Three-dimensional laser radar point cloud acquisition control method and system

Abstract

The invention discloses a three-dimensional laser radar point cloud acquisition control method and system, and belongs to the technical field of laser radar measurement and control. An existing low-cost laser radar point cloud acquisition method is large in angle accumulative error and cannot meet the real-time angle tracking requirement under the high-rotating-speed working condition. According to the three-dimensional laser radar point cloud acquisition control method, the coded disc design optimization model and the abrupt change signal capture model are established, the notch is designed in the coded disc of the three-dimensional laser radar, and the width of the notch of the coded disc is optimized, so that pulse signal period abrupt change can be accurately captured, angle accumulation errors can be effectively reduced, and the accuracy of the point cloud acquisition control is improved. And the angle synchronization precision is improved, and the method is especially suitable for real-time angle tracking requirements in a high-rotating-speed scene. Meanwhile, jitter interference of the laser radar can be eliminated, and spatio-temporal interpolation fusion can be performed on multi-circle acquired data, so that high-density point cloud data can be generated, and the angular resolution and the point cloud density can be effectively improved.

Description

Technical Field

[0001] The present invention relates to a method and system for controlling 3D lidar point cloud acquisition, belonging to the technical field of lidar measurement and control technology. Background Art

[0002] Chinese Patent Application (Publication No.: CN114627374A) discloses a point cloud acquisition system based on a lidar and a pan-tilt head and an insulator recognition and positioning method. The point cloud acquisition system includes a lidar, a computer, and a pan-tilt head. The lidar is vertically placed at the center of the pan-tilt head; the computer is connected to the lidar and the pan-tilt head respectively; the computer drives the lidar to scan through ROS, and at the same time, the computer drives the pan-tilt head to rotate at a constant speed at a fixed angle, so as to realize the scanning of the entire scene; after scanning, the lidar transmits the scanned point cloud data to the computer through the Ethernet protocol for subsequent processing.

[0003] The above solution is a low-cost lidar point cloud acquisition method. To save costs, it adopts a fixed-angle sampling strategy, which leads to a large angular cumulative error and insufficient angular synchronization accuracy. In particular, it cannot meet the real-time angle tracking requirements in high-speed rotation (greater than 10,000 RPM) scenarios.

[0004] Furthermore, repeating fixed-angle sampling in multi-circle scanning will cause the point cloud density of distant targets to decrease exponentially with the increase of distance, resulting in a sparse point cloud, which in turn affects the popularization and use of 3D lidars in low-cost solutions.

[0005] The information disclosed in this background art is only used to understand the background of the inventive concept, so it may include information that does not constitute prior art. Summary of the Invention

[0006] Aiming at the above problems or one of the above problems, the first object of the present invention is to provide a method and system for controlling 3D lidar point cloud acquisition. By creating a code disk design optimization model and a mutation signal capture model, notches are designed on the code disk of the 3D lidar, and the width of the notches on the code disk is optimized, so as to accurately capture the mutation of the pulse signal period and obtain the zero position information, thereby effectively reducing the angular cumulative error and improving the angular synchronization accuracy, especially suitable for the real-time angle tracking requirements in high-speed rotation scenarios; at the same time, by constructing a jitter interference elimination model, a sampling dynamic adjustment model, and a spatio-temporal interpolation fusion model, the jitter interference of the lidar is eliminated, the angular resolution of the lidar is calibrated, and through a multi-circle scanning accumulation mechanism, the spatio-temporal interpolation fusion of multi-circle acquisition data is performed to generate high-density point cloud data, thereby effectively improving the angular resolution and point cloud density, avoiding the exponential decrease of the point cloud density of distant targets with the increase of distance, and facilitating the popularization and use of 3D lidars in low-cost solutions.

[0007] In view of the above problems or one of the above problems, the second object of the present invention is to provide a three-dimensional lidar point cloud acquisition control method and system, which can achieve high-precision angle measurement without using a high-precision angle encoder and a high-speed signal processing unit, and can ensure the real-time performance of the lidar measurement system; at the same time, through frequency division interpolation, the angular resolution is effectively improved, and the point cloud spacing of distant targets is reduced, so that the hardware cost can be effectively reduced, meeting the needs of consumer robots for low-cost 3D lidars.

[0008] To achieve one of the above objects, the first technical solution of the present invention is: A three-dimensional lidar point cloud acquisition control method, comprising the following steps: Step 1, through a pre-created code disk design optimization model, based on the demand information of the zero-point marking of the three-dimensional lidar, design a notch on the code disk of the three-dimensional lidar, and optimize the width of the notch of the code disk to obtain notch design data; Step 2, using a pre-created mutation signal capture model, based on the notch design data, capture the mutation of the pulse signal period caused by the notch to obtain zero-point position information; Step 3, based on a pre-created jitter interference elimination model, according to the zero-point position information and combined with a filtering algorithm, obtain jitter interference elimination data for eliminating the jitter interference of the lidar; Step 4, using a pre-created sampling dynamic adjustment model, calibrate the angular resolution of the lidar that has eliminated the jitter interference to obtain sampling time adjustment information; Step 5, adopt a pre-created spatio-temporal interpolation fusion model, based on the sampling time adjustment information, and through a multi-turn scanning accumulation mechanism, perform spatio-temporal interpolation fusion on the multi-turn acquisition data to generate high-density point cloud data.

