Calibration method for time-of-flight ranging lidar

Abstract

The application discloses a kind of time-of-flight ranging laser radar's calibration method, it includes: to the receiving end detector of not being equipped with lens is calibrated, using constant power uniform light irradiation, record the charge quantity code value under different modulation phase under fixed integration time, complete calibration;Then, laser radar module is calibrated, including adjusting digital output result close to zero under no light condition, ensure that the same light intensity under pixel output is consistent, eliminate bad point, correct lens distortion, and determine focusing distance;Then, by moving the position of test card in two directions, adjust lens and detector position and inclination, ensure that image definition is consistent and located in field of view center;In addition, using fourier series fitting error between measured distance and actual distance, correction is carried out;Finally, to the second-order linear fitting of measured distance and temperature data, according to the fitting curve, temperature correction is carried out, to improve the ranging accuracy and stability of laser radar module.

Description

Technical Field

[0001] This invention relates to the field of lidar technology, and specifically to a calibration method for a time-of-flight ranging lidar. Background Technology

[0002] i-ToF (Indirect Time-of-Flight) lidar, as a new generation of lidar technology, operates on a significantly different principle from 2D imaging systems. It is a transceiver system that measures distance by emitting laser pulses and receiving the reflected signals. Due to this characteristic, error sources are limited not only at the transmitting end but also at the receiving end. Therefore, to improve the accuracy and precision of ranging, in addition to conventional camera intrinsic, extrinsic, and distortion calibrations, it is also necessary to calibrate parameters such as ranging accuracy, ranging precision, and temperature drift. Currently, only calibration methods for parameter indicators are available, while the calibration process for dual-frequency all-solid-state i-ToF lidar is a complex and meticulous process involving multiple steps and aspects of calibration. Therefore, how to improve the ranging accuracy and precision of lidar through a complete calibration process is a pressing technical problem that needs to be solved. Summary of the Invention

[0003] To overcome the shortcomings of the prior art, the present invention provides a calibration method for a time-of-flight ranging lidar, which can significantly improve ranging accuracy, optimize frequency selection, enhance system adaptability, and reduce testing costs.

[0004] This invention is achieved using the following technical solution:

[0005] A calibration method for a time-of-flight ranging lidar, comprising:

[0006] The receiver detector without a lens is calibrated by illuminating it with uniform light of constant power and obtaining the charge quantity code value of the receiver detector under different modulation phases at a fixed integration time. The calibration of the receiver detector is completed based on the charge quantity code value under different modulation phases.

[0007] The lidar module is calibrated by ensuring that the digital output of each pixel in the receiver detector is close to zero under no-light conditions, and that the output of any two pixels in the receiver detector is consistent under the same incident light intensity. Defective pixels are removed, and the lens distortion file is imported to correct the lens distortion. The position of the light-diffusing plate relative to the lens is moved to the pre-obtained focus position to ensure that the light-diffusing plate is parallel to the optical platform. Then, the focus distance of the lidar module lens is output. The test card is moved in the first and second directions respectively so that the ray from the test card in the third direction passes exactly through the center position of the lidar module lens in the first and second directions respectively. Then, the lens is moved in the third direction to make the obtained image clear. The position of the receiver detector is adjusted so that the image is at the center of the field of view. The tilt of the lens is adjusted so that the image clarity is consistent at the four position points.

[0008] The error between the measured distance and the actual distance is fitted using Fourier series, and the measured distance is corrected based on the function obtained from the fitting.

[0009] A second-order linear fit is performed on the measured distance and temperature data, and the measured distance is temperature-corrected based on the curve obtained from the fit.

[0010] The calibration of the receiver detector is completed by using the charge quantity code values ​​under different modulation phases, including: obtaining the charge quantity code value curves of the same pixel under different modulation phases within the fixed integration time; fitting the charge quantity code value curves with a first-order curve function to minimize the linear deviation between the first-order curve function and the charge quantity code value curves; and traversing all pixels of the receiver detector to obtain the fitting curve coefficients of all pixels.

[0011] The fitting curve coefficients include the equivalent conversion efficiency correction coefficient and the background noise voltage code value. The equivalent conversion efficiency correction coefficient is the ratio of the mode of the equivalent conversion efficiency under different modulation phases at different integration times to the equivalent conversion efficiency under different modulation phases.

[0012] The equivalent conversion efficiency is the ratio of the actual charge code obtained by illuminating the receiver detector without a lens with uniform light through the integrating sphere for an integral time under different modulation phases to the ideal photogenerated electronic code value.

[0013] The background noise voltage code value is the vertical intercept of the charge quantity code value curve.

[0014] The formula for the ideal photogenerated electronic digital value is as follows:

[0015] Among them, P pixel_signal Let λ be the incident light power on the pixel, λ be the wavelength, c be the speed of light, h be Planck's constant, and T be the wavelength.int The time for integration.

[0016] The process of removing defective pixels includes: collecting the charge quantity code values ​​output by all pixels of the receiver detector in a dark environment, calculating the standard deviation and mean of the charge quantity code values, calculating the absolute value of the difference between the charge quantity code values ​​of all pixels and the mean, determining whether the absolute value is greater than a preset threshold, wherein the preset threshold is a multiple of the standard deviation; if so, determining that the pixel is a defective pixel and removing the pixel.

[0017] The test card is moved in the first and second directions respectively, so that the ray from the test card in the third direction passes exactly through the center position of the lens of the lidar module in the first and second directions respectively. Then, the lens is moved in the third direction to make the obtained image clear. The position of the receiver detector is adjusted so that the image is at the center of the field of view. This includes: marking the center position of the test card in the first, second, and third directions using the ray in the third direction of the level; moving the axis of the three-axis displacement stage in the first direction so that the ray in the third direction passes exactly through the center position of the first direction axis of the lens; and fixing the position of the first direction axis of the three-axis displacement stage; marking the center position of the test card in the second direction using the ray in the third direction of the level; moving the axis of the three-axis displacement stage in the second direction so that the ray in the third direction passes exactly through the center position of the second direction axis of the lens; and fixing the position of the second direction axis of the three-axis displacement stage.

[0018] Adjusting the position of the receiver detector so that the image is at the center of the field of view includes: when the image is not at the center of the field of view, moving the receiver detector in the opposite direction to the image's deviation from the center of the field of view.

[0019] Adjusting the tilt of the lens to ensure consistent image sharpness at the four locations includes adjusting the rotation angle of the lens in the opposite direction to the direction of image sharpness.

[0020] The error between the measured distance and the actual distance is fitted using Fourier series. The measured distance is then corrected based on the fitted function, including setting the lidar module under test and the reflectivity plate at a fixed distance and calibrating the fixed distance.

