A coherent imaging device and method

Abstract

This invention provides a coherent imaging device and method, belonging to the field of photoelectric detection and three-dimensional imaging technology. The device includes: an FMCW laser outputting swept light; a beam splitter dividing the swept light into a measurement branch light, a local oscillator branch light, and a reference branch sampling light; a measurement interferometric branch outputting a measurement beat frequency signal; a reference interferometric branch outputting a reference beat frequency signal using an optical frequency comb; a digital processing module compensating the measurement beat frequency signal based on the reference beat frequency signal and generating a scanner drift observation; using the compensation result and the drift observation to jointly estimate geometric parameter errors and phase drift errors; performing geometric mapping correction and fine phase compensation based on the estimation results; and outputting a coherent three-dimensional point cloud or a coherent detection image. Using this invention can improve the phase consistency and three-dimensional point cloud quality of coherent imaging.

Description

Technical Field

[0001] This invention relates to the field of photoelectric detection and three-dimensional imaging technology, specifically to a coherent imaging device and method. Background Technology

[0002] Frequency-modulated continuous-wave (FMCW) laser coherent imaging technology acquires the target's distance, velocity, and phase information simultaneously by transmitting a linearly modulated continuous-wave laser and coherently mixing it with a local oscillator. This type of system relies on phase consistency across time and scanning positions for coherent accumulation, thereby achieving high-resolution 3D point cloud reconstruction and image focusing. Existing FMCW coherent imaging systems typically employ the following methods: first, geometrically correcting the angle and pose of the scanner (such as a galvanometer or rotary stage) using an encoder or inertial measurement unit; second, post-processing compensation for the laser's sweep frequency nonlinearity using an auxiliary interferometer or numerical autofocusing method; and third, extracting phase noise based on a fixed-reference interferometer to compensate for the phase of the echo signal.

[0003] In the process of realizing this invention, the inventors discovered that existing methods have at least the following drawbacks: First, sweep frequency nonlinearity, laser phase noise, and common optical path disturbances (such as temperature drift and vibration) cause echo phase jitter over time, accumulating deviations between different scanning positions. Existing post-processing compensation methods struggle to measure common path phase noise in real time, leading to coherent accumulation mismatch and image defocusing. Second, scanner angular bias, installation errors, and pose disturbances cause inconsistencies between the propagation distance model and the actual geometric relationship. Furthermore, this geometric error is coupled with the phase error caused by slowly varying environmental drift. Existing methods struggle to effectively distinguish and correct these two errors separately, easily getting trapped in local optima, resulting in increased sidelobes and enhanced artifacts. Third, existing solutions mostly employ open-loop post-processing, with sweep frequency linearization relying on offline calibration. Scanner drift lacks an effective closed-loop feedback mechanism, leading to a significant performance degradation after long-term system operation, failing to meet the requirements for high-quality coherent imaging and point cloud stability under complex operating conditions.

[0004] Therefore, how to measure the phase noise of the common path in real time, compensate for the sweep frequency nonlinearity, phase noise, geometric error and slow drift, and output high-quality coherent 3D point cloud or detection image is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] In view of this, it is necessary to provide a coherent imaging device and method to solve the technical problems of poor coherent imaging quality caused by the inability of existing methods to measure common path phase noise in real time and the difficulty in decoupling and correcting geometric errors and slow phase drift.

[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a coherent imaging apparatus, comprising: FMCW laser, outputting frequency-modulated continuous wave swept light; A beam splitter, connected to the FMCW laser, splits the swept beam into a measurement branch beam, a local oscillator branch beam, and a reference branch sampling beam. The measurement interferometer branch receives the light from the measurement branch and the local oscillator branch, detects the target, and outputs a measurement beat frequency signal; The reference interference branch receives the sampled light from the reference branch and the reference light provided by the optical frequency comb, and outputs a reference beat frequency signal. The digital processing module is connected to the measurement interferometric branch and the reference interferometric branch respectively. It is used to compensate the measurement beat frequency signal based on the reference beat frequency signal and generate scanner drift observations. It uses the compensation results and the drift observations to estimate the geometric parameter error and phase drift error, and performs geometric mapping correction and fine phase compensation based on the estimation results to output coherent three-dimensional point cloud or coherent detection image.

[0007] In one possible implementation, the scanner is one or a combination of a galvanometer, a rotary stage, or a fiber optic scanner.

[0008] In one possible implementation, the coherent imaging device includes a closed-loop control module connected to the digital processing module, which receives the residual phase error output by the digital processing module and selectively outputs feedback control values ​​to the FMCW laser and / or the scanner.

[0009] On the other hand, the present invention also provides a coherent imaging method, comprising: The swept light output from the FMCW laser is acquired and divided into measurement branch light, local oscillator branch light and reference branch sampling light. The target is detected based on the measurement branch light and the local oscillator branch light to obtain the measurement beat frequency signal, and the reference beat frequency signal is obtained by interfering with the reference branch sampling light and the reference light provided by the optical frequency comb. The measured beat frequency signal is subjected to sweep frequency linearization compensation and phase noise compensation to obtain the compensated measurement complex echo; Extract the reference phase noise from the reference beat frequency signal, and discretize and map the reference phase noise into scanner drift observations; A joint estimation model is constructed, which uses the compensated measurement echo and the drift observation to jointly estimate the scanner's geometric parameter error and discrete equivalent phase drift error. Based on the estimation results, the geometric mapping of the propagation distance is corrected, and the compensated measurement echo is subjected to fine phase compensation, outputting a coherent three-dimensional point cloud or a coherent detection image.

[0010] In one possible implementation, the step of performing sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measured complex echo includes: The reference beat frequency signal is demodulated to obtain a reference complex signal; The instantaneous frequency estimate is obtained by expanding and differentiating the phase of the reference complex signal; Based on the instantaneous frequency estimate and the equivalent fixed delay of the reference delay line, a time remapping function is constructed; The measured complex echo is resampled according to the time remapping function to obtain the linearized measured complex echo.

[0011] In one possible implementation, the step of performing sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measured complex echo includes: Extract the reference phase noise estimate from the reference beat frequency signal; Based on the reference phase noise estimation, the linearized measurement complex echo is subjected to phase rotation compensation at each sampling point to obtain the phase-compensated measurement complex echo.

[0012] In one possible implementation, discretizing the reference phase noise into a scanner drift observation includes: Obtain the scan position index and corresponding event stamp recorded by the scanner encoder; By using a mapping operator, the reference phase noise in the continuous time domain is converted into drift observations in the scan index domain.

[0013] In one possible implementation, the joint estimation of scanner geometric parameter error and discrete equivalent phase drift error using the compensated measurement echo and the drift observation includes: A target function is constructed based on the compensated measurement echo and the drift observation; The objective function is optimized by using weighted least squares or Kalman filtering to obtain estimates of the scanner's geometric parameter error and discrete equivalent phase drift error.