[0009] Through continuous exploration and experiments, the present invention designs a notch on the code disk of the three-dimensional lidar by creating a code disk design optimization model and a mutation signal capture model, and optimizes the width of the notch of the code disk, so as to accurately capture the mutation of the pulse signal period and obtain zero-point position information, thus effectively reducing the angle accumulation error and improving the angle synchronization accuracy, which is especially suitable for the real-time angle tracking requirements in high-speed rotation (greater than 10,000 RPM) scenarios; at the same time, by constructing a jitter interference elimination model, a sampling dynamic adjustment model and a spatio-temporal interpolation fusion model, the jitter interference of the lidar is eliminated, the angular resolution of the lidar is calibrated, and through a multi-turn scanning accumulation mechanism, spatio-temporal interpolation fusion is performed on the multi-turn acquisition data to generate high-density point cloud data, thereby effectively improving the angular resolution and point cloud density, avoiding the situation that the point cloud density of distant targets decreases exponentially with the increase of distance, and facilitating the popularization and use of three-dimensional lidars in low-cost solutions.

[0010] Furthermore, by optimizing the code disk design, the present invention can accurately capture the sudden change of the pulse signal period and achieve dynamic compensation of the motor speed. At the same time, by using the point cloud spatio-temporal interpolation algorithm, the point cloud density is increased, so it is applicable to the angle synchronization calibration and point cloud resolution enhancement scenarios under high-speed working conditions.

[0011] Moreover, the present invention can achieve high-precision and low-cost angle measurement without using a high-precision angle encoder (such as a 2048-line magnetic encoder) and a high-speed signal processing unit (FPGA), and can ensure the real-time performance of the radar measurement system. At the same time, through frequency division interpolation, the angular resolution is effectively improved, and the point cloud spacing of distant targets is reduced, so the hardware cost can be effectively reduced to meet the requirements of consumer robots for low-cost 3D radars.

[0012] As a preferred technical measure: Step 1: Based on the demand information of the zero-point marking of the 3D lidar, design a notch on the code disk of the 3D lidar through the pre-created code disk design optimization model, and optimize the width of the notch on the code disk to obtain the notch design data, which is as follows: Obtain the number of teeth and tooth position information of the code disk; Calculate several equally divided positions of the code disk according to the number of teeth and tooth position information; Take one or more of the equally divided positions as the position of the notch; Collect the demand information of the code disk zero-point marking, which includes the design demand information and signal detection requirement information of the code disk; Quantify the design demand information and signal detection requirement information to obtain several variable parameters, and construct multi-parameter constraint information; Based on several variable parameters and multi-parameter constraint information, construct an objective function for optimizing the width of the notch on the code disk; solve the objective function to obtain the optimized notch width value; Summarize the position and width value of the notch to obtain the notch design data.

[0013] As a preferred technical measure: The method for quantifying the design demand information and signal detection requirement information to obtain several variable parameters is as follows: Collect the design demand information, which includes mechanical parameters and dynamic parameters; Obtain the signal detection requirement information, which includes optical parameters and environmental parameters; Quantify the mechanical parameters to obtain the code disk diameter, total number of teeth and standard number of teeth; Quantify the dynamic parameters to obtain the maximum working speed and sensor response time; Quantify the optical parameters to obtain the light source wavelength and spot diameter; Quantify the environmental parameters to obtain the signal-to-noise ratio threshold.

[0014] As a preferred technical measure: The method for constructing multi-parameter constraint information is as follows: According to the notch design requirements, set the spot opening angle and the angle drift compensation value caused by the rotation speed; Based on the spot opening angle and the angle drift compensation value, establish a minimum detection angle constraint so that the notch can cover the projection of the spot in the tangential direction of the code disk; Obtain the maximum torque of the code disk and the material yield strength; According to the maximum torque and the material yield strength, establish a mechanical strength safety constraint so that the notch width can meet the requirements of the code disk's torsional strength; Obtain the linear velocity of the code disk and the jitter tolerance of the optical sensor; Based on the linear velocity of the code disk and the jitter tolerance of the optical sensor, establish a signal width constraint so that the notch pulse can be distinguished from the normal tooth pulse; Summarize the minimum detection angle constraint, the mechanical strength safety constraint and the signal width constraint to obtain the multi-parameter constraint information.

[0015] As a preferred technical measure: The method for constructing the objective function is as follows: Obtain the multi-parameter constraint information, which includes the minimum detection angle constraint, the mechanical strength safety constraint and the signal width constraint; According to the mechanical strength safety constraint, the signal width constraint and the standard number of teeth, construct a design reference expression; Based on the minimum detection angle constraint, the code disk diameter and the optical sensitivity weight, construct an expression for minimizing the missed detection risk; According to the material yield strength and the material safety factor, construct an expression for suppressing stress concentration; Couple the design reference expression, the expression for minimizing the missed detection risk and the expression for suppressing stress concentration to obtain the optimized objective function.