[0021] Two curves, ideal ranging and actual ranging, are obtained at the first demodulation frequency. An error curve between the ideal ranging and actual ranging is obtained, and a Fourier series is used to fit the error curve at the first frequency.

[0022] Based on the unambiguous distance formula of dual-frequency ranging, a second demodulation frequency that meets the conditions is selected. Two curves, ideal ranging and actual ranging, are obtained at the second demodulation frequency. The error curve between ideal ranging and actual ranging is obtained. Fourier series is used to fit the error curve at the second frequency. The second frequency corresponding to the minimum peak value of the error curve after error elimination is obtained as the second frequency of dual-frequency ranging.

[0023] The measurement distance and temperature data are subjected to second-order linear fitting, and the measurement distance is corrected for temperature based on the fitted curve. This includes: placing the lidar module in a temperature chamber with low-reflectivity baffles on all sides; adjusting the temperature of the temperature chamber according to a preset temperature step size to obtain a curve of measurement distance versus temperature; and performing second-order linear fitting on the curve of measurement distance versus temperature.

[0024] This invention provides a dual-frequency ranging error calibration and compensation method for i-ToF (Indirect Time-of-Flight) lidar systems, achieving precise correction of system ranging errors and significantly improving ranging accuracy and stability. The following are the beneficial effects of this invention:

[0025] 1. Improved ranging accuracy: By fitting the residual curve with Fourier series, this invention can accurately identify and compensate for oscillation errors, thereby significantly improving the ranging accuracy.

[0026] 2. Optimized dual-frequency ranging frequency selection: Based on the unambiguous distance formula for dual-frequency ranging, this invention scientifically selects multiple sets of suitable second demodulation frequencies f2, and chooses the optimal dual-frequency ranging frequency by comparing the peak-to-peak values ​​of the residual curves. This method not only simplifies the frequency selection process but also ensures the accuracy and stability of dual-frequency ranging.

[0027] 3. Enhanced system adaptability: The method proposed in this invention is not only applicable to specific i-ToF lidar systems, but can also be widely applied to other similar ranging systems. By adjusting the parameters in the calibration process, it can flexibly adapt to the needs of different systems, improving the versatility and practicality of the method.

[0028] 4. Reduced testing costs: During testing, this invention effectively reduces testing costs by optimizing temperature stabilization time and the number of data acquisitions. Simultaneously, precise temperature control and data acquisition processing ensure the accuracy and reliability of the test results.

[0029] In summary, this invention proposes an efficient, accurate, and universal dual-frequency ranging error calibration method for i-ToF lidar systems, which can significantly improve ranging accuracy, optimize frequency selection, enhance system adaptability, and reduce testing costs. Attached Figure Description

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0031] The term "comprising" and its variations as used herein signify open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "at least partially based on". The terms "one example embodiment" and "one embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0032] Figure 1 A schematic flowchart illustrating the calibration method for a time-of-flight ranging lidar provided in the first embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of the Q-value correction process;

[0034] Figure 3 A schematic diagram showing the number of electrons acquired in a four-phase signal when the duty cycle is inconsistent;

[0035] Figure 4 This is a schematic diagram showing the theoretical positional relationship between the lens, lens mount, and detector chip.

[0036] Figure 5 A schematic diagram illustrating optical axis deviation caused by lens assembly;

[0037] Figure 6 A schematic diagram illustrating assembly deviations between the lens and detector chip;

[0038] Figure 7 Schematic diagram of the lens and detector chip positions after active alignment process;

[0039] Figure 8 This is a schematic diagram of the actual manual AA platform.

[0040] Figure 9 This is a schematic diagram showing the ideal optical axis and the position of the test card;

[0041] Figure 10 This is a schematic diagram of the swing error testing system;

[0042] Figure 11 This is a schematic diagram of a phase delay method data acquisition board;

[0043] Figure 12 This is a schematic diagram comparing the theoretical distance and the measured distance.

[0044] Figure 13 This is a schematic diagram of the error curve;

[0045] Figure 14 This is a schematic diagram of the residual curve after Waggling+FPPN elimination;

[0046] Figure 15 This is a schematic diagram of a temperature calibration test scenario;

[0047] Figure 16 This is a schematic diagram of the measured distance curves before and after temperature correction;

[0048] Figure 17 This is a schematic diagram of the effect of temperature on the measured distance;

[0049] Figure label:

[0050] Lens group 401;

[0051] Optical axis 402;

[0052] Mirror mount 403;

[0053] Focal length 404;

[0054] Depth of focus 405;

[0055] Detector center 406;

[0056] Detector chip 407.

[0057] Optical axis 501;

[0058] Focal length 502;

[0059] Detector center 503;

[0060] Optical axis 601;

[0061] Focal length 602;

[0062] Optical axis 701;

[0063] UV adhesive 702;

[0064] Focal length 703;

[0065] Detector center 704;

[0066] Light homogenizer 801;

[0067] Transmissive ISO12233 resolution test chart 802;

[0068] XYZ three-axis and θ-axis displacement stage 803;

[0069] Infrared light source 804;

[0070] Fixture and hardware board moving stage 805;

[0071] Computer 806;

[0072] Optical vibration isolation stage 807;

[0073] Data transmission cable 808;

[0074] The i-ToF module is labeled as 1901.

[0075] Receive light path 1902;

[0076] Light-absorbing paper tube 1903;

[0077] Incubator 1904;

[0078] Reflector 1905. Detailed Implementation

[0079] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0080] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0081] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0082] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0083] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0084] Please see Figure 1 The first embodiment of the present invention provides a calibration method for a time-of-flight ranging lidar, which can be executed by an electronic device to achieve the following steps:

[0085] S101, calibrate the receiver detector without a lens by illuminating the receiver detector with uniform light of constant power, obtaining the charge quantity code value of the receiver detector under different modulation phases at a fixed integration time, and completing the calibration of the receiver detector based on the charge quantity code value under different modulation phases.

[0086] In this embodiment, i-ToF uses continuous wave modulation and demodulation to obtain the time difference information between the transmitted and returned waves. It calculates the phase information, converts the phase into a time-of-flight variation, and finally calculates the round-trip distance of the light. In practical applications, two sets of sampled signals with a 180° phase difference can be used to demodulate the echo signal; this method is called the two-phase method. There are also schemes that utilize four phase angles—0°, 90°, 180°, and 270°—to obtain the target distance.

[0087] When using multi-phase sampling of signals, the ideal received waveforms for different phases should be square waves with the same duty cycle. However, due to the instability of laser emission power or demodulated signal, the duty cycles of different sampled signals may be inconsistent, which directly affects the inconsistency of the four-phase sampled signals. Furthermore, 2-Tap or multi-Tap pixel structures cannot guarantee equal signal output gain across multiple integration nodes, further exacerbating the inconsistency of the four-phase sampled signals. This inconsistency causes an unavoidable spatial error. Currently, there is no solution to this spatial error. This application proposes a charge quantity code value correction method to effectively suppress the spatial error caused by the inconsistency of duty cycle and multi-integration node signal output gain. For ease of explanation, the charge quantity code value error caused by the inconsistency of the four-phase signals is referred to as the Q-value error in this application.