[0014] In one possible implementation, the objective function includes: The echo consistency term is constructed based on the difference between the model echo and the compensated measured echo. The reference anchor term is constructed based on the difference between the drift observation and the discrete equivalent phase drift error; The smoothed prior terms are constructed based on the difference of the discrete equivalent phase drift error of adjacent scan position indices.

[0015] In one possible implementation, after outputting a coherent 3D point cloud or a coherent detection image, the coherent imaging method includes: Calculate the residual phase error; Based on the residual phase error, update at least one of the following parameters: the scaling factor for mapping the reference phase noise to the measurement branch phase disturbance, the equivalent alignment deviation between the reference branch and the measurement branch, the reference anchoring term weight, the smoothing prior term weight, and the mapping operator used for discretization mapping. The low-frequency component of the frequency sweep linearization error is output to the frequency control terminal of the FMCW laser. The low-frequency drift error during the scanner's movement is extracted, and the low-frequency drift error is output to the scanner's drive terminal.

[0016] The beneficial effects of this invention are as follows: The coherent imaging method provided by this invention constructs a reference interference branch by introducing an optical frequency comb, extracts the reference beat frequency signal in real time to compensate for the measured beat frequency signal, and can directly capture and suppress phase jitter on the common path; at the same time, it maps the reference phase noise to the scanner drift observation, jointly estimates the geometric parameter error and the phase drift error, and then corrects the geometric mapping and performs fine phase compensation based on the estimation results, so that the originally coupled scanner bias and the slowly varying environmental drift can be corrected separately, and the phase consistency across scanning positions is significantly improved, which is conducive to improving the focusing quality of coherent imaging and the spatial accuracy of three-dimensional point clouds, thereby improving the quality of coherent imaging. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of an embodiment of the coherent imaging device provided by the present invention; Figure 2 A schematic diagram of a specific application structure of the coherent imaging device provided by the present invention; Figure 3 This is a schematic flowchart of an embodiment of the coherent imaging method provided by the present invention; Figure 4 For the present invention Figure 3 A schematic diagram of an embodiment of S303; Figure 5 For the present invention Figure 3 A schematic diagram of another embodiment of S303; Figure 6 For the present invention Figure 3 A schematic diagram of an embodiment of S304; Figure 7 For the present invention Figure 3 A schematic diagram of an embodiment of S305; Figure 8 A schematic diagram of an embodiment of the synchronous sampling triggering timing of measuring beat frequency and reference beat frequency in the coherent imaging method provided by the present invention; Figure 9 This is a comparative schematic diagram showing the change of measurement distance error with scan index in the coherent imaging method provided by the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0021] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides a coherent imaging device and method. The technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0024] Figure 1 This is a schematic diagram of an embodiment of the coherent imaging device provided by the present invention, as shown below. Figure 1 As shown, the coherent imaging device 100 includes: FMCW laser 101 outputs frequency-modulated continuous wave swept light; The beam splitter 102 is connected to the FMCW laser 101 and splits the frequency sweep light into measurement branch light, local oscillator branch light and reference branch sampling light. The measurement interferometer branch 103 receives the light from the measurement branch and the local oscillator branch, detects the target, and outputs the measurement beat frequency signal. Reference interferometer branch 104 receives the reference light from the reference branch sampling light and the reference light provided by the optical frequency comb, and outputs a reference beat frequency signal. The digital processing module 105 is connected to the measurement interferometric branch 103 and the reference interferometric branch 104 respectively. It is used to compensate the measurement beat frequency signal based on the reference beat frequency signal and generate scanner drift observation. It uses the compensation result and the drift observation to jointly estimate the geometric parameter error and phase drift error, and performs geometric mapping correction and fine phase compensation based on the estimation result to output coherent three-dimensional point cloud or coherent detection image.

[0025] Specifically, the FMCW laser 101 is used to output a frequency-modulated continuous wave swept beam. This swept beam has linear frequency modulation characteristics in time, and the sweep period and sweep bandwidth can be set according to the measurement distance and resolution requirements. For example, for the inspection of industrial parts within a range of several meters, the sweep bandwidth can be set to 1 GHz, corresponding to a theoretical distance resolution of approximately 15 cm.

[0026] Beam splitter 102 connects to FMCW laser 101 and splits the received swept light into three paths: measurement branch light, local oscillator branch light, and reference branch sampling light. Beam splitter 102 can be a fiber coupler or a free-space beam splitter prism, and its splitting ratio can be adjusted according to the optical power requirements of each branch. As a preferred method, the splitting ratio of the measurement branch light to the reference branch sampling light can be set to 90:10 to ensure the measurement signal strength while providing sufficient optical power for the reference interferometry.

[0027] The measurement interferometer branch 103 receives the measurement branch light and the local oscillator branch light, detects the target, and outputs a measurement beat frequency signal. Specifically, the measurement branch light is emitted to the target surface by the scanner after passing through an optical circulator. The echo light reflected or scattered by the target returns along the original optical path, is combined with the local oscillator branch light in a coupler after passing through the optical circulator, and is then converted into an electrical signal by a balanced photodetector. This electrical signal is the measurement beat frequency signal. The frequency of the measurement beat frequency signal is proportional to the target distance, while its phase carries fine position information of the target.

[0028] Reference interferometer branch 104 receives the sampled light from the reference branch and the reference light provided by the optical frequency comb, and outputs a reference beat frequency signal. The optical frequency comb can generate a series of stable reference lights with equal frequency intervals, and its frequency stability is much higher than that of ordinary lasers. The sampled light from the reference branch first passes through a reference delay line (e.g., a 10-meter-long fiber delay line) to introduce a fixed equivalent delay, then is combined with the reference light from the optical frequency comb in a second coupler, and finally converted into a reference beat frequency signal by a photodetector. This reference beat frequency signal reflects the sweep frequency nonlinearity, phase noise, and common optical path disturbance of the FMCW laser in real time.

[0029] The digital processing module 105 is connected to the measurement interferometry branch 103 and the reference interferometry branch 104, respectively, and is used to receive the measurement beat frequency signal and the reference beat frequency signal. The digital processing module 105 first compensates for the measurement beat frequency signal based on the reference beat frequency signal, and simultaneously generates scanner drift observations. Next, it uses the compensation results and drift observations to jointly estimate geometric parameter errors and phase drift errors, and performs geometric mapping correction and fine phase compensation based on the estimation results, finally outputting a coherent 3D point cloud or a coherent detection image. It should be noted that geometric mapping correction refers to correcting the nonlinear relationship between the actual angle of the scanner and the theoretical propagation distance, such as the distance error introduced by the zero-bias angle of the galvanometer or the installation eccentricity of the rotary stage.

