In the rotating mirror binocular vision laser tracking light interruption recovery system, the two cameras of the binocular vision module are symmetrically installed on both sides of the laser optical axis M. The first camera visual axis M1 and the second camera visual axis M2 after reflection by the tracking rotating mirror 4 are symmetrical with respect to the center of the reflected laser optical axis M, and the intersection of the first camera visual axis M1 and the second camera visual axis M2 is located at infinity; Based on the camera calibration and imaging principles, the geometric relationship between the two camera visual axes and the laser optical axis M is used to construct a virtual imaging plane model in which the camera visual axis coincides with the laser optical axis.

[0116] 2) Calibrate the zero point position of the target ball on the virtual imaging plane, and correct the angle between the virtual imaging plane coordinates and the two-dimensional turntable coordinates of the laser tracker;

[0117] In step 2), the correction of the angle between the virtual imaging plane coordinates and the laser tracker two-dimensional turntable coordinate axis is as follows: E The light spot position deviation caused by the control of the azimuth motor 3 makes the imaging point of the target mirror 5 move from point C along the X direction until the imaging point is located at the edge of the image. X Coordinates and zero point C X Coordinates in the Y direction of the virtual imaging plane D YR The deviation on O is D; D varies with O X Linear change of coordinates:

[0118] D=K D ×(O X -C X )

[0119] In the formula, K D is the coefficient. Then when moving in the X direction, the deviation △Y of the imaging point in the Y direction can be calculated as:

[0120] ΔY=cosa E ×K D ×(O X -C X )

[0121] Angle between axes a E is a small enough angle, then cos a E can be considered as 1. Then the Y coordinate after compensation is O YC :

[0122] O YC =O Y +ΔY

[0123] Similarly, the pitch motor 2 is controlled to rotate to compensate for the deviation of the target mirror position in the X direction to obtain O XC .

[0124] 3) By modulating the infrared light source of the binocular vision system, the target mirror image is obtained at the modulation peak and modulation trough respectively, and the two images are subtracted to obtain a frame difference image with only the infrared light source illumination information, and the frame difference image is processed to identify the target mirror and obtain its position information;

[0125] In step 3), the binocular vision module identifies the target mirror and obtains the position information as follows:

[0126] The system emits a first infrared light beam through a first infrared illumination light source 8, and emits a second infrared light beam through a second infrared illumination light source 9. After receiving the first infrared light beam and the second infrared light beam, the target mirror 5 reflects the first infrared light beam and the second infrared light beam along the original path, and the first camera 6 and the second camera 7 of the binocular vision module receive the reflected first infrared light beam and the second infrared light beam respectively.

[0127] The first infrared illumination light source 8 and the second infrared illumination light source 9 are modulated respectively, and the first camera 6 obtains the foreground image and the background image respectively when the modulation wave is at the peak and the trough, and the foreground image and the background image are subtracted to obtain the frame difference image, and the frame difference image is processed to obtain the target mirror position O carried by the first infrared beam. 1 Similarly, the second camera frame difference image is processed to obtain the target mirror position O carried by the second infrared beam 2 ;

[0128] 4) According to the target mirror positions on the imaging planes of the two cameras, the target mirror positions on the virtual imaging plane are obtained, and the predicted light-break recovery rotation angle is calculated according to its two-dimensional offset, and the predicted light-break recovery rotation angle is compensated in combination with the error model;

[0129] In step 4), the process of obtaining the light-off recovery rotation angle is as follows:

[0130] On the virtual imaging plane where the camera visual axis coincides with the laser optical axis, the X-direction coordinate of the target mirror position is (P LX +P RX ) / 2, the Y coordinate of the target mirror position is (P LY +P RY ) / 2, where P LX and P RX represents the X-direction coordinates of the target mirror positions of the first camera 6 and the second camera 7 imaging planes, respectively, LY and P RY Respectively represent the Y-direction coordinates of the target mirror positions of the first camera and the second camera imaging planes;

[0131] The predicted azimuth rotation angle is:

[0132] The predicted pitch rotation angle is:

[0133] Among them, C X and C Y They represent the X-direction coordinate and Y-direction coordinate of the zero point position C of the target mirror on the virtual imaging plane, respectively; p is the distance coefficient from the calibrated virtual imaging plane coordinate to the actual measurement space; and f is the focal length of the first camera and the second camera;

[0134] In step 4), the compensation of the prediction error according to the imaging plane target mirror position error model is as follows:

[0135] When the laser beam is aimed at the target mirror 5, the zero point position C of the target mirror on the virtual imaging plane is calibrated, and the precise angle reading of the angle grating of the laser tracker is used to control the precise angle stepping of the motor, obtain the error between the predicted light-off recovery rotation angle and the stepping angle, and construct a linear relationship model between the target mirror position and the predicted error, and compensate for the predicted error of the light-off recovery rotation angle caused by the divergence angle of the illumination light;

[0136] 5) Drive the corresponding rotating mechanism to rotate the corresponding light-off recovery angle so that the laser beam is realigned to the target mirror 5.