[0088] In this embodiment, the following is adopted: Figure 2 The method shown corrects for Q-value error. Figure 2 This is a schematic diagram of the Q-value correction process. (Example) Figure 2 As shown, it includes the following steps:

[0089] S201: Calibrate the receiver detector without a lens by illuminating it with uniform light of constant power.

[0090] Specifically, an integrating sphere is an optical device coated with a highly reflective material to ensure that light is uniformly distributed after multiple reflections within the sphere. When a light source (such as an LED or laser) shines into the entrance of the integrating sphere, the light undergoes multiple diffuse reflections within the sphere and finally exits uniformly from the exit or other openings. The LiDAR receiver detector chip is fixed on the test platform, ensuring its surface is clean and horizontal. The integrating sphere is placed above the LiDAR chip, ensuring its exit is in close contact with the photosensitive surface of the chip or at an appropriate distance. The light source inside the integrating sphere is activated, and its intensity and wavelength are adjusted to meet the operating requirements of the LiDAR chip. The adjusting mechanism of the integrating sphere ensures that the light emitted from its exit illuminates the photosensitive surface of the LiDAR chip uniformly and stably.

[0091] S202: Obtain the charge quantity code value of the receiver detector under different modulation phases at a fixed integration time.

[0092] Specifically, timing control is crucial to ensuring the orderly execution of the entire calibration process. Through timing control, the integration process of the LiDAR chip can be triggered sequentially, and the output of the four-phase sampling signal can be read at the end of each integration period. Before the start of each integration cycle, it is necessary to ensure that the LiDAR chip is in a reset state to accurately record the number of photogenerated electrons within each integration cycle. By iterating through different integration times, the response characteristics of the four-phase sampling signal at different integration times can be observed.

[0093] During the traversal, it is necessary to ensure that the integration time setting range is sufficiently wide to cover all lighting conditions that the LiDAR chip may encounter. Simultaneously, the integration time needs to be adjusted gradually to accurately observe the variation of the four-phase sampling signal with the integration time.

[0094] At the end of each integration time period, by reading the output signal of the lidar chip, the number of photogenerated electrons received by the four-phase sampling signals can be obtained as N0, N... 90 N 180 N 270 .

[0095] When the integration time is too long or the light intensity is too high, the four-phase sampling signal may reach saturation, leading to inaccurate data. To avoid signal saturation, the output intensity of the four-phase sampling signal needs to be monitored in real time during calibration, and the integration time or light source intensity needs to be adjusted as needed to prevent saturation.

[0096] S203: The receiver detector is calibrated based on the charge quantity code value under different modulation phases.

[0097] Specifically, the actual incident light power and integration time are substituted into formula (1) to calculate the ideal photogenerated electron digital value, denoted as N. ideal :

[0098]

[0099] Among them, P pixel_signal Let λ be the incident light power on the pixel, λ be the wavelength, c be the speed of light, h be Planck's constant, and T be the wavelength. int The time for integration.

[0100] In the scenario depicting the phase signal curves, the horizontal axis represents the desired number of photogenerated electrons under different integration time conditions, while the vertical axis records the actual number of photogenerated electrons received through four-phase sampling technology. Under this setup, the slope of each curve directly reflects the magnitude of the quantum efficiency (η), which quantifies the efficiency of photon conversion into electrons. Simultaneously, the intercept on the vertical axis corresponds to the system's background noise electron count (N). noise ), which represents the number of electrons generated by a system under conditions of no light due to various factors (such as dark current, thermal noise, etc.).

[0101] Ideally, if the performance of a four-phase sampling system is completely consistent, the signal curves corresponding to the four phases should have the same slope. However, in practical applications, due to various factors such as imbalances in the duty cycles of each phase and differences in charge transfer efficiency, the slopes of the obtained curves may exhibit some fluctuations or inconsistencies. See also Figure 3 , Figure 3 This is a schematic diagram of the number of electrons acquired by the four-phase signal when the duty cycle is inconsistent.

[0102] In order to achieve Q-value error correction and ensure accurate overlap of the four-phase photogenerated electron curves, this embodiment adopts the following steps to correct the curve slope.

[0103] First, according to formula (2), calculate the equivalent conversion efficiency for each phase (0°, 90°, 180°, 270°), denoted as η0, η1, η2, η3, η4, η5, η6, η7, η8, η9, η1, η1, η2, η1, η2, η3, η4, η5, η6, η7, η8, η 90 η 180 η 270These equivalent conversion efficiencies reflect the differences in efficiency between different phases in the process of converting photons into electrons.

[0104]

[0105] Next, the mode of these equivalent conversion efficiencies at different integration times is calculated, i.e., the value with the highest frequency, and is used as the standard equivalent conversion efficiency η for the entire system. This value represents the most likely conversion efficiency of the system under different conditions and is used as a benchmark in the subsequent correction process.

[0106] Then, the ratio of the equivalent conversion efficiency to the standard equivalent conversion efficiency for each phase is calculated to obtain four correction coefficients k0, k... 90 k 180 k 270 These are correction coefficients corresponding to the 0°, 90°, 180°, and 270° phases, respectively. These correction coefficients reflect the conversion efficiency deviation of each phase relative to the system standard.

[0107] Simultaneously, the background electron number of each phase is converted into background quantization code values, denoted as b0, b... 90 b 180 b 270 These background quantization values ​​represent the digital code values ​​corresponding to the number of electrons generated by system noise in each phase under conditions of no light.

[0108] Finally, we use the following formula (3) to correct the digital code values ​​of the four phases:

[0109] D calibre =D actual k n -b n (3)

[0110] Among them, D calibre D is the Q-value corrected digital code value. actual k is the digital code value before correction. n The equivalent conversion efficiency correction coefficients (corresponding to k0, k... for each phase) 90 k 180 k 270 ), b n The background noise voltage code value (corresponding to b0, b1, b2, ..., b3 for each phase) 90 b 180 b 270 ).

[0111] This calibration process effectively eliminates the differences in conversion efficiency between phases and the influence of background noise, enabling the photogenerated electron curves of the four phases to coincide precisely after calibration, thereby improving the measurement accuracy and stability of the system.

[0112] Furthermore, to analyze the relationship between the integration time and output code value of the same pixel for demodulated signals with different phases, a corresponding curve was plotted and labeled as y = f(x). Subsequently, a first-order curve fitting technique was employed, aiming to find a fitting function z = Ax + B that minimizes the linear deviation between z and y. This fitting process was performed on every pixel of the entire chip, and the fitting coefficients were recorded for each pixel.