[0016] As a preferred technical measure: Step 2, use the previously created mutant signal capture model, based on the notch design data, to capture the pulse signal period mutation caused by the notch, and the method for obtaining the zero position information is as follows: According to the signal attenuation situation under different working conditions and based on the rotation speed, adjust the pulse interval threshold in real time; Based on the pulse interval threshold, obtain the pulse signal output by the photoelectric encoder; Based on the notch design data, capture the pulse signal period mutation caused by the notch through the edge detection circuit to obtain the edge signal; Determine the zero position according to the edge signal and in combination with the pulse interval threshold to obtain zero position information; The method for real-time adjusting the pulse interval threshold is as follows: Capture the rising edge signal output by the optoelectronic encoder; Identify the effective edge time of the rising edge signal through a digital comparator to obtain a timestamp; Based on the output signal of the optoelectronic encoder, obtain the current linear velocity; According to the current linear velocity, the minimum resolution distance, the acceleration compensation factor, and the timestamp, construct a dynamic calculation formula for the pulse interval; solve the dynamic calculation formula for the pulse interval to obtain the pulse interval threshold.

[0017] As a preferred technical measure: Step three, based on the previously created jitter interference cancellation model, according to the zero position information and in combination with a filtering algorithm, the method for obtaining jitter interference cancellation data is as follows: Based on the real-time rotational speed, the number of lines of the code disk, and the process tolerance, predict the rotational speed time sequence and calculate the expected arrival time of the pulse: According to the expected arrival time of the pulse and the actual pulse time, measure the time sequence deviation and calculate the predicted value deviation: According to the filtering algorithm, use the predicted value deviation, the time sequence deviation sensitivity coefficient, the current signal-to-noise ratio, and the rotational speed change rate suppression factor to establish a multi-factor fusion decision formula; Solve the multi-factor fusion decision formula to obtain a dynamic confidence value; According to the confidence interval and the dynamic confidence value, set a filtering decision mechanism; When the dynamic confidence value is greater than or equal to the confidence threshold, directly output the pulse signal; When the dynamic confidence value is less than the confidence threshold, perform counter debounce on the pulse signal to obtain jitter interference cancellation data.

[0018] As a preferred technical measure: Step four, using the previously created sampling dynamic adjustment model, perform angular resolution calibration on the lidar that has already eliminated jitter interference to obtain the sampling time adjustment information, and the method is as follows: Obtain the code disk pulse signal output by the optoelectronic encoder; After shaping the code disk pulse signal using a Schmitt trigger, trigger an external interrupt of the microcontroller to obtain the time interval between adjacent code disk signals; Based on the time interval, calculate the real-time rotational speed of the 3D lidar; According to the real-time rotational speed, calculate the corresponding angular velocity; Based on the target angular resolution, calculate the radian resolution; Calculate the sampling time according to the relationship between the angular velocity and the radian resolution; Process the sampling time, and based on the timer clock cycle and the prescaler coefficient, obtain the count value of the timer; According to the count value of the timer, update the overflow time of the timer interrupt in real time, so as to determine the sampling time adjustment information.

[0019] As a preferred technical measure: Step five, adopt the pre-created spatio-temporal interpolation fusion model, based on the sampling time adjustment information, and through the multi-turn scanning accumulation mechanism, the method for spatio-temporal interpolation fusion of multi-turn collected data to generate high-density point cloud data is as follows: Obtain the sampling time adjustment information, which includes the basic sampling period; Calculate the required angular resolution according to the point cloud spacing requirement; Based on the angular resolution, calculate the number of sampling points per turn; Determine the required division factor according to the number of sampling points per turn; Based on the division factor, divide the basic sampling period to generate multiple fine trigger moments; Through the multi-turn scanning accumulation mechanism, and according to multiple fine trigger moments, simultaneously superimpose the division offset, and set the ranging trigger moment for each turn of scanning to cover multiple fine positions; Based on the ranging trigger moment, perform several turns of scanning so that all fine positions are completely covered to obtain multi-turn collected data; perform spatio-temporal interpolation fusion on the multi-turn data to generate high-density point cloud data.

[0020] To achieve one of the above purposes, the second technical solution of the present invention is: A three-dimensional lidar point cloud acquisition control system, which includes: One or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned three-dimensional lidar point cloud acquisition control method.

[0021] Compared with the prior art solution, the present invention has the following beneficial effects: Through continuous exploration and experimentation, the present invention designs notches on the code disk of a 3D lidar by creating an optimized model for code disk design and a model for capturing mutation signals, and optimizes the width of the notches on the code disk, so as to accurately capture the mutation of the pulse signal period, obtain the zero position information, and thus can effectively reduce the angle accumulation error and improve the angle synchronization accuracy, which is particularly suitable for the real-time angle tracking requirements in high-speed rotation (greater than 10,000 RPM) scenarios; at the same time, by constructing a jitter interference elimination model, a sampling dynamic adjustment model and a spatio-temporal interpolation fusion model, the jitter interference of the lidar is eliminated, the angular resolution of the lidar is calibrated, and through a multi-turn scanning accumulation mechanism, the spatio-temporal interpolation fusion of the multi-turn collected data is performed to generate high-density point cloud data, so as to effectively improve the angular resolution and the point cloud density, and avoid the exponential decrease of the point cloud density of distant targets with the increase of distance, which is convenient for the popularization and use of 3D lidars in low-cost solutions.