[0137] In step 5), the corresponding rotating mechanism is driven to rotate the corresponding light-off recovery angle as follows: drive the azimuth motor 3 to rotate θ X ' angle, since the pitch rotation angle of the laser beam of the rotating mirror laser tracker is twice the rotation angle of the pitch motor, the pitch motor is driven to rotate θ Y ' / 2 angle, so that the laser beam pitches and rotates θ Y ' angle, and finally the laser beam is re-aimed at the target mirror. X '、θ Y 'respectively represent the azimuth rotation angle and the pitch rotation angle after compensation.

[0138] The above specific implementation modes are used to explain the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A method for light failure recovery based on a rotating mirror binocular vision laser tracking light failure recovery system, characterized in that: The following steps are involved: S1) when the laser tracker is interrupted, the infrared beams emitted by the two infrared illumination sources in the binocular vision module are modulated, and then a camera corresponding to each infrared illumination source is used to obtain a target mirror image at the modulation peak and modulation trough of the infrared beam, and a frame difference image of the two target mirror images is obtained. After image processing, the target mirror position on the camera imaging plane is obtained; S2) using the pre-constructed corrected virtual imaging plane model, according to the target mirror positions on the two camera imaging planes, obtaining the target mirror position on the virtual imaging plane and the predicted light-break recovery rotation angle; the predicted light-break recovery rotation angle includes a predicted azimuth rotation angle and a predicted pitch rotation angle; S3) outputting the target mirror position on the virtual imaging plane obtained in step S2 to a pre-constructed error model to obtain a corresponding prediction error, and compensating the predicted light-break recovery rotation angle obtained in step S2 by the prediction error to obtain a compensated light-break recovery rotation angle; The compensation post-break light recovery rotation angle includes a compensation post-azimuth rotation angle and a compensation post-pitch rotation angle; S4) controlling the rotating mirror laser tracker according to the rotation angle of the light failure recovery after compensation, so as to realign the laser beam to the target mirror (5).

2. The light failure recovery method according to claim 1, characterized in that: In step S2, the target mirror positions on the two camera imaging planes are processed by the following formula to obtain the target mirror positions on the virtual imaging plane: O X =(P LX +P RX ) / 2 O Y =(P LY +P RY ) / 2 In the formula, O X , O Y Respectively represent the X and Y coordinates of the target mirror position on the virtual imaging plane; P LX , P LY Respectively represent the X and Y coordinates of the target mirror position on the first camera imaging plane, P RX , P RY They respectively represent the X and Y coordinates of the target mirror position on the second camera imaging plane.

3. The light failure recovery method according to claim 1, characterized in that: In step S2, the predicted light-off recovery rotation angle is obtained by the following formula: In the formula, θ X Represents the predicted azimuth rotation angle, θ Y represents the predicted pitch rotation angle, P LX , P LY They represent the X and Y coordinates of the target mirror position on the first camera imaging plane, respectively. RX , P RY They represent the X and Y coordinates of the target mirror position on the second camera imaging plane, respectively. X , C Y They represent the X and Y coordinates of the zero point position of the target mirror on the virtual imaging plane, p represents the distance coefficient from the calibrated virtual imaging plane coordinates to the actual measurement space, and f represents the focal length of the camera lens.

4. The light failure recovery method according to claim 1, characterized in that: In the step S2, in the corrected virtual imaging plane model, the camera visual axis coincides with the laser optical axis; the construction process of the corrected virtual imaging plane model is specifically as follows: establishing a virtual imaging plane model in which the camera visual axis coincides with the laser optical axis; aiming the laser beam at the target mirror (5), calibrating the zero point position of the target mirror on the virtual imaging plane in the virtual imaging plane model; and correcting the inter-axis angle between the virtual imaging plane coordinates and the two-dimensional turntable coordinates of the laser tracker.

5. The light failure recovery method according to claim 1, characterized in that: The step S3 is specifically as follows: inputting the target mirror position on the virtual imaging plane obtained in step S2 into two linear relationship models in the error model respectively, and obtaining the corresponding azimuth rotation angle prediction error and pitch rotation angle prediction error; the two linear relationship models are linear relationship models of the target mirror position on the virtual imaging plane and the azimuth rotation angle prediction error and the pitch rotation angle prediction error respectively; The predicted azimuth rotation angle and the predicted pitch rotation angle are compensated using the azimuth rotation angle prediction error and the pitch rotation angle prediction error respectively to obtain the compensated azimuth rotation angle and the compensated pitch rotation angle, which are combined to obtain the compensated light-off recovery rotation angle.