[0113] S102, calibrate the lidar module. Under no-light conditions, ensure that the digital output of each pixel in the receiver detector is close to zero. Under the condition of the same incident light intensity, ensure that the output of any two pixels in the receiver detector is consistent. Remove bad pixels. Import the lens distortion file to correct the lens distortion. Move the position of the light-diffusing plate away from the lens to the pre-obtained focus position to ensure that the light-diffusing plate is parallel to the optical platform. Then output the focus distance of the lidar module lens. Move the test card in the first direction and the second direction respectively so that the ray of the test card in the third direction passes exactly through the center position of the lidar module lens in the first direction and the second direction respectively. Then move the lens in the third direction to make the obtained image clear. Adjust the position of the receiver detector so that the image is at the center of the field of view. Adjust the tilt of the lens to make the image clarity consistent at the four position points.

[0114] In this embodiment, to optimize the imaging clarity of the lidar module, ensuring optimal alignment of the optical axes of the detector chip and optical components is crucial. Traditional passive calibration methods, such as Zhang's calibration method, are typically performed after the camera module (including the lens) is assembled. This means that errors that may be introduced during lens assembly cannot be adjusted during calibration and can only be compensated for by subsequent internal and external parameter calibration.

[0115] To reduce the impact of assembly errors on the imaging quality of a lidar system at its source, this embodiment proposes a solution using active lens alignment (AA) technology. AA technology can adjust the lens position in real time during lens installation to achieve optimal alignment with the optical axis of the detector chip. This method not only improves the accuracy of alignment but also directly corrects errors generated during assembly, thereby significantly enhancing the imaging clarity and overall performance of the lidar system.

[0116] Specifically, AA (Analog-Assembly) technology aims to precisely determine the relative positions of components during the assembly process. In the complex process of lens packaging, multiple components such as the detector chip, lens mount, PCB (Printed Circuit Board), lens, and circuit board need to be precisely assembled multiple times. As the assembly steps increase, the cumulative tolerances between components gradually accumulate, leading to a larger overall tolerance.

[0117] This accumulation of tolerances can lead to a series of problems, such as the image being off-center or significant differences in sharpness at the four corners of the captured area. These issues not only affect the visual quality of the image but can also negatively impact its resolution and accuracy.

[0118] In Flash-based LiDAR systems, the application of anti-aliasing (AA) technology is equally crucial. LiDAR generates 3D point cloud data by emitting a laser beam and receiving the reflected light signals. When there are deviations between the lens and the detector chip in six degrees of freedom (i.e., three translational directions and three rotational angles), these deviations directly affect the LiDAR's imaging quality and the accuracy of the 3D point cloud.

[0119] Therefore, in the manufacturing process of Flash LiDAR, employing AA (Alignment and Alignment) technology to precisely align the lens and detector chip is crucial. By accurately adjusting their relative positions, deviations can be minimized, thereby improving the imaging quality and accuracy of the 3D point cloud. See details... Figure 4 , Figure 4 This is a schematic diagram illustrating the theoretical positional relationship between the lens, lens mount, and detector chip. (Example) Figure 4 As shown, in the ideal optical path concept, the lens and detector chip should be perfectly parallel, with the optical axis center of the lens precisely overlapping the optical axis center of the detector chip, and the optical axis center maintaining a perfect perpendicular relationship with the detector surface. However, in actual manufacturing, production, and assembly processes, various error factors, such as limitations in processing precision and assembly process deviations, make achieving this ideal state extremely difficult. These errors may lead to slight misalignment between the lens and detector chip, inaccurate alignment of the optical axis centers, and even affect the perpendicularity of the optical axis to the detector. The combined effect of these factors makes it difficult for the final produced camera module to achieve the theoretically perfect state. See details... Figures 5-6 . Figure 5 This is a schematic diagram illustrating optical axis misalignment caused by lens assembly. Figure 5As shown, this illustrates optical axis misalignment during lens assembly. During the stacking and assembly of lens elements, various factors (such as manufacturing precision and assembly processes) can cause the optical axis of the lens to shift, resulting in the optical center of the lens not being precisely aligned with the central axis of the lens. This optical axis misalignment negatively impacts the lens's image quality. Ideally, the optical center of the lens should be precisely located on the central axis of the lens to ensure accurate focusing and a sharp image. However, when the optical axis shifts, light cannot be properly focused, leading to problems such as image blurring and distortion.

[0120] Figure 6 This is a schematic diagram illustrating assembly deviations between the lens and detector chip. Figure 6 The diagram illustrates how, during detector chip assembly, the center of the detector chip shifts relative to the optical center of the lens due to mounting tolerances, and the chip may also develop a certain tilt angle. This displacement and tilt angle directly impact detector performance. Ideally, the center of the detector chip should be precisely aligned with the optical center of the lens to ensure that the received optical signal is converted into a maximum effective electrical signal. However, when displacement and tilt angle occur, some optical signal may deviate from the sensitive area of ​​the detector chip, leading to signal loss and a decrease in image quality.

[0121] Active alignment technology can effectively compensate for tolerances generated during the front-end assembly process in five degrees of freedom. These degrees of freedom include offsets on the X and Y axes, as well as correction of tilt offsets, thereby ensuring precise alignment between the central axis of the lens and the central axis of the detector chip.

[0122] This assembly process is completed while the camera module is lit and operational. First, active alignment technology is used to adjust the image quality of the center point and four corners of the detector chip to ensure optimal sharpness. Once the adjustment is complete, UV adhesive is used to fix the lens in position to ensure its stability and durability.

[0123] The positions of each module after active alignment technology, such as Figure 7 As shown, Figure 7 A schematic diagram showing the positions of the lens and detector chip after active alignment process.

[0124] The calibration process of the lidar module provided in the embodiments of this application will be described in detail below.

[0125] Given the lack of dedicated automated platform equipment, this embodiment employs a manual active alignment (MAA) method in the lidar lens calibration process. The MAA platform integrates several key components to support this precise calibration process. These components primarily include: an infrared light source: used to provide stable infrared illumination, simulating the lidar's operating environment.

[0126] Light-diffusing plate: Ensures uniform light distribution, reducing the impact of uneven light on calibration results. Transmissive ISO12233 resolution test chart: Serves as a standard reference during calibration, used to evaluate lens resolution performance. Three-axis displacement stage: Allows fine-tuning of the lens in the X (first direction), Y (second direction), and Z (third direction) directions for precise alignment. Theta-axis rotary stage: Provides the lens with the freedom of rotation around the optical axis, used to correct angular deviations. Lens mount clamp and hardware board moving stage: Securely holds the lens, ensuring its stability during calibration, used to adjust the position of the detector chip or related components for optimal alignment with the lens. Computer: Acts as the control center, running calibration software and receiving and processing data from the test chart. Data transmission cable: Connects the computer to various hardware components, ensuring real-time data transmission and processing. Optical vibration isolation stage: Ensures the entire platform is level, reducing errors caused by tilt and vibration.