[0022] Furthermore, by optimizing the code disk design, the present invention can accurately capture the mutation of the pulse signal period and achieve dynamic compensation of the motor speed; at the same time, by using the point cloud spatio-temporal interpolation algorithm, the point cloud density is improved, so it is applicable to the scenarios of angle synchronization calibration and point cloud resolution enhancement under high-speed operating conditions.

[0023] Moreover, the present invention can achieve high-precision and low-cost angle measurement without using a high-precision angle encoder (such as a 2048-line magnetic encoder) and a high-speed signal processing unit (FPGA), and can ensure the real-time performance of the radar measurement system; at the same time, through frequency division interpolation, the angular resolution is effectively improved and the point cloud spacing of distant targets is reduced, so the hardware cost can be effectively reduced to meet the requirements of consumer robots for low-cost 3D radars. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic flow chart of a cloud acquisition control method of the present invention; Figure 2 is a schematic structural diagram of a mechanical code disk of the present invention; Figure 3 is a schematic diagram for determining the zero position of the present invention; Figure 4 is a schematic diagram for zero point detection of the present invention; Figure 5 is a schematic timing diagram after n-turn scanning of the present invention; Figure 6 is a schematic diagram of a point cloud effect collected by applying the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0026] On the contrary, the present invention covers any alternatives, modifications, equivalent methods and solutions made within the spirit and scope of the present invention as defined by the claims. Further, in order to enable the public to better understand the present invention, some specific details are described in detail in the following detailed description of the present invention. Those skilled in the art can fully understand the present invention without the description of these details.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0028] As Figure 1 shown, the first specific embodiment of the three-dimensional lidar point cloud acquisition control method of the present invention: A three-dimensional lidar point cloud acquisition control method includes the following steps: Based on the demand information of the three-dimensional lidar zero point marker, a notch is designed on the encoder disk of the three-dimensional lidar, and the width of the notch on the encoder disk is optimized to obtain notch design data; Based on the notch design data, capture the sudden change in the pulse signal period caused by the notch to obtain zero point position information; According to the zero point position information and combined with a filtering algorithm, obtain jitter interference cancellation data for eliminating the jitter interference of the lidar; For the lidar whose jitter interference has been eliminated, perform angular resolution calibration to obtain sampling time adjustment information; Based on the sampling time adjustment information, and through a multi-turn scanning accumulation mechanism, perform spatio-temporal interpolation fusion on the multi-turn acquisition data to generate high-density point cloud data.

[0029] The second specific embodiment of the three-dimensional lidar point cloud acquisition control method of the present invention: A three-dimensional lidar point cloud acquisition control method includes the following steps: Step 1, through a pre-created encoder disk design optimization model, based on the demand information of the mechanical encoder disk zero point marker, optimize the width of the notch on the mechanical encoder disk to obtain a notch width value; Step 2, use a pre-created sudden change signal capture model, based on the notch width value, capture the sudden change in the pulse signal period caused by the notch to obtain zero point position information; Step 3: Based on the pre-created jitter interference cancellation model, and according to the zero point position information and combined with the time window filtering algorithm, obtain the jitter interference cancellation data for eliminating the jitter interference of the lidar; Step 4: Use the pre-created sampling dynamic adjustment model to calibrate the angular resolution of the lidar that has already eliminated the jitter interference, and obtain the sampling time adjustment information; Step 5: Adopt the pre-created spatio-temporal interpolation fusion model. Based on the sampling time adjustment information, and through the multi-loop scanning accumulation mechanism, perform spatio-temporal interpolation fusion on the multi-loop data to generate high-density point cloud data.

[0030] In this embodiment, in Step 1, through the pre-created code disk design optimization model, based on the requirement information of the mechanical code disk zero point marking, optimize the notch width of the mechanical code disk. The method for obtaining the notch width value is as follows: Based on the design requirements of the mechanical code disk zero point marking, combined with the signal detection robustness and high-speed dynamic response characteristics, construct a code disk design optimization model with multi-parameter constraints. By quantifying mechanical, optical, and dynamic variables, output the solution of the minimum safe notch width to ensure stable recognition at high rotational speeds. The input parameters of this embodiment can be seen in Table 1.

[0031] Table 1: Input Parameters

[0032] The multi-parameter constraints include the minimum detectable angle constraint, the mechanical strength safety constraint, and the signal width constraint.

[0033] Minimum Detectable Angle The constraint is used to ensure that the notch can cover the projection of the light spot in the tangential direction of the code disk. Its expression is as follows:

[0034] Among them, the first term is the light spot angular spread, and the second term is the angle drift compensation caused by the rotational speed.