[0030] In some embodiments of the present invention, the scanner may employ one or more combinations of a galvanometer, a rotary stage, or a fiber optic scanner. A galvanometer is suitable for rapid scanning over a small area, a rotary stage is suitable for large-area circular scanning, and a fiber optic scanner is suitable for flexible detection paths.

[0031] For example, in industrial online inspection scenarios, a two-dimensional grating scan of the target can be achieved using a dual-axis galvanometer, with a scanning rate of up to several thousand points per second.

[0032] Furthermore, the coherent imaging device also includes a closed-loop control module. This closed-loop control module is connected to the digital processing module 105, receives the residual phase error output by the digital processing module 105, and selectively outputs feedback control quantities to the FMCW laser 101 and / or the scanner. The residual phase error reflects the low-frequency drift component that still exists after digital compensation. The closed-loop control module feeds this low-frequency component back to the frequency control terminal of the FMCW laser 101, which can adjust the sweep start frequency or sweep slope of the laser in real time, thereby suppressing long-term drift of sweep nonlinearity. At the same time, feeding back the low-frequency deviation to the drive terminal of the scanner can compensate for the slow angular drift of the scanner caused by mechanical wear or temperature drift.

[0033] In a specific example, suppose a three-dimensional surface inspection of a large wind turbine blade is performed, and the scanner uses a rotary table to drive a one-dimensional galvanometer. During continuous scanning for several hours, changes in ambient temperature cause the laser center frequency to drift slowly, while the temperature rise of the rotary table bearings causes the angle zero-point drift. In this embodiment, the reference interferometer branch 104 continuously outputs a reference beat frequency signal, and the digital processing module 105 extracts the laser frequency drift information and the slowly changing angle error of the rotary table in real time. On the one hand, it corrects the phase of the measured complex echo of the current frame through a compensation algorithm; on the other hand, it feeds back the correction amount to the laser frequency control terminal and the rotary table drive through a closed-loop control module, thereby ensuring that the point clouds of each frame obtained by long-term scanning maintain phase consistency and avoid splicing misalignment or image blurring.

[0034] In a specific application scenario, such as Figure 2 As shown, Figure 2 The specific application example of the coherent imaging device provided in this embodiment includes an FMCW laser 101, a beam splitter 102, an optical circulator 1031, a scanner 1032, a first coupler 1033, a balanced photodetector 1034, a reference delay line 1041, an optical frequency comb 1042, a second coupler 1043, a photodetector 1044, a digital processing module 105, and a closed-loop control module 106. It should be noted that 107 is the target and is determined according to actual needs.

[0035] The FMCW laser 101 outputs frequency-modulated continuous wave swept light. The input of the beam splitter 102 is connected to the output of the FMCW laser 101, splitting the swept light into three paths: measurement branch light S1, local oscillator branch light S2, and reference branch sampling light S3. Among them, the measurement branch light S1 enters port 1 of the optical circulator 1031, the local oscillator branch light S2 is directly sent to one input of the first coupler 1033, and the reference branch sampling light S3 is sent to the input of the reference delay line 1041.

[0036] Port 2 of the optical circulator 1031 is connected to the scanner 1032, which guides the measurement branch light S1 to the scanner 1032, and the scanner 1032 emits a scanning beam towards the target. The echo light generated by the target returns to the scanner 1032 along the original optical path, and then enters port 2 of the optical circulator 1031. The optical circulator 1031 extracts the echo light from its port 3, denoted as the echo light R. The function of the optical circulator 1031 is to achieve directional isolation between the emitted light and the received echo light, preventing the emitted light from directly entering the receiving channel and causing interference.

[0037] The scanner 1032 receives measurement light from the optical circulator 1031, performs spatial scanning of the target, and receives the echoes reflected or scattered by the target. The scanner 1032 can be one or more combinations of a galvanometer, a rotating stage, or a fiber optic scanner. For example, in a two-dimensional imaging scene, two orthogonal galvanometers can be used to achieve grating scanning; in a large field-of-view circular scan, a rotating stage can be used to drive the scanning head.

[0038] The two input terminals of the first coupler 1033 are respectively connected to port 3 of the optical circulator 1031 (receiving the echo light R) and the local oscillator branch of the beam splitter 102 (receiving the local oscillator light S2). The first coupler 1033 combines the echo light and the local oscillator light, and its output terminal is connected to the balanced photodetector 1034. The balanced photodetector 1034 converts the combined optical signal into a measurement beat frequency electrical signal and outputs it to the digital processing module 105. The balanced detection method can effectively suppress common-mode noise and improve the signal-to-noise ratio.

[0039] The input of reference delay line 1041 receives the reference branch sampled light S3 from beam splitter 102, and its output is connected to one input of second coupler 1043. Reference delay line 1041 introduces a fixed equivalent delay into the reference branch sampled light S3, which can be achieved, for example, using a section of fiber of known length or a free-space optical path. This fixed delay is used for calculating the instantaneous sweep slope in subsequent sweep frequency linearization compensation.

[0040] The optical frequency comb 1042 outputs a stable reference light, and its output end is connected to the other input end of the second coupler 1043. The optical frequency comb 1042 provides a uniformly spaced and long-term stable optical frequency reference, and its frequency stability is much higher than that of ordinary FMCW lasers.

[0041] The second coupler 1043 combines the sampled swept light S3 after passing through the reference delay line 1041 with the reference light output from the optical frequency comb 1042, and its output is connected to the photodetector 1044. The photodetector 1044 converts the combined optical signal into a reference beat frequency electrical signal. IF ref ( t ), and output to the digital processing module 105.

[0042] The digital processing module 105 receives the measurement beat frequency signal output from the balanced photodetector 1034 and the reference beat frequency signal output from the photodetector 1044, respectively. The digital processing module 105 compensates for the measurement beat frequency signal based on the reference beat frequency signal and simultaneously generates scanner drift observations. It uses the compensation results and drift observations to jointly estimate geometric parameter errors and phase drift errors, and performs geometric mapping correction and fine phase compensation based on the estimation results, ultimately outputting a coherent 3D point cloud or a coherent detection image. The digital processing module 105 also outputs the residual phase error to the closed-loop control module 106.

[0043] The closed-loop control module 106 is connected to the output of the digital processing module 105, and can also be selectively connected to the frequency control terminal of the FMCW laser 101 and / or the drive terminal of the scanner 1032. Based on the received residual phase error, the closed-loop control module 106 updates its internal parameters (such as the proportional gain, alignment deviation, weights, etc.), feeds back the low-frequency component of the sweep linearization error to the FMCW laser 101, and feeds back the low-frequency drift error of the scanner to the drive terminal of the scanner 1032, thereby forming a compensation closed loop and improving the long-term stability of the system.