[0127] The layout of the manual AA platform used in this embodiment is as follows: Figure 8 As shown, Figure 8 This is a schematic diagram of the actual manual AA platform. Figure 8 As shown, this platform, by integrating the aforementioned components, provides the necessary accuracy and flexibility for lidar lens calibration.

[0128] The complete active alignment process is divided into two stages. The first stage is grayscale image correction. The main goal of this step is to eliminate any non-ideal factors that may exist in the grayscale image, ensuring that they do not interfere with the accurate judgment of focus sharpness. Through this correction, the focusing process can be ensured to proceed smoothly, laying the foundation for subsequent steps.

[0129] Active calibration technology relies on image sharpness to assess the accuracy of the positional relationship between the lens and the detector chip. To ensure calibration accuracy, the consistency of the LiDAR output image must be compensated. The core of this compensation process lies in acquiring and processing background noise to eliminate the offset between pixels, so that the digital output of each pixel approaches zero under dark conditions.

[0130] The specific steps are as follows: First, in a dark environment, collect the output digital code values ​​of all pixels. Since offset is usually more pronounced between pixel columns, centering is performed on each column of data. This involves calculating the mean of the pixel output for each column and subtracting this mean from the output data of each pixel in that column. This step effectively reduces inconsistencies between columns and improves the accuracy of the image data. The calibrated image shows a significant improvement in uniformity and consistency, which supports accurate subsequent judgments of image sharpness.

[0131] Conversion gain correction ensures that, within the field of view, when two pixels receive incident light of the same intensity, the corresponding pixels on the detector chip output consistent results. To achieve this, conversion gain correction is necessary to eliminate output inconsistencies caused by differences in slope between pixels.

[0132] The specific operation steps are as follows: First, collect the response curves of all pixels in the no-light and saturated light intensity ranges. These response curves reflect the output characteristics of each pixel under different light intensities. Next, normalize the response curve of each pixel so that, under the same light intensity, the output code values ​​of each pixel can cover the same range. The purpose of this step is to ensure that the output results of all pixels remain consistent under the same light intensity conditions.

[0133] Upon completion of this step, a conversion gain correction file will be generated. This file records the normalized correction parameters for each pixel and can be used in subsequent image processing to ensure accurate judgment of image sharpness.

[0134] Before focusing begins, removing defective pixels from the detector chip is a crucial step. Due to factors such as manufacturing processes, the conversion gain of individual pixels may be problematic. If these defective pixels appear near the center of the field of view, they will significantly affect the centering determination of the detector chip.

[0135] To effectively remove these bad pixels, the following specific steps are taken: First, the output code values ​​of all pixels are acquired in a dark environment, and the standard deviation σ and mean μ of these code values ​​are calculated. Next, the absolute value x of the difference between the output code value of each pixel and the mean μ is calculated. If x is greater than a preset threshold, which can be a multiple of σ, such as 3σ, 4σ, etc., then the pixel is identified as a bad pixel and removed from subsequent processing.

[0136] After one round of defective pixel removal, the above steps are repeated to screen the remaining pixels again until the x-value of all pixels is less than a preset threshold. This process ensures that only stable, normal pixels are used for subsequent image processing and focusing.

[0137] Once the defective pixel removal process is complete, a defective pixel correction file will be generated. This file records information on all removed defective pixels and can be used in subsequent image processing and system calibration to ensure the performance and accuracy of the detector chip.

[0138] During the design and manufacturing stages of lenses, barrel distortion or pincushion distortion is often unavoidable. These two types of distortion correspond to the "-" and "+" signs in the lens specifications, respectively. The larger the absolute value of the distortion parameter, the more significant the distortion. Therefore, correcting lens distortion is crucial for achieving accurate images and focusing. Typically, lens manufacturers provide detailed lens distortion parameters at the factory for users to perform subsequent corrections. The presence of distortion severely affects the accurate judgment of the field of view and center position during focusing, greatly complicating the focusing operation. To solve this problem, a lens distortion correction file can be directly imported. This file contains distortion correction parameters and algorithms specific to the lens, effectively eliminating or mitigating distortion.

[0139] The above embodiments are all designed to ensure the uniformity and consistency of the detector array, which is a prerequisite for successful focusing. Specifically, by correcting the grayscale image, eliminating inter-pixel offset, performing conversion gain correction, and removing bad pixels, the stability and accuracy of the detector array can be significantly improved.

[0140] Meanwhile, lens distortion correction is also crucial. Distortion interferes with the accurate judgment of the field of view and center position during focusing, thus affecting the precision and reliability of focusing. Therefore, by importing and applying lens distortion correction files, we can eliminate or reduce distortion, providing more accurate image information for focusing.

[0141] The next step is focal length determination and optical axis alignment. In this stage, the lens focal length is determined based on the corrected grayscale image, and through precise adjustments, the optical axes of the lens and the detector chip are optimally aligned. This step is crucial for achieving optimal positioning between the lens and the detector chip, and is essential for improving the imaging quality and performance of the entire optical system.

[0142] To achieve focusing, the depth of field of the lens must first be calculated. This is a crucial step in determining the object distance for focusing, ensuring that the LiDAR system can produce a clear image within the set ranging range. Simply put, depth of field refers to the distance between the closest and farthest points in a scene that produce a relatively sharp image. The length of the depth of field directly affects the range of the scene that can be rendered in sharp focus: the longer the depth of field, the wider the range of sharpness.

[0143] With a constant focal length and aperture, the depth of field is closely related to the distance between the object and the lens. When the object is closer to the lens, the depth of field is relatively shallow; conversely, when the object is farther from the lens, the depth of field is relatively large. This phenomenon occurs because the size of the image is directly affected by the object distance. When the object is closer to the lens, the image size increases accordingly, resulting in a coarser circle of confusion, thus making the depth of field shallower as the object distance increases. As the distance between the object and the lens increases, the circle of confusion gradually decreases, and therefore the depth of field gradually increases accordingly.

[0144] Foreground depth of field can be expressed as:

[0145] The depth of field behind a lens can be expressed as:

[0146] Lens depth of field (DOF) can be expressed as:

[0147]

[0148] Where L is the shooting distance, δ is the diameter of the circle of confusion, F is the lens aperture value, and f is the lens focal length.

[0149] According to lens imaging theory, an image is ultimately formed by multiple points. Even if the image formed by each point is slightly "blurred," the final image will still be clearly visible. Therefore, the size of the actual allowable "blurred" points is called the allowable circle of confusion. The formula for calculating the diameter δ of the allowable circle of confusion is:

[0150] δ=d / 1730 (7)

[0151] Where d is the diagonal length of the detector chip.