[0035] The mechanical strength safety constraint is used to ensure that the notch width can meet the torsional strength requirements of the code disk. Its expression is as follows:

[0036] Among them is the maximum torque, is the material yield strength.

[0037] The signal width constraint is used to ensure that the notch pulse can be significantly distinguished from the normal tooth pulse. Its expression is as follows:

[0038] Among them, k is the safety factor, which is used to reserve design redundancy in the signal pulse width constraint, v is the linear velocity of the code disk, is the jitter tolerance of the optical sensor.

[0039] Based on multi-parameter constraints, an optimization objective function is constructed, and its expression is as follows:

[0040] Among them, the first term of the expression is used to maintain the design benchmark, generally 2 times the tooth width; the second term of the expression is used to minimize the risk of missed detection, is the optical sensitivity weight; the third term of the expression is used to suppress stress concentration, is the safety factor, is the working stress.

[0041] In this embodiment, a mechanical code disk with 11 teeth in a single turn is adopted, and a notch is set at the 12th equal division position of the mechanical code disk. For the specific notch position, refer to Figure 2 ; the diameter of the code disk is D = 50 mm, and the rotation speed is ; the yield strength of the material is The spot diameter is ; the signal-to-noise ratio is SNR = 20 dB. Furthermore, the parameter calculation results can be seen in Table 2.

[0042] Table 2: Parameter calculation results

[0043] Therefore, in this embodiment, high-robustness zero-point detection is realized through the difference in notch width, and stable identification can still be achieved even at high rotation speeds (greater than 10,000 RPM). Furthermore, the present invention can realize zero-point identification by using a low-cost code disk, thereby reducing the hardware cost by 60%.

[0044] In this embodiment, the method for zero-point detection is as follows: Utilize the sudden change in the pulse signal period caused by the notch, capture the edge signal through the interruption of the microcontroller MCU, and combine the time window filtering algorithm to eliminate the jitter interference. The method is as follows: Output a pulse signal through the optical encoder, and the interval threshold of the pulse signal is 1.5 times the average period; then trigger the interruption of the microcontroller MCU through the edge detection circuit, and determine the zero-point position through the time interval threshold. Refer to Figure 3 . The signal processing flow mainly includes the following processes: Adjust the pulse interval threshold in real time according to the rotation speed to adapt to the signal attenuation under different working conditions (such as the working condition where electromagnetic interference increases during high-speed rotation).

[0045] ​Configure a multi-level interrupt nesting mechanism in the microcontroller MCU to ensure the real-time performance of the angle sampling task, and the interrupt response time is less than 1 microsecond 。

[0046] In this embodiment, the method for signal debounce filtering is as follows: For the signal processing of the optical encoder, a code disk signal debounce filtering method and system based on rotational speed prediction and confidence value judgment are adopted to solve the defects of strong hardware dependence and insufficient dynamic adaptability of the traditional code disk debounce scheme. The specific contents are as follows: Using the rotational speed time series prediction model, based on the real-time rotational speed n, whose unit is revolutions per minute (RPM), calculate the expected arrival time of the pulse The calculation formula is as follows:

[0047] where is the number of lines of the code disk, is the process tolerance limit.

[0048] Based on the actual pulse time measure the timing deviation to obtain the prediction value deviation The calculation formula is as follows:

[0049] Construct a multi-factor fusion decision formula to dynamically generate the confidence value The calculation formula is as follows:

[0050] where represents the timing deviation sensitivity coefficient; represents the current signal-to-noise ratio; represents the reference signal-to-noise ratio, which is the quality reference value measured under ideal working conditions of the system; represents the rotational speed change rate; represents the rotational speed change rate suppression factor, which depends on the rotational speed increase increment of the code disk motor.

[0051] According to the confidence value and the confidence interval, construct a filtering decision mechanism, which includes the following contents: When the confidence value is greater than or equal to 0.8, the signal does not need to be filtered and can be directly output.

[0052] When the confidence value is less than 0.8, the signal is filtered by the counter debounce method.

[0053] Furthermore, based on a 12-line code disk and a rotational speed of 12000 RPM for verification, the specific debounce verification results can be seen in Table 3.

[0054] Table 3: Debounce Verification Results

[0055] In this embodiment, the method for adjusting the laser pulse interval threshold is as follows: Capture the rising edge of the sensor output signal in real time, identify the effective edge time through a digital comparator, and record the timestamp . And obtain the current linear velocity through the sensor output signal or angular velocity , and calculate the instantaneous velocity vector.

[0056] Calculate the pulse interval The dynamic formula of is as follows:

[0057] where is the minimum resolvable distance (constrained by the spot diameter); is the acceleration compensation factor (to suppress the influence of velocity mutation); is the sensor response time; represents the rate of change of velocity; is the maximum allowable acceleration; is the response time correction factor.