[0044] It should be noted that the optical path connections between the above components can be achieved using single-mode fiber or free-space optical paths, depending on the actual application scenario and cost requirements. For example, in integrated small systems, fiber optic devices are preferred to simplify optical path alignment; in ultra-high power or extreme environment applications, free-space optical paths can be used to withstand higher optical power or adapt to special environments.

[0045] The coherent imaging device provided in this embodiment, by introducing an optical frequency comb as a reference and employing a digital processing module to simultaneously perform phase compensation, drift observation, joint estimation, and geometric correction, can effectively suppress various disturbances such as sweep frequency nonlinearity, laser phase noise, scanner geometric errors, and slowly varying environmental drift. The addition of a closed-loop control module further enhances the system's long-term adaptive capability and output stability. This significantly improves phase consistency across scanning positions, increases the repeatability accuracy of the 3D point cloud, and thus helps meet the practical requirements of high quality and high stability for precision industrial inspection and long-distance coherent imaging.

[0046] Figure 3 This is a schematic flowchart of an embodiment of the coherent imaging method provided by the present invention, as shown below. Figure 3 As shown, the coherent imaging method includes: S301. Obtain the swept light output from the FMCW laser and divide the swept light into measurement branch light, local oscillator branch light and reference branch sampling light; S302. The target is detected based on the measurement branch light and the local oscillator branch light to obtain the measurement beat frequency signal, and the reference beat frequency signal is obtained by interfering with the reference branch sampling light and the reference light provided by the optical frequency comb. S303. Extract the reference phase noise from the reference beat frequency signal and discretize the reference phase noise to map it into scanner drift observations; S304. Perform sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measurement echo; S305. Construct a joint estimation model, use the compensated measurement echo and drift observation to jointly estimate the scanner's geometric parameter error and discrete equivalent phase drift error, and correct the geometric mapping of the propagation distance based on the estimation results and perform fine phase compensation on the compensated measurement echo to output a coherent three-dimensional point cloud or coherent detection image.

[0047] The coherent imaging method in this embodiment can be applied to the coherent imaging device described in the aforementioned device embodiment. The specific steps are explained below.

[0048] First, the swept light output from the FMCW laser 101 is acquired, and then split into a measurement branch light, a local oscillator branch light, and a reference branch sampling light by a beam splitter 102. The splitting ratio can be adjusted according to actual needs, for example, the measurement branch light accounts for 90%, and the local oscillator branch light and the reference branch sampling light each account for 5%.

[0049] Preferably, this embodiment synchronously triggers multi-channel synchronous sampling. The boundary of each sweep cycle is defined by a sweep trigger signal (SYNC), and sampling occurs within the same sampling clock. clk s The following method utilizes a digital processing module to synchronously sample and measure the beat frequency signal of the measurement branch. IF mea ( t (and the reference branch beat frequency signal, and record the scan position index) m This index is obtained from the scanner's encoder. Simultaneously, the effective acquisition window is defined to elicit the sweep transition segment. The sweep slope of the FMCW laser is defined as:

[0050] in, T For the frequency sweep period, B This is the sweep bandwidth.

[0051] Next, the measurement branch light is emitted to the target surface via an optical circulator and scanner. The echo reflected or scattered by the target is combined with the local oscillator branch light in a coupler, and then converted into an electrical signal by a balanced photodetector, which is the measurement beat frequency signal. Simultaneously, the reference branch sampling light and the reference light provided by the optical frequency comb are interfered with to obtain the reference beat frequency signal. The reference branch sampling light is then combined with the stable reference light from the optical frequency comb after passing through a fixed delay line, and converted into a reference beat frequency signal by a photodetector. For example, when detecting surface defects on large wind turbine blades, the sweep bandwidth is set to 2 GHz and the sampling rate to 100 MS / s to ensure the accuracy of subsequent phase compensation.

[0052] Furthermore, a digital processing module is used to process the beat frequency signal of the measurement branch. IF mea ( t After down-conversion and filtering, the measured complex echo is obtained:

[0053] in, This indicates the instantaneous amplitude of the measured complex echo. This represents the instantaneous phase of the measured complex echo, where j represents the imaginary unit, defined as a number whose square equals -1.

[0054] And according to the scan index m Constructing continuous-time complex echoes into a discrete complex sample set { s m},in:

[0055] in, s m Corresponding to the m The echo at each scan position, Indicates the first m Echo intensity at each scan position, Indicates the first m The echo phase at each scan position.

[0056] After obtaining the measured beat frequency signal and the reference beat frequency signal, the measured beat frequency signal is subjected to sweep frequency linearization compensation and phase noise compensation to obtain the compensated measured complex echo. Sweep frequency linearization compensation is used to correct the beat frequency broadening caused by the sweep frequency nonlinearity of the FMCW laser, while phase noise compensation is used to suppress the phase noise of the laser itself and the phase jitter introduced by common optical path disturbances (such as temperature drift and vibration). Both compensations rely on the reference phase information extracted from the reference beat frequency signal.

[0057] Then, the reference phase noise is extracted from the reference beat frequency signal and discretized to map it as a scanner drift observation. The reference phase noise reflects the phase disturbance on the common optical path; discretizing it according to the scan position index recorded by the scanner encoder yields the drift observation corresponding to each scan position. This observation is subsequently used to constrain the estimation of the discrete equivalent phase drift error.

[0058] Finally, a joint estimation model is constructed, utilizing the compensated measurement echoes and drift observations to jointly estimate the scanner's geometric parameter errors and discrete equivalent phase drift errors. Scanner geometric parameter errors include angular zero bias, installation eccentricity, and shaft wobble, while the discrete equivalent phase drift error represents the incompletely compensated slow-varying phase remnants. Based on the estimation results, the geometric mapping of the propagation distance is corrected (i.e., the nonlinear relationship between the scanner command angle and the actual propagation distance is corrected), and fine phase compensation is performed on the compensated measurement echoes. Finally, a coherent 3D point cloud or coherent detection image is output.

[0059] As an example, scanning imaging was performed on a precision-cast aero-engine blade. The blade surface has high curvature and specular reflection characteristics. The scanner uses a biaxial galvanometer with a sweep period of 1 ms, acquiring 2048 sampling points per cycle. Using the above method, the nonlinear jitter of the laser at the beginning of the sweep and the common phase noise caused by environmental vibration were captured in real time by the reference beat frequency signal. The standard deviation of the measured complex echo phase after compensation was reduced from 0.35 rad to 0.06 rad. The joint estimation model further decoupled the geometric error (approximately 0.05 degrees) caused by galvanometer mounting eccentricity and the slowly varying phase drift (approximately 0.2 rad / hour) caused by temperature drift. The final output 3D point cloud showed significant elimination of ghosting at the blade edge, and the point cloud thickness was reduced from the original 1.2 m to 0.3 m.