6. The light failure recovery method according to claim 1, characterized in that: The construction process of the error model is specifically as follows: 1) aligning the laser beam with the target mirror (5), and calibrating the zero point position of the target mirror on the virtual imaging plane; 2) controlling the azimuth motor (3) to step according to a preset step length; After each step, the actual light-break recovery angle is obtained using the angle grating, the target mirror positions on the imaging planes of the two cameras are obtained according to step S1, and then the target mirror positions and predicted azimuth rotation angles on the virtual imaging plane are obtained according to step S2; Obtain the difference between the actual light-break recovery angle and the predicted azimuth rotation angle, and obtain the azimuth rotation angle prediction error corresponding to the current position of the target mirror on the virtual imaging plane; 3) A fitting method is used to process the azimuth rotation angle prediction error corresponding to different positions of the target mirror on the virtual imaging plane, and a linear relationship model between the position of the target mirror on the virtual imaging plane and the azimuth rotation angle prediction error is obtained; 4) Through the same process as steps 2 to 3, a linear relationship model of the target mirror position and the pitch rotation angle prediction error on the virtual imaging plane is obtained; 5) The two linear relationship models obtained in step 3 and step 4 constitute the error model.

7. The light failure recovery method according to claim 1, characterized in that: The step S4 specifically comprises: driving the azimuth motor (3) to rotate by the same angle according to the compensated azimuth rotation angle, and driving the pitch motor (2) to rotate by half the angle according to the compensated pitch rotation angle, so that the laser beam can be realigned with the target mirror (5).

8. A rotating mirror binocular vision laser tracking light failure recovery system applied to the light failure recovery method according to any one of claims 1 to 7, characterized in that: It includes a rotating mirror laser tracker and a binocular vision module; the binocular vision module includes two recognition modules, each recognition module includes a camera and an infrared lighting source, and the cameras in the two recognition modules are symmetrically arranged on both sides of the laser optical axis of the rotating mirror laser tracker.

9. The rotating mirror binocular vision laser tracking light-off recovery system according to claim 8, characterized in that: The rotating mirror laser tracker comprises a laser ranging module (1), a pitch motor (2), an azimuth motor (3), a tracking rotating mirror (4) and a target mirror (5); the laser ranging module (1) is used to emit and receive laser beams; an azimuth motor (3) is arranged above the laser ranging module (1); a pitch motor (2) is mounted on the azimuth motor (3); the azimuth motor (3) and the pitch motor (2) form a two-dimensional turntable; a tracking rotating mirror (4) is mounted on the pitch motor (2); the laser ranging module (1) is used to emit and receive laser beams; an azimuth motor (3) is arranged above the laser ranging module (1); a pitch motor (2) is mounted on the azimuth motor (3); the azimuth motor (3) and the pitch motor (2) form a two-dimensional turntable; a tracking rotating mirror (4) is mounted on the pitch motor (2); The laser beam emitted by the distance module (1) is reflected by the tracking mirror (4) and then incident on the target mirror (5), and the laser beam returned from the target mirror (5) returns to the laser distance measurement module (1); two binocular vision modules are symmetrically arranged on both sides of the laser beam emitted by the laser distance measurement module (1), and are both connected to the azimuth motor (3); the infrared illumination light source in each recognition module emits an infrared beam, which is reflected by the tracking mirror (4) and the target mirror (5) in turn, and then returns to the camera in the recognition module along the original optical path and is received by the camera.

10. The rotating mirror binocular vision laser tracking light-off recovery system according to claim 9, characterized in that: Also includes: A data processing main control module (10), wherein a first input end and a second input end are respectively connected to the two cameras for communication, and is used to receive a peak target mirror image and a trough target mirror image collected by each camera, and obtain a compensated light-break recovery rotation angle after processing using a corrected virtual imaging plane model and an error model; A motor drive control module (11), the input end of which is communicatively connected to the first output end of the data processing main control module (10), and the output end of which is electrically connected to the control ends of the pitch motor (2) and the azimuth motor (3); Used to receive the compensation light failure recovery rotation angle from the data processing main control module (10), and control the two motors to rotate corresponding angles according to the compensation light failure recovery rotation angle; The upper computer (12) is used for receiving the compensation light-off recovery rotation angle from the data processing main control module (10).

Images