[0058] In this embodiment, the method for dynamic overloading of the timer is as follows: After detecting the edge event, immediately write the pulse interval into the timer comparison register and reload the initial count value for the next cycle; if the pulse interval is less than the dead time, enable the pulse queue buffer and schedule the trigger timing according to the priority. The dynamic overloading effect can be seen in Table 4.

[0059] Table 4: Dynamic Overloading Effect

[0060] In this example, the method for dynamic angle calibration is as follows: During the motor startup process, when the rotational speed reaches 70% of the rated speed, and under normal operating conditions, when the motor speed fluctuates ( 5%), by adjusting the sampling interval time in real time, ensure that the angular resolution is strictly constant at , and eliminate the problem of angular resolution fluctuation caused by adjusting the number of sampling times in the traditional scheme. The specific method is as follows: Obtain the real-time rotational speed as (unit: RPM), and convert the real-time rotational speed to angular velocity (unit: radians per second rad / s), and its calculation formula is as follows:

[0061] Set the target angular resolution and obtain the radian resolution based on the target angular resolution The calculation formula is as follows:

[0062] Derive the sampling time according to the relationship between angular velocity and resolution The calculation formula is as follows:

[0063]

[0064] Since the sampling time is inversely proportional to the rotational speed When the rotational speed increases, the sampling interval is shortened; when the rotational speed decreases, the interval is extended to ensure that the angular increment per point is constantly .

[0065] In this embodiment, the method for real-time rotational speed detection and dynamic calculation is as follows: After the output pulses of the optoelectronic encoder are shaped by a Schmidt trigger, they trigger an external interrupt (rising edge detection) of the microcontroller MCU. Record the time interval between adjacent code disk signals (excluding the zero notch teeth).

[0066] Calculate the real-time rotational speed according to the time interval. The calculation formula is as follows:

[0067] Convert the calculated sampling time into the count value of the timer through the following formula to update the sampling time. The calculation formula is as follows:

[0068] where is the timer clock period and PSC is the prescaler coefficient. Update the overflow time of the timer interrupt in real time at the encoder trigger point, execute the laser ranging task, ensure the stability of the angular resolution, and perform calibration at each rising edge to recalculate the sampling interval. See Figure 4 .

[0069] Furthermore, the test results of the angular resolution stability can be obtained. Under the rotational speed fluctuation of %, the angular resolution error % is significantly better than the traditional scheme ( 5%), see Table 5.

[0070] Table 5: Test Results of Angular Resolution Stability

[0071] In this embodiment, the method for performing spatio-temporal interpolation fusion on multi-loop data to generate high-density point clouds is as follows: Using the frequency-division interpolation mode, without increasing the single-loop sampling frequency, through multi-loop scanning accumulation and time interpolation, the equivalent angular resolution is increased to one nth of the angular resolution . The problem of sparse point clouds caused by traditional fixed-angle sampling can be effectively solved. Divide the basic sampling period by n (n = 2, 4, 8) to generate n sub-trigger moments. Through continuous n-loop scanning, ranging is triggered at different moments after frequency division in each loop, and finally all sub-positions are covered. Spatio-temporal interpolation fusion is performed on the multi-loop data to generate high-density point clouds. It includes the following contents: Set the frequency-division trigger timing, which includes the basic sampling period , the frequency-division coefficient , .

[0072] When the angular resolution , ; Based on the frequency-division coefficient and the basic sampling period, calculate the sub-period , and its calculation formula is as follows:

[0073] According to the sub-period, generate multi-loop trigger moments. In the k-th loop scan , the laser ranging trigger moment is:

[0074] In each loop scan, on the basis of the basic period , superimpose the frequency-division offset , covering n sub-positions. After continuous n-loop scanning, all sub-positions are completely covered, and m represents the serial number of the sub-sampling period. The scanning effects of the first loop, the second loop, the third loop, the fourth loop, and the cumulative 4 loops can be seen in Figure 5 .

[0075] When the angular resolution , calculate the number of single-loop sampling points , and its calculation formula is as follows:

[0076] Based on the number of single-loop sampling points, calculate the equivalent number of points after n-frequency division , and its calculation formula is as follows:

[0077] When n = 4, the equivalent angular resolution is equal to 0.15 degrees. The point cloud spacing at 10 meters is reduced from 10 m to 2.5 cm, effectively improving the equivalent angular resolution.

[0078] In summary, the present invention can achieve high-precision and low-cost angle measurement, realizing a dynamic resolution of 0.15 degrees on a single-turn 12-equal-part code disk with a rotational speed exceeding 10,000 RPM. By using fewer code disk teeth, the overhead of interrupts on the microcontroller MCU is effectively reduced, ensuring the real-time performance of the radar measurement system.

[0079] Meanwhile, the present invention can effectively improve the point cloud density. Through n-frequency division interpolation, the angular resolution is improved to 0.15 degrees within 4-circle scanning. The point cloud spacing at 10 meters is reduced from 10 cm to 2.5 cm. For the effect, see Figure 6 , Figure 6 In it, the red line represents the X-axis, the green line represents the Y-axis, and the blue line represents the Z-axis. Its color scheme follows the correspondence between the three primary colors RGB and the three-dimensional rectangular coordinate axes.