[0060] The coherent imaging method provided in this embodiment introduces an optical frequency comb as a reference standard to acquire reference phase noise in real time. It uses the compensated measurement echo and drift observation to jointly estimate geometric and drift errors. Combined with geometric mapping correction and fine phase compensation, it can effectively suppress phase inconsistencies across scanning positions, improve the focusing quality of coherent imaging and the geometric accuracy of three-dimensional point clouds, and help meet the actual needs of high-precision three-dimensional inspection in fields such as aviation, energy, and precision manufacturing.

[0061] In some embodiments of the present invention, such as Figure 4 As shown, step S303 performs sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measurement complex echo, including: S401. Demodulate the reference beat frequency signal to obtain the reference complex signal; S402. Expand and differentiate the phase of the reference complex signal to obtain the instantaneous frequency estimate; S403. Construct a time remapping function based on the instantaneous frequency estimate and the equivalent fixed delay of the reference delay line; S404. Based on the time remapping function, the measured complex echo is resampled to obtain the linearized measured complex echo.

[0062] This embodiment further explains the sweep frequency linearization compensation in step S303. This compensation process utilizes the phase information in the reference beat frequency signal to eliminate the influence of the FMCW laser sweep frequency nonlinearity on the measured beat frequency signal.

[0063] In step S401, demodulation typically employs Hilbert transform or quadrature down-conversion to convert the real-valued reference beat frequency signal into a complex exponential form reference complex signal in order to extract its instantaneous phase. Specifically, the reference beat frequency signal... IF ref ( tAfter digital down-conversion and low-pass filtering, it can be expressed as:

[0064] in, Instantaneous amplitude It is the instantaneous phase.

[0065] In step S402, due to the instantaneous phase When obtained using the arctangent function, it will be wrapped in ( Within the interval [π, π], a phase expansion operation is required to eliminate phase jumps and make the phase a continuous function of time. Then, the time derivative of the expanded phase is calculated to obtain the instantaneous frequency estimate of the reference beat frequency.

[0066] in, This represents taking the phase angle with respect to the reference complex signal, d() / dt represents differentiation, and unwrap{ } indicates that the principal phase sequence is de-jumped and expanded so that the phase changes continuously with time.

[0067] Subsequently, the equivalent instantaneous sweep slope is obtained from the reference delay:

[0068] in, τ ref The equivalent fixed delay is used as a reference branch.

[0069] The reference phase noise estimate is extracted from the reference beat frequency signal. The instantaneous phase of the reference complex signal consists of three components: deterministic sweep phase, reference phase noise, and constant phase.

[0070] Then, according to Construct a time remapping function such that the "equivalent frequency sweep" satisfies the nominal slope after remapping. k :

[0071] The above equation obtains the equivalent instantaneous sweep slope from the reference delay. Integrate over the interval (0, t), then divide by the previously defined sweep slope of the FMCW laser. k Based on this, the measured complex echo y(t) is resampled into linearized data. This ensures that the subsequent "beat frequency-distance" mapping and coherent processing are not distorted due to frequency sweep nonlinearity.

[0072] It should be understood that this embodiment uses a time remapping function to process the reference beat frequency signal, estimate the sweep frequency slope, and resample the measurement signal to compensate for the sweep frequency nonlinearity, so that the subsequent imaging data can maintain high quality.

[0073] Then, first use Deterministic sweep phase term for reference:

[0074] Finally, the reference phase noise estimate is obtained:

[0075] in, This is an estimate of the constant phase term.

[0076] It should be noted that there is a hardware link delay difference between the optical paths of the reference branch and the measurement branch. Therefore, there is a fixed alignment deviation between the reference phase noise estimate and the measurement echo on the time axis. To ensure the reference phase noise... The measurement echo is correctly applied in time to the linearized signal. Therefore, the closed-loop compensation module first obtains the equivalent alignment deviation Δ between the reference branch and the measurement branch based on the hardware link delay of the two channels. t This is used to compensate for the inconsistency between the time axes of the measurement branch and the reference branch. Then, the scaling factor for mapping the phase noise observed in the reference branch to the phase disturbance in the measurement branch is calibrated. α .

[0077] As an example, assume the nominal sweep slope of the FMCW laser is 10 MHz / μs, and the reference delay line length is 20 meters (corresponding to an equivalent delay of approximately 0.1 μs). Actual measurements revealed that the instantaneous frequency of the reference beat frequency was 5% lower at the beginning of the sweep and 3% higher at the end. After constructing a time remapping function using the method described above, the measured complex echo was resampled. The beat frequency spectrum, originally broadened due to nonlinearity, was compressed to near the theoretical bandwidth, and the beat frequency peak corresponding to the target distance became sharper. Experiments showed that the range resolution after resampling improved from approximately 30 cm to nearly the theoretical value of 15 cm.

[0078] This embodiment corrects the actual nonlinear sweep frequency signal into an equivalent linear sweep frequency by taking the phase derivative of the reference complex signal and remapping the time, thereby eliminating the negative impact of sweep frequency nonlinearity on distance measurement and coherent accumulation, which is beneficial to improving the range resolution and cross-cycle phase consistency of the FMCW coherent imaging system.

[0079] In some embodiments of the present invention, such as Figure 5As shown, step S303 performs sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measurement complex echo, which includes: S501. Extract reference phase noise estimation from reference beat frequency signal; S502. Based on the reference phase noise estimation, perform phase rotation compensation on the linearized measurement complex echo point by point to obtain the phase-compensated measurement complex echo.

[0080] This embodiment further explains the phase noise compensation in step S303. This compensation process uses the phase noise estimation extracted from the reference beat frequency signal to perform point-by-point phase correction on the measurement complex echo that has undergone frequency sweep linearization processing, in order to suppress phase jitter caused by laser phase noise and common optical path disturbances.

[0081] Based on the reference phase noise estimation, a sampling-point-by-sampling phase rotation compensation is performed on the linearized measurement complex echo to obtain the phase-compensated measurement complex echo. This is to suppress the common phase jitter of the measurement complex echo signal (mainly including phase noise of the FMCW laser, phase jitter caused by frequency sweep nonlinearity, and phase jitter caused by temperature drift or vibration in the common optical path before the beam splitter).

[0082] in, y c ( t The signal is the measurement echo after point-by-point phase compensation. This is the measurement recovery echo after frequency sweeping, linearization, and resampling. This is a phase rotation factor used to cancel common phase jitter in the measured echo. This phase rotation is performed independently at each sampling point, and is a point-by-point complex multiplication operation, resulting in high computational efficiency.

[0083] This embodiment extracts phase noise estimation from the reference beat frequency signal and performs phase rotation compensation on the linearized measurement complex echo point by point. This effectively suppresses broadband phase jitter introduced by laser phase noise, sweep frequency residual nonlinearity, and common optical path temperature drift and vibration, significantly improving phase consistency across scanning positions and sweep frequency cycles, and providing clean complex echo data for subsequent joint estimation and high-precision 3D reconstruction.