[0152] In short, the depth of field of a lens is directly determined by the position of the focus point. Adjusting the focus point will change the depth of field accordingly. Specifically, by setting a specific focus distance, the foreground depth of field and the background depth of field can be calculated separately. The ranges of these two dimensions together constitute the area of ​​sharp image, which must be greater than or equal to the ranging range of the LiDAR.

[0153] After determining the appropriate focusing distance, the next step is to adjust the relative position of the light-diffusing plate and the lens to achieve the focusing position. During this process, the parallelism of the light-diffusing plate needs to be checked to ensure it remains parallel to the optical platform. This step is crucial because it directly affects the quality and accuracy of the image.

[0154] After completing the above adjustments, the distance between the light-diffusing plate and the detector chip can be further adjusted using the ISO12233 test chart to optimize the imaging effect. Finally, the focusing distance parameters of the module lens will be output.

[0155] In the focusing process, optical axis alignment is a crucial step. First, using the Z-axis ray of a level, the center position of the test piece on the XYZ axis needs to be precisely marked. Next, by adjusting the X-axis of the XYZ three-axis displacement stage, it is ensured that the Z-axis ray accurately passes through the center of the lens's X-axis, and the X-axis is fixed at this position.

[0156] Subsequently, the center position of the test card on the Y-axis was marked again using the Z-direction ray of the level. By adjusting the Y-axis of the XYZ three-axis displacement stage, the Z-direction ray could accurately pass through the center of the lens's Y-axis, and the Y-axis was fixed at this position.

[0157] After the above steps, the center of the lens and the test card were aligned, completing the optical axis alignment. See also Figure 9 , Figure 9 This is a schematic diagram showing the ideal optical axis and the position of the test card.

[0158] After completing the aforementioned steps, first move the lens along the Z-axis until the image is basically clear. If the observed image is not in the center of the field of view at this point, and no matter how the lens is tilted left, right, up, or down along the Z-axis, the image cannot be centered, this usually means that the center of the lens mount is not aligned with the center of the detector chip.

[0159] In this situation, the positional relationship between the lens mount and the detector chip needs to be readjusted. Specifically, if the test card center is consistently located to the left of the image, the detector chip needs to be adjusted to the left because the chip's output data is reversed in the left-right direction; if the test card center is to the right, the chip needs to be adjusted to the right. Similarly, if the test card center is consistently located to the top of the image, the detector chip needs to be adjusted downwards because the chip's output data in the vertical direction is correct; if the test card center is to the bottom, the chip needs to be adjusted upwards.

[0160] During the adjustment process, it is necessary to continuously observe whether the image has moved to the center of the field of view. If the image center cannot be aligned with the center of the screen without moving the XYZ three-axis displacement stage (for example, considering that the chip resolution is 320×240, the center pixel should be located at the 120th pixel in the row direction and the 160th pixel in the column direction), then the current focusing material is considered unsuitable, and a new material needs to be replaced and the focusing process repeated until the image center is aligned with the center of the screen.

[0161] Next, adjust the rotation angle to ensure consistent image sharpness at all four positions (top, bottom, left, and right). If the left side of the image is sharp, tilt the lens to the right; if the right side is sharp, tilt it to the left. Similarly, if the top of the image is sharp, tilt the lens downwards; if the bottom of the image is sharp, tilt it upwards.

[0162] After completing the above adjustments, the lens mount and the center of the detector chip are aligned, which means the optical axis alignment is complete. At this point, the positional relationship between the lens mount and the detector chip needs to be fixed, and the position and rotation angle θ of the XYZ three-axis displacement stage should be recorded. Finally, the lens is fixed with adhesive to ensure the stability and durability of the focusing result.

[0163] S103, use Fourier series to fit the error between the measured distance and the actual distance, and correct the measured distance based on the function obtained from the fitting.

[0164] Specifically, the analysis in the above embodiments shows that known non-ideal modulation waveforms introduce nonlinear errors into the final ranging results; these errors are called waggling errors. On the other hand, Fixed Phase Pattern Noise (FPPN) is a phase shift unique to each pixel, originating from the pixel's specific position in the array, and is therefore also known as pixel depth non-uniformity. FPPN is mainly affected by the modulation frequency and operating temperature, and exists as a fixed deviation, making it easy to correct through calibration.

[0165] To calibrate the Waggling error and FPPN, a look-up table method is often used. In practice, the lidar system needs to be securely mounted on a guide rail, and the reflectivity plate needs to be placed within the lidar's field of view. Then, by setting the guide rail's movement step size, a series of measurements are performed to construct the look-up table, thereby achieving accurate calibration of these two errors.

[0166] Specifically, after completing all actual distance measurements, a comparison curve between the actual measured distance and the ideal distance can be plotted, and an error curve can be generated based on the error values. Next, by recording the error values ​​of all sampling points, a lookup table is constructed, and the calibrated data is calculated using formula (8):

[0167] D calibre =D measure +D offset (8)

[0168] It should be noted that the correction accuracy of lookup tables typically increases with the number of sampling points, but this also increases storage space consumption. Therefore, in practical applications, a balance needs to be found between storage depth and system accuracy.

[0169] By observing the principle of error generation and the actual error curve, it can be found that the error curve exhibits periodic changes and its shape is similar to a sine wave. In order to describe this error more effectively, this embodiment uses the Fourier series fitting method to fit the error curve, and the specific formula is shown in (9):

[0170]

[0171] In this formula, the constant term A0 represents the correction parameter for FPPN error, while a n b n ω and ω are the correction parameters for the Waggling error. Once the order of fitting n is determined, it is only necessary to solve for A0, a n b n With these unknowns, ω, the error curve can be fully plotted. This method significantly reduces the storage requirements for correction data while maintaining accuracy.

[0172] Furthermore, when collecting data at different distances using the guide rail, a challenge arises: the echo signal energy is weaker at longer distances, which may increase accuracy errors. Analysis of the system's ranging accuracy shows that weakened echo energy exacerbates accuracy errors. Therefore, when correcting for oscillation errors, it is necessary to eliminate interference from accuracy errors at different ranging distances. While this error can be reduced by collecting a large amount of data at the same distance and averaging it, this decreases calibration efficiency. To address this issue, this embodiment proposes a phase delay method. By adjusting the phase changes of the transmitted wave and the echo, the distance variation is simulated, thereby decoupling accuracy errors from precision errors. This method ensures both calibration accuracy and improved calibration efficiency. (See figure) Figure 10 and Figure 11 , Figure 10 This is a schematic diagram of the swing error testing system. Figure 11 This is a schematic diagram of a phase delay method data acquisition board.