[0080] Furthermore, the present invention has strong robustness. Through a real-time rotational speed compensation algorithm, the angular error is stabilized within ±0.33%, suitable for scenarios with a rotational speed fluctuation of ±5%.

[0081] An equipment embodiment applying the method of the present invention: An electronic device, which includes: One or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the above-mentioned three-dimensional lidar point cloud acquisition control method.

[0082] A computer medium embodiment applying the method of the present invention: A computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above-mentioned three-dimensional lidar point cloud acquisition control method is implemented.

[0083] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0084] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce means for implementing the functions specified in one Figure 1 flow or multiple flows or / and blocks Figure 1 or multiple blocks.

[0085] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one Figure 1 flow or multiple flows or / and blocks Figure 1 or multiple blocks.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 flow or multiple flows or / and blocks Figure 1 or multiple blocks.

[0087] The model in the present application is an object that constitutes an objective description of the morphological structure by means of an entity or a virtual representation. The object is not equal to an object and is not limited to entities and virtuals. It can be a data processing function, a software program, a processing mode, a usage method, an operation mode, a work process, an application process, electronic hardware, a circuit module, a processing system, a system imitation, or a simulation object.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art can still modify or equivalently replace the specific implementation manners of the present invention. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for controlling the acquisition of 3D lidar point cloud, characterized in that: It includes the following steps: Step 1, through a pre-created encoder design optimization model, based on the requirement information of the 3D lidar zero point marker, design a notch on the encoder of the 3D lidar, and optimize the width of the notch on the encoder to obtain notch design data; Step 2, use a pre-created mutation signal capture model, based on the notch design data, capture the mutation of the pulse signal period caused by the notch to obtain zero point position information; Step 3, based on a pre-created jitter interference elimination model, according to the zero point position information and combined with a filtering algorithm, obtain jitter interference elimination data for eliminating the jitter interference of the lidar; Step 4, use a pre-created sampling dynamic adjustment model to calibrate the angular resolution of the lidar whose jitter interference has been eliminated to obtain sampling time adjustment information; Step 5, adopt a pre-created spatio-temporal interpolation fusion model, based on the sampling time adjustment information, and through a multi-loop scanning accumulation mechanism, perform spatio-temporal interpolation fusion on the multi-loop acquisition data to generate high-density point cloud data.

2. A method for controlling the acquisition of 3D lidar point cloud as described in claim 1, characterized in that: The method of Step 1, through a pre-created encoder design optimization model, based on the requirement information of the 3D lidar zero point marker, design a notch on the encoder of the 3D lidar, and optimize the width of the notch on the encoder to obtain notch design data: Obtain the number of teeth and tooth position information of the encoder; Calculate several equally divided positions of the encoder according to the number of teeth and tooth position information; Take one or more of the equally divided positions as the position of the notch; Collect the requirement information of the encoder zero point marker, which includes the design requirement information and signal detection requirement information of the encoder; Quantify the design requirement information and signal detection requirement information to obtain several variable parameters, and construct multi-parameter constraint information; Based on several variable parameters and multi-parameter constraint information, construct an objective function for optimizing the width of the notch on the encoder; Solve the objective function to obtain the optimized notch width value; Summarize the position of the notch and the notch width value to obtain notch design data.

3. A method for controlling the acquisition of 3D lidar point cloud as described in claim 2, characterized in that: The method of quantifying the design requirement information and signal detection requirement information to obtain several variable parameters is as follows: Collect the design requirement information, which includes mechanical parameters and dynamic parameters; Obtain the signal detection requirement information, which includes optical parameters and environmental parameters; Quantify the mechanical parameters to obtain the encoder diameter, total number of teeth and standard number of teeth; Quantify the dynamic parameters to obtain the maximum working speed and sensor response time; Quantify the optical parameters to obtain the light source wavelength and spot diameter; Quantify the environmental parameters to obtain the signal-to-noise ratio threshold.

4. A method for controlling the acquisition of 3D lidar point cloud as described in claim 2, characterized in that: The method of constructing multi-parameter constraint information is as follows: According to the notch design requirements, set the spot opening angle and the angle drift compensation value caused by the rotation speed; Based on the spot angular spread and the angle drift compensation value, establish the minimum detection angle constraint so that the notch can cover the projection of the spot in the tangential direction of the code disk; Obtain the maximum torque of the code disk and the material yield strength; According to the maximum torque and the material yield strength, establish the mechanical strength safety constraint so that the notch width can meet the torsional strength requirements of the code disk; Obtain the linear velocity of the code disk and the jitter tolerance of the optical sensor; Based on the linear velocity of the code disk and the jitter tolerance of the optical sensor, establish the signal width constraint so that the notch pulse can be distinguished from the normal tooth pulse; Summarize the minimum detection angle constraint, the mechanical strength safety constraint, and the signal width constraint to obtain the multi-parameter constraint information.