[0084] In some embodiments of the present invention, such as Figure 6 As shown, step S304 discretizes the reference phase noise and maps it to scanner drift observations, including: S601. Obtain the scan position index and corresponding event stamp recorded by the scanner encoder; S602. Using the mapping operator, the reference phase noise in the continuous time domain is converted into the drift observation in the scan index domain.

[0085] This embodiment further explains the discretization mapping in step S304. This process converts the reference phase noise in the continuous time domain into drift observations associated with the scanner position index, providing a data basis for subsequent joint estimation in the scan position domain.

[0086] Retrieve the scan position index and corresponding event stamp recorded by the scanner encoder. During spatial scanning, the scanner's internal encoder records the current scan position in real time (e.g., the deflection angle of the galvanometer or the azimuth angle of the rotary table) and assigns an incrementing index to each scan position. m Meanwhile, the data acquisition system records each index. m The corresponding sampling time is called the event stamp. t m Event stamps are typically measured in sampling clock cycles to ensure an accurate correspondence between the timeline and the scan position index. For example, in 2D raster scanning, the encoder outputs a line synchronization signal after each scan line is completed, and the event stamp corresponding to this signal is recorded as follows: t m ,in m This is a row index.

[0087] Using a mapping operator, the reference phase noise in the continuous-time domain is converted into a drift observation in the scan-index domain. (Reference phase noise) It is a discrete-time series obtained under continuously sampled clocks, where each sampling point corresponds to a time coordinate. Mapping operator M The function of {} is to determine the event stamp. t m Extract or interpolate the corresponding scan position from the continuous time domain. m The reference phase noise value. Specifically, nearest neighbor decimation or linear interpolation methods can be used:

[0088] in, z m That is, the first m The drift observation at the first scan position. This observation reflects the scanner's drift during the execution of the first scan position. m At each position, the low-frequency phase perturbation component in the common optical path.

[0089] It should be noted that the mapping operator may also include filtering or downsampling operations to remove high-frequency components (such as shot noise) from the reference phase noise, retaining only the slowly varying components related to scanner drift. As a preferred approach, before mapping, the high-frequency components are first filtered or downsampled. A low-pass filter with a cutoff frequency of 5 times the scan frame rate is applied, and then the data is extracted based on the event stamp to obtain a smoother drift observation.

[0090] As an example, suppose the scanner is a dual-axis galvanometer that performs raster scanning at a frame rate of 10 Hz, with each frame containing 500 × 500 scan points. The encoder generates a pulse at the start of each scan point, and the event is stamped. t m The data is latched by a hardware timer with an accuracy of 10 nanoseconds. A reference phase noise sequence is acquired at a sampling rate of 100 MHz, with each frame lasting 10 milliseconds. Using a mapping operator, the reference phase noise value corresponding to each event stamp is extracted, forming a drift observation sequence of length 250,000. Experiments show that the peak amplitude in this sequence can reach 0.3 rad, mainly originating from the low-frequency components of environmental vibrations.

[0091] This embodiment obtains the scan position index and event stamp, and uses a mapping operator to convert the reference phase noise in the continuous time domain into the drift observation in the scan index domain. This provides a reliable observation input for the reference anchoring term in the subsequent joint estimation model, which is beneficial for distinguishing and correcting geometric errors and drift errors in the scan position domain, thereby improving the parameter estimation accuracy of the coherent imaging system.

[0092] In some embodiments of the present invention, such as Figure 7 As shown, step S305, which uses the compensated measurement echo and drift observation to jointly estimate the scanner's geometric parameter error and discrete equivalent phase drift error, includes: S701. Construct an objective function based on the compensated measurement echo and drift observation; S702. Using weighted least squares or Kalman filtering, the objective function is optimized to obtain the estimated values ​​of the scanner's geometric parameter error and the discrete equivalent phase drift error.

[0093] In some embodiments of the present invention, the objective function includes: The echo consistency term is constructed based on the difference between the model echo and the compensated measured echo. The reference anchoring term is constructed based on the difference between the drift observation and the discrete equivalent phase drift error. The smoothed prior terms are constructed based on the difference of the discrete equivalent phase drift error of adjacent scan position indices.

[0094] This embodiment further explains the joint estimation in step S305. After phase rotation compensation, the broadband common phase jitter in the measurement data has been significantly suppressed, but two main types of errors still remain in the system: one is the geometric model deviation caused by the angle and pose errors of the scanner. First, the propagation distance is inconsistent with the true value; second, the differential optical path disturbance not fully covered by the reference branch and the slowly varying environmental drift cause phase drift to still exist at the discrete scanning positions. This embodiment utilizes the compensated measurement echo and drift observation to construct an objective function and optimize its solution, thereby achieving decoupled estimation of scanner geometric parameter errors and discrete equivalent phase drift errors.

[0095] First, an objective function is constructed based on the compensated measurement echo and drift observations. The compensated measurement echo has undergone sweep linearization compensation and phase noise compensation, but geometric parameter errors of the scanner itself (such as angular zero offset, installation eccentricity, and shaft wobble) and incompletely compensated slow-varying phase drift still exist. Therefore, a discrete phase model is established to describe the... m Echo phase at each scan position:

[0096] in, m This is the scan position index, used to identify the scanner's position in the discrete sampling sequence. m One location; λ The center wavelength of the FMCW laser; The propagation distance is determined by the scanner's scanning angle parameters and is used to reflect the impact of scanner angular offset, installation error, pose disturbance, etc. on the distance. For the first m The equivalent phase shift at each position is used to reflect slow-varying optical path disturbances, residual phase shifts, etc. This is the noise term, which includes random disturbances such as detection noise, phase extraction error, and speckle noise.

[0097] Then, the objective function is jointly estimated using methods such as weighted least squares or Kalman filtering to obtain... and Preferably, this is achieved by minimizing the following objective function:

[0098] Among them, the first item This is an echo consistency term used to ensure model echo consistency. Compared with the measured echo s m The consistency directly determines the quality of coherent focusing; the second item For reference anchoring, it indicates that the observation is performed using the reference branch. z m Constraint drift To prevent drift estimation from drifting away and to avoid geometric errors being absorbed by drift; the third term To smooth the prior terms, this means utilizing the slow-varying drift characteristics to suppress drift jumps caused by noise and improve estimation stability. β The reference anchor weights are used to adjust the observations of the reference branch. z m The degree of fit,β A larger value indicates a higher reference branch observation value. z m The more credible; γ To smooth out prior weights, used to constrain the randomness m The change is slow. γ The larger the value, the more likely it is to represent The smoother the surface, the stronger its noise resistance.

[0099] Finally, using precise phase compensation to discrete complex samples s m This is used to correct the scanner's geometric model or scan angle mapping to compensate for distance deviations and phase drift caused by the scanner's angle and pose errors. Specifically, based on... Fine phase compensation is performed on discrete complex samples as follows: ; based on and Output high-quality 3D point clouds.