[0173] like Figure 10 As shown, at the start of the test, the distance between the module under test and the reflectivity plate is first precisely fixed at a relatively close position (e.g., 1 meter) using an electrically controlled guide rail, and this distance is precisely calibrated using a single-point rangefinder. After determining the fixed test distance, a suitable modulation frequency f (first frequency) needs to be selected. Next, by configuring the FPGA (Field Programmable Gate Array) on the phase acquisition board, the phase shift of the transmitted modulation waveform can be realized, with a phase shift accuracy of t. delay This step is crucial because it allows us to accurately measure and obtain the total 1 / f1*t at the demodulation frequency f1. delay Data points.

[0174] After data collection is complete, two curves can be plotted based on this data: one is the ideal distance measurement curve, and the other is the actual distance measurement curve. A comparison of these two curves is provided. Figure 12 As shown, Figure 12This is a schematic diagram comparing the theoretical distance and the measured distance. An error curve can be plotted by subtracting the measured distance from the theoretical distance, as shown below. Figure 13 As shown, Figure 13 This is a schematic diagram of the error curve. In this embodiment, a Fourier series was used to fit the residual curve. By observing the fitted graph, it can be seen that the fitting effect is quite ideal when the Fourier series reaches level 7. After correcting the oscillation error using the fitted curve of level 7 Fourier series, the residual curve was redrawn, as shown below. Figure 14 As shown. Figure 14 This is a schematic diagram of the residual curve after Waggling+FPPN elimination.

[0175] After completing the linearity and error calibration of the first demodulation frequency f1, multiple sets of second demodulation frequencies f2 that meet the conditions are selected according to the unambiguous distance formula (15) for dual-frequency ranging. The formula is as follows:

[0176]

[0177] Where f1 and f2 represent the two modulation frequencies in dual-frequency ranging, f E For the equivalent modulation frequency, D E For equivalent fuzzy distance, gcd represents finding the greatest common divisor of f1 and f2.

[0178] For each f2 that meets the conditions, the above calibration process is repeated, and the residual curve is redrawn after the oscillation error is eliminated. We record the peak-to-peak value of each residual curve, denoted as E. 1PP E 2PP ,…,E nPP Finally, from these candidate f2 frequencies, the peak-to-peak value E is selected. pp The smallest one is used as the final dual-frequency ranging frequency to ensure the accuracy and stability of ranging.

[0179] S104, perform second-order linear fitting on the measured distance and temperature data, and perform temperature correction on the measured distance based on the curve obtained from the fitting.

[0180] Specifically, ambient temperature directly affects the laser's emission power, which in turn indirectly affects the measurement distance of i-ToF technology; this phenomenon is known as thermal drift or temperature drift. Fortunately, temperature drift is linear and has a consistent effect on all pixels, but it varies with different modulation frequencies. To compensate for temperature-induced drift, it is first necessary to fully characterize the system's temperature characteristics and then implement precise correction measures for ranging errors accordingly.

[0181] To quantify the drift of distance measurements with temperature variations, i-ToF lidar systems are typically tested in a temperature chamber. However, the limited space and predominantly metal walls of the chamber make it highly susceptible to multipath interference. To address this, low-reflectivity baffles are often installed around the chamber. These baffles cover the entire path from the camera lens to the target object, effectively suppressing multipath propagation. The test environment setup is as follows: Figure 15 As shown, Figure 15 This is a schematic diagram of a temperature calibration test scenario. Since temperature drift has the same effect on each pixel, collecting pixel data from only the central region of the image is sufficient to meet the test requirements.

[0182] After setting up the test environment, the ambient temperature can be adjusted within the target working area, and the distance information of the target object can be recorded simultaneously. For example, the ambient temperature of the system under test can be controlled within the range of -40℃ to 85℃. During this process, considering the significant impact of temperature stabilization time on test reliability, a balance needs to be found between test efficiency and cost. If the temperature stabilization time is too short, it may lead to inaccurate temperature measurements, thus affecting the accuracy of the calibration results; while if the time is too long, it will increase the total calibration time required and increase test costs.

[0183] To improve efficiency while ensuring test accuracy, this embodiment uses a preset temperature step size of 20°C for temperature adjustment, and sets the stabilization time for each temperature group to 20 minutes. After the temperature reaches a stable state, more than 30 sets of data are collected at the current temperature. By averaging these data, the impact of measurement errors on the collected data can be further reduced, thereby improving the reliability of the data. The final test data has been compiled and is shown in Table (1):

[0184] Table 1. Test data at different temperatures

[0185]

[0186] A curve showing temperature versus measured distance was plotted using measured data, as shown below. Figure 16 As shown, Figure 16 This is a schematic diagram of the measured distance curves before and after temperature correction.

[0187] The expression for a second-order linear fit on the data is:

[0188] D Temperature_offset = a + bT + cT 2 (0-1)

[0189] Where a, b, and c are temperature correction coefficients, and T is the temperature. After solving for a, b, and c using the least squares method, the error curves before and after temperature correction can be plotted as follows: Figure 17 As shown, Figure 17 This is a schematic diagram of the effect of temperature on the measured distance. The data shows that within the temperature range of -40℃ to 85℃, after temperature correction, the accuracy deviation caused by temperature drift is corrected to within 4mm.

[0190] In summary, compared with the prior art, this embodiment has the following advantages:

[0191] 1. Improved ranging accuracy: By fitting the residual curve with Fourier series, this invention can accurately identify and compensate for oscillation errors, thereby significantly improving the ranging accuracy.

[0192] 2. Optimized dual-frequency ranging frequency selection: Based on the unambiguous distance formula for dual-frequency ranging, this invention scientifically selects multiple sets of suitable second demodulation frequencies f2, and chooses the optimal dual-frequency ranging frequency by comparing the peak-to-peak values ​​of the residual curves. This method not only simplifies the frequency selection process but also ensures the accuracy and stability of dual-frequency ranging.

[0193] 3. Enhanced system adaptability: The method proposed in this embodiment is not only applicable to specific i-ToF lidar systems, but can also be widely applied to other similar ranging systems. By adjusting the parameters in the calibration process, it can flexibly adapt to the needs of different systems, improving the versatility and practicality of the method.

[0194] 4. Reduced testing costs: During the testing process, this embodiment effectively reduced testing costs by optimizing temperature stabilization time and the number of data acquisitions. Simultaneously, precise temperature control and data acquisition processing ensured the accuracy and reliability of the test results.

[0195] In summary, this embodiment proposes an efficient, accurate, and universal method for calibrating the dual-frequency ranging error of an i-ToF lidar system. This method can significantly improve ranging accuracy, optimize frequency selection, enhance system adaptability, and reduce testing costs.