5. The method for controlling the point cloud acquisition of a three-dimensional lidar according to claim 2, wherein: The method for constructing the objective function is as follows: Obtain the multi-parameter constraint information, which includes the minimum detection angle constraint, the mechanical strength safety constraint, and the signal width constraint; According to the mechanical strength safety constraint, the signal width constraint, and the standard number of teeth, construct the design reference expression; Based on the minimum detection angle constraint, the code disk diameter, and the optical sensitivity weight, construct the expression for minimizing the missed detection risk; According to the material yield strength and the material safety factor, construct the expression for suppressing stress concentration; Couple the design reference expression, the expression for minimizing the missed detection risk, and the expression for suppressing stress concentration to obtain the optimized objective function.

6. The method for controlling the point cloud acquisition of a three-dimensional lidar according to claim 1, wherein: Step two, use the previously created mutation signal capture model, and based on the notch design data, capture the mutation of the pulse signal period caused by the notch to obtain the zero point position information. The method is as follows: According to the signal attenuation situation under different working conditions and based on the rotational speed, adjust the pulse interval threshold in real time; Based on the pulse interval threshold, obtain the pulse signal output by the photoelectric encoder; Based on the notch design data, capture the mutation of the pulse signal period caused by the notch through the edge detection circuit to obtain the edge signal; According to the edge signal and combined with the pulse interval threshold, determine the zero point position to obtain the zero point position information; The method for adjusting the pulse interval threshold in real time is as follows: Capture the rising edge signal output by the photoelectric encoder; Identify the effective edge time of the rising edge signal through a digital comparator to obtain the time stamp; Based on the signal output by the photoelectric encoder, obtain the current linear velocity; According to the current linear velocity, the minimum resolution distance, the acceleration compensation factor, and the time stamp, construct the dynamic calculation formula for the pulse interval; Solve the dynamic calculation formula for the pulse interval to obtain the pulse interval threshold.

7. The method for controlling the point cloud acquisition of a three-dimensional lidar according to claim 1, wherein: Step three, based on the previously created jitter interference elimination model, according to the zero point position information and combined with the filtering algorithm, obtain the jitter interference elimination data. The method is as follows: Based on the real-time rotational speed, the number of lines of the code disk, and the process tolerance, predict the rotational speed time series and calculate the expected arrival time of the pulse: According to the expected arrival time of the pulse and the actual pulse time, measure the time series deviation and calculate the predicted value deviation: According to the filtering algorithm, a multi-factor fusion decision formula is established by using the prediction value deviation, the timing deviation sensitivity coefficient, the current signal-to-noise ratio, and the rotation speed change rate suppression factor; Solve the multi-factor fusion decision formula to obtain the dynamic confidence value; Set the filtering decision mechanism according to the confidence interval and the dynamic confidence value; When the dynamic confidence value is greater than or equal to the confidence threshold, directly output the pulse signal; When the dynamic confidence value is less than the confidence threshold, perform counter debouncing on the pulse signal to obtain the jitter interference cancellation data.

8. A three-dimensional lidar point cloud acquisition control method according to claim 1, characterized in that: Step four, using the pre-created sampling dynamic adjustment model, calibrate the angular resolution of the lidar that has eliminated jitter interference to obtain the sampling time adjustment information, and the method is as follows: Obtain the encoder pulse signal output by the optoelectronic encoder; After shaping the encoder pulse signal by using a Schmitt trigger, trigger the external interrupt of the microcontroller to obtain the time interval between adjacent encoder signals; Based on the time interval, calculate the real-time rotation speed of the three-dimensional lidar; Calculate the corresponding angular velocity according to the real-time rotation speed; Based on the target angular resolution, calculate the radian resolution; Calculate the sampling time according to the relationship between the angular velocity and the radian resolution; Process the sampling time, and based on the timer clock cycle and the prescaler coefficient, obtain the counter value of the timer; According to the counter value of the timer, update the overflow time of the timer interrupt in real time to determine the sampling time adjustment information.

9. A three-dimensional lidar point cloud acquisition control method according to claim 1, characterized in that: Step five, adopt the pre-created spatio-temporal interpolation fusion model, based on the sampling time adjustment information, and through the multi-loop scanning accumulation mechanism, perform spatio-temporal interpolation fusion on the multi-loop acquisition data to generate high-density point cloud data, and the method is as follows: Obtain the sampling time adjustment information, which includes the basic sampling period; Calculate the required angular resolution according to the point cloud spacing requirement; Based on the angular resolution, calculate the number of sampling points per circle; Determine the required division factor according to the number of sampling points per circle; Based on the division factor, divide the basic sampling period to generate multiple sub-trigger times; Through the multi-loop scanning accumulation mechanism, and according to multiple sub-trigger times, simultaneously superimpose the division offset to set the ranging trigger time for each loop scan to cover multiple sub-positions; Based on the ranging trigger time, perform several loop scans so that all sub-positions are completely covered to obtain multi-loop acquisition data; perform spatio-temporal interpolation fusion on the multi-loop data to generate high-density point cloud data.

10. A three-dimensional lidar point cloud acquisition control system, characterized in that: It includes: One or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement a three-dimensional lidar point cloud acquisition control method according to any one of claims 1-9.

Images