[0100] Furthermore, the objective function is designed to make the model echo as consistent as possible with the measured echo, while using drift observations to constrain drift estimation and leveraging the slow-varying characteristics of drift to apply smoothing constraints. Specifically, the objective function includes three terms: The first term is the echo consistency term, which is constructed based on the difference between the model echo and the compensated measured complex echo. It is usually expressed as the sum of squares of the complex signal difference or in the form of log-likelihood, and is used to ensure that the estimated values ​​of geometric parameters and drift parameters can accurately fit the actual measurement data.

[0101] The second term is the reference anchoring term, constructed based on the difference between the drift observation and the discrete equivalent phase drift error. This penalty term forces the estimated drift observation to approximate the drift observation obtained from the reference phase noise mapping, thereby preventing overall drift in the drift estimate and avoiding the incorrect absorption of geometric parameter errors by the drift term.

[0102] The third term is the smoothing prior, constructed based on the difference in discrete equivalent phase drift errors between adjacent scan position indices, i.e., the constraint magnitude. This prior reflects the characteristic that slowly varying drifts (such as temperature drift and mechanical creep) in the physical environment change smoothly between adjacent scan positions, effectively suppressing drift estimation jumps caused by noise.

[0103] This embodiment constructs an objective function that includes an echo consistency term, a reference anchoring term, and a smoothing prior term, and uses weighted least squares or Kalman filtering for optimization. This effectively decouples the scanner's geometric parameter errors from the slowly varying phase drift, obtaining high-precision estimates for each. This provides a reliable basis for subsequent geometric mapping correction and fine phase compensation, which is beneficial for improving the measurement accuracy and environmental adaptability of the coherent imaging system.

[0104] In some embodiments of the present invention, after outputting a coherent three-dimensional point cloud or a coherent detection image, the coherent imaging method includes: Calculate the residual phase error; Based on the residual phase error, update at least one of the following parameters: the scaling factor for mapping the reference phase noise to the measurement branch phase disturbance, the equivalent alignment deviation between the reference branch and the measurement branch, the reference anchoring term weight, the smoothing prior term weight, and the mapping operator used for discretization mapping. The low-frequency component of the frequency sweep linearization error is output to the frequency control terminal of the FMCW laser. Extract the low-frequency drift error during scanner movement and output the low-frequency drift error to the scanner's driver.

[0105] This embodiment further explains the closed-loop feedback step after outputting the coherent 3D point cloud or coherent detection image. This step constitutes an independent feedback loop, used to update system parameters online and output correction values ​​to the hardware actuators to ensure the long-term stability of the system.

[0106] After outputting the 3D point cloud for the current frame or scan cycle, the digital processing module first calculates the residual phase error. The residual phase error refers to the phase difference that still exists between the model echo and the measured echo after the aforementioned frequency sweep linearization compensation, phase noise compensation, joint estimation, and fine phase compensation. This residual can be obtained by statistically analyzing the phase difference values ​​in the echo consistency term, for example, by calculating the average phase deviation or root mean square phase difference across all scan positions. The magnitude of the residual phase error reflects the degree of matching between the current system parameters (such as scaling factor, alignment deviation, weights, etc.) and the actual optical path state.

[0107] Based on the residual phase error, update at least one of the following parameters: the scaling factor for mapping the reference phase noise to the phase disturbance of the measurement branch. α Equivalent alignment deviation between reference branch and measurement branch Δt Reference anchor weight β Smoothing prior weights γ and mapping operators used for discretization mapping. M {·}.

[0108] Specifically, when the residual phase error remains consistently large, it indicates that... α and Δt The calibrated value has deviated from the actual value and can be fine-tuned using gradient descent or Kalman filtering. If there is a systematic deviation between the drift observation and the finely compensated phase residual, it indicates... β Smaller or γIf the values ​​are too large, the weighting coefficients should be adjusted accordingly. If there is a cumulative error between the time point extracted by the mapping operator and the scanner event stamp, the interpolation coefficients of the mapping operator need to be corrected. Parameter updates can be performed using an online recursive algorithm (such as recursive least squares), updating once per frame, so that the system gradually approaches the optimal operating point.

[0109] The low-frequency component of the sweep linearization error is output to the frequency control terminal of the FMCW laser. The sweep linearization error refers to the residual deviation between the actual sweep slope and the nominal slope after time remapping. The low-frequency component of this error (e.g., the portion below 0.1 Hz) mainly originates from slowly varying factors such as laser temperature drift and power supply fluctuations. By outputting this component to the laser's analog modulation port or digital control interface via a digital-to-analog converter, the sweep start frequency or sweep slope of the laser can be fine-tuned in real time, suppressing long-term drift of sweep nonlinearity at its source.

[0110] The low-frequency drift error during scanner movement is extracted and output to the scanner's drive. During long-term operation, due to factors such as bearing wear, temperature changes, and zero drift in the drive circuit, a slowly changing deviation occurs between the actual angle and the commanded angle. This low-frequency drift error can be identified from the discrete equivalent phase drift error. ψ m Extract from, specifically, to ψ m The sequence is low-pass filtered (with a cutoff frequency much lower than the scan frame rate) to obtain the low-frequency drift component corresponding to each scan position. This component is converted into a voltage or current signal and output to the servo drive input of the scanner to compensate for the slow angle drift of the scanner.

[0111] This embodiment calculates the residual phase error and dynamically updates the system parameters. Simultaneously, it feeds back the low-frequency component of the sweep linearization error to the laser and the low-frequency drift error of the scanner to its driver, forming a complete compensation mechanism. This mechanism can adaptively track the slow drift of the optical path and scanner state, significantly improving the imaging consistency and point cloud repeatability under continuous long-term operating conditions. This is beneficial for meeting the application requirements of unattended, long-term automatic detection in industrial settings.