[0196] Those skilled in the art will understand that the embodiments described in this specification can be provided as methods, systems, or computer program products. Therefore, those skilled in the art will realize that the functional modules / units or controllers and related method steps described in the above embodiments can be implemented in software, hardware, or a combination of both.

[0197] Unless explicitly stated otherwise, the actions or steps of the methods and procedures described in the embodiments of the present invention do not necessarily have to be performed in a specific order and can still achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0198] Several embodiments of the present invention have been described, but for the sake of brevity, the descriptions of the embodiments are not exhaustive, and identical or similar features or parts between the embodiments may be omitted. In this document, "one embodiment," "some embodiments," "example," "specific example," or "some examples" refers to embodiments applicable to at least one, but not all, of the present invention. The above terms do not necessarily refer to the same embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described herein, as well as the features of the different embodiments or examples.

[0199] The exemplary systems and methods of the present invention have been specifically shown and described with reference to the foregoing embodiments, which are merely examples of the best mode for implementing the systems and methods. Those skilled in the art will understand that various changes can be made to the embodiments of the systems and methods described herein without departing from the spirit and scope of the invention as defined in the appended claims when implementing the systems and / or methods.

Claims

1. A calibration method for a time-of-flight ranging lidar, characterized in that, include: The receiver detector without a lens is calibrated by illuminating it with uniform light of constant power and obtaining the charge quantity code value of the receiver detector under different modulation phases at a fixed integration time. The calibration of the receiver detector is completed based on the charge quantity code value under different modulation phases. The lidar module is calibrated by ensuring that the digital output of each pixel in the receiver detector is close to zero under no-light conditions, and that the output of any two pixels in the receiver detector is consistent under the same incident light intensity. Defective pixels are removed, and the lens distortion file is imported to correct the lens distortion. The position of the light-diffusing plate relative to the lens is moved to the pre-obtained focus position to ensure that the light-diffusing plate is parallel to the optical platform. Then, the focus distance of the lidar module lens is output. The test card is moved in the first and second directions respectively so that the ray from the test card in the third direction passes exactly through the center position of the lidar module lens in the first and second directions respectively. Then, the lens is moved in the third direction to make the obtained image clear. The position of the receiver detector is adjusted so that the image is at the center of the field of view. The tilt of the lens is adjusted so that the image clarity is consistent at the four position points. The error between the measured distance and the actual distance is fitted using Fourier series, and the measured distance is corrected based on the function obtained from the fitting. A second-order linear fit is performed on the measured distance and temperature data, and the measured distance is temperature-corrected based on the curve obtained from the fit.

2. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, The calibration of the receiver detector is completed based on the charge quantity code values ​​under different modulation phases, including: obtaining the charge quantity code value curves of the same pixel under different modulation phases within a fixed integration time; fitting the charge quantity code value curves with a first-order curve function to minimize the linear deviation between the first-order curve function and the charge quantity code value curves; and traversing all pixels of the receiver detector to obtain the fitting curve coefficients of all pixels.

3. The calibration method for a time-of-flight ranging lidar according to claim 2, characterized in that, The fitting curve coefficients include the equivalent conversion efficiency correction coefficient and the background noise voltage code value. The equivalent conversion efficiency correction coefficient is the ratio of the mode of the equivalent conversion efficiency under different modulation phases at different integration times to the equivalent conversion efficiency under different modulation phases. The equivalent conversion efficiency is the ratio of the actual charge code obtained by illuminating the receiver detector without a lens with uniform light through the integrating sphere for an integral time under different modulation phases to the ideal photogenerated electronic code value. The background noise voltage code value is the vertical intercept of the charge quantity code value curve.

4. The calibration method for a time-of-flight ranging lidar according to claim 3, characterized in that, The formula for the ideal photogenerated electronic digital value is as follows: Among them, P pixel_signal Let λ be the incident light power on the pixel, λ be the wavelength, c be the speed of light, h be Planck's constant, and T be the wavelength. int The time for integration.

5. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, The process of removing bad pixels includes: collecting the charge quantity code values ​​output by all pixels of the receiver detector in a dark environment, calculating the standard deviation and mean of the charge quantity code values, calculating the absolute value of the difference between the charge quantity code values ​​of all pixels and the mean, and determining whether the absolute value is greater than a preset threshold, wherein the preset threshold is a multiple of the standard deviation. If so, the pixel is determined to be a bad pixel and is removed.

6. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, The test card is moved in the first and second directions respectively so that the ray of the test card in the third direction passes exactly through the center position of the lens of the lidar module in the first and second directions respectively. Then, the lens is moved in the third direction to make the obtained image clear. The position of the receiver detector is adjusted so that the image is at the center position of the field of view. This includes: marking the center position of the axis of the test card in the first, second and third directions with the ray of the third direction of the level; moving the axis of the three-axis displacement stage in the first direction so that the ray of the third direction passes exactly through the center position of the first direction axis of the lens; and fixing the position of the first direction axis of the three-axis displacement stage. Mark the center position of the second direction axis of the test card with the third direction ray of the level. Move the second direction axis of the three-axis displacement stage so that the third direction ray passes exactly through the center position of the second direction axis of the lens. Fix the position of the second direction axis of the three-axis displacement stage.

7. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, Adjusting the position of the receiver detector so that the image is at the center of the field of view includes: when the image is not at the center of the field of view, moving the receiver detector in the opposite direction to the image's deviation from the center of the field of view.

8. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, Adjusting the tilt of the lens to ensure consistent image sharpness at four locations includes adjusting the rotation angle of the lens in the opposite direction to the direction of image sharpness.

9. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, The error between the measured distance and the actual distance is fitted using Fourier series. The measured distance is then corrected based on the fitted function, including setting the lidar module under test and the reflectivity plate at a fixed distance and calibrating the fixed distance. Two curves, ideal ranging and actual ranging, are obtained at the first demodulation frequency. An error curve between the ideal ranging and actual ranging is obtained, and a Fourier series is used to fit the error curve at the first frequency. Based on the unambiguous distance formula of dual-frequency ranging, a second demodulation frequency that meets the conditions is selected. Two curves, ideal ranging and actual ranging, are obtained at the second demodulation frequency. The error curve between ideal ranging and actual ranging is obtained. Fourier series is used to fit the error curve at the second frequency. The second frequency corresponding to the minimum peak value of the error curve after error elimination is obtained as the second frequency of dual-frequency ranging.

10. The calibration method for a time-of-flight ranging lidar according to claim 1, characterized in that, The measurement distance and temperature data are subjected to second-order linear fitting, and the measurement distance is corrected for temperature based on the fitted curve. This includes: placing the lidar module in a temperature chamber with low-reflectivity baffles on all sides; adjusting the temperature of the temperature chamber according to a preset temperature step size to obtain a curve of measurement distance versus temperature; and performing second-order linear fitting on the curve of measurement distance versus temperature.

Images