[0112] Please see Figure 8 , Figure 8 This diagram illustrates an embodiment of the synchronous sampling triggering timing of the measured beat frequency and the reference beat frequency in the coherent imaging method provided by the present invention. It serves to explain the co-source triggering and co-clocking sampling of the two beat frequency signals within each sweep cycle, and their correspondence with the scan position index. From top to bottom, the diagram shows: the sweep trigger signal SYNC, the scanner encoder position update timing, and the scan position index. m The progressive relationship, measurement of beat frequency signal IFmea ( t Reference beat frequency signal IF ref ( t ), sampling clock clk s This includes the sampling pulse sequence and the sampling window settings within each sweep cycle. Specifically, each trigger pulse of SYNC corresponds to times T1~T3, and its rising edge is used to define the boundary between adjacent sweep cycles. The vertical dashed lines in the figure represent the boundaries between each sweep cycle. Within each cycle, the scanner driver / encoder output generates a corresponding position update, making the scan position index... m It increments in predetermined steps within a period and accumulates continuously between adjacent periods. Figure 8 In the example, the indices at the periodic boundaries are as follows: m =0, 4, 8, indicating that a certain number of scan position samples are acquired within each sweep frequency cycle (measuring the beat frequency). IF mea ( t ) and reference beat frequency IF ref ( t All are sampled on the same clock. clk s Under control, data is sampled synchronously using the same time reference, ensuring a one-to-one correspondence between the two data streams. Simultaneously, a sampling window is set within each frequency sweep cycle, extracting and processing data only from the effective range covered by the sampling window to avoid the initial transition segment or non-steady-state intervals of the frequency sweep. This is achieved through the aforementioned SYNC periodic framing and sharing. clk s Synchronous sampling and scanning index m Alignment with the sampling window provides a consistent time reference and data registration conditions for subsequent extraction of reference phase information, measurement phase compensation, and joint estimation of the scan position domain.

[0113] Figure 9 The measurement distance error in the coherent imaging method provided by this invention varies with the scan index. m The diagram illustrates the changes, where the solid black line in the upper diagram represents the distance error curve without phase compensation, and the dashed black line in the lower diagram represents the distance error curve after applying the phase compensation method of this invention. Figure 9As can be seen, the distance error under uncompensated conditions exhibits obvious random fluctuations and slow drift superposition characteristics during the scanning process, reflecting the influence of common phase jitter such as sweep frequency light source phase noise / sweep frequency nonlinearity and common optical path disturbance on the ranging results. After introducing reference branch phase information and completing time alignment and mapping calibration, this invention performs phase correction on the measured complex echo point by point, which reduces the overall fluctuation amplitude of the compensated distance error curve and makes the fluctuation more stable, indicating that the influence of common phase disturbance on the ranging results is effectively suppressed.

[0114] The coherent imaging apparatus and method provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A coherent imaging device, characterized in that, include: FMCW laser, outputting frequency-modulated continuous wave swept light; A beam splitter, connected to the FMCW laser, splits the swept beam into a measurement branch beam, a local oscillator branch beam, and a reference branch sampling beam. The measurement interferometer branch receives the light from the measurement branch and the local oscillator branch, detects the target, and outputs a measurement beat frequency signal; The reference interference branch receives the sampled light from the reference branch and the reference light provided by the optical frequency comb, and outputs a reference beat frequency signal. A digital processing module, connected to the measurement interferometer branch and the reference interferometer branch respectively, is used to compensate the measurement beat frequency signal based on the reference beat frequency signal, and at the same time generate scanner drift observations; The geometric parameter error and phase drift error are estimated by combining the compensation result with the drift observation. Based on the estimation result, geometric mapping correction and fine phase compensation are performed to output a coherent three-dimensional point cloud or coherent detection image.

2. The coherent imaging apparatus according to claim 1, characterized in that, The scanner is one or a combination of a galvanometer, a rotary stage, or a fiber optic scanner.

3. The coherent imaging apparatus according to claim 1, characterized in that, The coherent imaging device includes a closed-loop control module connected to the digital processing module. The closed-loop control module receives the residual phase error output by the digital processing module and selectively outputs feedback control quantities to the FMCW laser and / or the scanner.

4. A coherent imaging method, characterized in that, include: The swept light output from the FMCW laser is acquired and divided into measurement branch light, local oscillator branch light and reference branch sampling light. The target is detected based on the measurement branch light and the local oscillator branch light to obtain the measurement beat frequency signal, and the reference beat frequency signal is obtained by interfering with the reference branch sampling light and the reference light provided by the optical frequency comb. Extract the reference phase noise from the reference beat frequency signal, and discretize and map the reference phase noise into scanner drift observations; The measured beat frequency signal is subjected to sweep frequency linearization compensation and phase noise compensation to obtain the compensated measurement complex echo; A joint estimation model is constructed, which uses the compensated measurement echo and the drift observation to jointly estimate the scanner's geometric parameter error and discrete equivalent phase drift error. Based on the estimation results, the geometric mapping of the propagation distance is corrected, and the compensated measurement echo is subjected to fine phase compensation, outputting a coherent three-dimensional point cloud or a coherent detection image.

5. The method of coherent imaging according to claim 4, characterized in that, The step of performing sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measured complex echo includes: The reference beat frequency signal is demodulated to obtain a reference complex signal; The instantaneous frequency estimate is obtained by expanding and differentiating the phase of the reference complex signal; Based on the instantaneous frequency estimate and the equivalent fixed delay of the reference delay line, a time remapping function is constructed; The measured complex echo is resampled according to the time remapping function to obtain the linearized measured complex echo.

6. The method of coherent imaging according to claim 4, characterized in that, The step of performing sweep frequency linearization compensation and phase noise compensation on the measured beat frequency signal to obtain the compensated measured complex echo includes: Extract the reference phase noise estimate from the reference beat frequency signal; Based on the reference phase noise estimation, the linearized measurement complex echo is subjected to phase rotation compensation at each sampling point to obtain the phase-compensated measurement complex echo.

7. The method of coherent imaging according to claim 4, characterized in that, The step of discretizing the reference phase noise and mapping it to a scanner drift observation includes: Obtain the scan position index and corresponding event stamp recorded by the scanner encoder; By using a mapping operator, the reference phase noise in the continuous time domain is converted into drift observations in the scan index domain.

8. The coherent imaging method according to claim 4, characterized in that, The method of jointly estimating the scanner geometric parameter error and discrete equivalent phase drift error using the compensated measurement echo and the drift observation includes: A target function is constructed based on the compensated measurement echo and the drift observation; The objective function is optimized by using weighted least squares or Kalman filtering to obtain estimates of the scanner's geometric parameter error and discrete equivalent phase drift error.

9. The method of coherent imaging according to claim 8, characterized in that, The objective function includes: The echo consistency term is constructed based on the difference between the model echo and the compensated measured echo. The reference anchor term is constructed based on the difference between the drift observation and the discrete equivalent phase drift error; The smoothed prior terms are constructed based on the difference of the discrete equivalent phase drift error of adjacent scan position indices.

10. The coherent imaging method according to any one of claims 4 to 9, characterized in that, After outputting a coherent 3D point cloud or a coherent detection image, the coherent imaging method includes: Calculate the residual phase error; Based on the residual phase error, update at least one of the following parameters: the scaling factor for mapping the reference phase noise to the measurement branch phase disturbance, the equivalent alignment deviation between the reference branch and the measurement branch, the reference anchoring term weight, the smoothing prior term weight, and the mapping operator used for discretization mapping. The low-frequency component of the frequency sweep linearization error is output to the frequency control terminal of the FMCW laser. The low-frequency drift error during the scanner's movement is extracted, and the low-frequency drift error is output to the scanner's drive terminal.

Images