Solid-state laser radar line width and temperature combined automatic calibration method

Abstract

In view of the deficiencies of the prior art, the present application provides a solid-state laser radar pulse width and temperature combined automatic calibration method to overcome the problems of long calibration time, repetitive operation and poor compensation consistency of different batches. In the present application, the principle of "priority pulse width compensation, temperature compensation following" is adopted, and in the pulse width compensation process, the dynamic compensation method is adopted, and after the single pulse width compensation is completed, the temperature compensation data is collected, and after a single temperature cycle, the temperature compensation calculation is started under the condition that the pulse width compensation has been completed, the temperature compensation adopts the linear fitting method, the centroid of the imaging point is calculated, the centroid offset relative to the normal temperature under different temperatures is compensated, the temperature compensation parameter file is generated and applied to the subsequent algorithm, and the pulse width and temperature combined automatic calibration of the solid-state laser radar is realized.

Description

Technical Field

[0001] This invention relates to lidar calibration technology, and more particularly to an automatic calibration method for pulse width and temperature in a solid-state lidar production line. Background Technology

[0002] Solid-state lidar has entered the mass production stage for automotive applications. Pure solid-state lidar based on dToF (direct time of flight) typically uses a VCSEL array as the transmitter and a SPAD (single-photon avalanche diode) array as the receiver.

[0003] Furthermore, the pulse width of VCSELs drifts with temperature, current, and batch variations. This drift is converted into a trigger timing offset after passing through the leading edge discrimination circuit, causing the ranging value of the same target to shift back and forth depending on the temperature or channel, forming a non-negligible fixed deviation. Targets with different reflectivities will amplify this deviation, causing white walls and black pillars to be measured as having different depths at the same distance. Due to inconsistent pulse widths, multi-channel chips exhibit layered misalignment in the point cloud, affecting subsequent fusion and localization.

[0004] Temperature is the primary external factor contributing to the aforementioned drift. The vehicle environment requires the radar to operate at full performance between -40°C and 85°C. Experiments show that temperature alone can cause a change of approximately 0.3 ns in the equivalent pulse width of the VCSEL, corresponding to a near 5 cm "moving wall" illusion. Simultaneously, the SPAD breakdown voltage decreases with increasing temperature, relatively raising the discrimination threshold and further amplifying inter-channel dispersion. The two temperature coefficients, with opposite directions and different amplitudes, result in a non-zero slope in the single-module full-temperature ranging curve after superposition, with inconsistent slopes across channels, causing overall point cloud shift and internal misalignment.

[0005] Existing methods for pulse width and temperature compensation of solid-state lidar mostly rely on manual operation and separate compensation stations, which suffer from drawbacks such as inconsistent compensation, repetitive work, and low degree of automation.

[0006] Therefore, this invention proposes an automatic calibration method for pulse width and temperature in solid-state lidar production lines, which compresses the ranging error of solid-state lidar to a range acceptable for autonomous driving. Summary of the Invention

[0007] In view of the deficiencies of the prior art described above, the present invention provides a joint automatic calibration method for pulse width and temperature of solid-state LiDAR on the production line, so as to overcome the problems of long calibration time for different batches, repetitive operations, and poor compensation consistency. In the present invention, the principle of "pulse width compensation first, temperature compensation follows" is adopted. In the pulse width compensation process, a dynamic compensation method is used, and data for temperature compensation acquisition starts after a single pulse width compensation is completed. After one full temperature cycle, in the case that the pulse width compensation has been finished, the temperature compensation calculation is started. The temperature compensation adopts a linear fitting method, which calculates the centroid of the imaging point, compensates for the centroid offset relative to the normal temperature at different temperatures, and generates a temperature compensation parameter file applied to subsequent algorithms, so as to realize the joint automatic calibration of pulse width and temperature of the solid-state LiDAR.

[0008] To achieve the above purpose, the specific implementation steps of the present invention are as follows:

[0009] S1: Set the calibration temperature range and variation step size;

[0010] S2: Collect and calculate the full width at half maximum F of the laser output at normal temperature;

[0011] S3: Set the temperature of the temperature chamber, collect the full width at half maximum F_i at different temperatures, and calculate |F-F_i|;

[0012] S4: Perform pulse width compensation according to the result of |F-F_i|;

[0013] S5: Collect temperature compensation data;

[0014] S6: Perform temperature compensation calculation.

[0015] In S1, set the minimum temperature Tmin and maximum temperature Tmax of the calibration temperature, set the temperature variation step of the temperature chamber as ΔT, then the number of temperature points collected in one temperature cycle is K=(Tmax-Tmin) / ΔT.

[0016] In S2, at normal temperature, deploy the data collection environment in the temperature chamber, collect one frame of raw data, and calculate the full width at half maximum F.

[0017] In S3, set the temperature of the temperature chamber to Tmin, wait until the module is stabilized, collect one frame of raw data, calculate the full width at half maximum F_i, and calculate |F-F_i|.

[0018] In S4, when |F-F_i|>200ps, perform pulse width compensation.

[0019] S401: When F<F_i, perform negative compensation on the LiDAR, re-collect raw data after compensation, and repeatedly execute S401 until |F-F_i|<=200ps, then the pulse width compensation at this temperature is finished, and the process proceeds to S5.

[0020] S402: When F>F_i, perform positive compensation on the lidar, and then re-acquire raw data. Execute S401 repeatedly until |F-F_i|<=200ps, at which point pulse width compensation ends and proceeds to S5.

[0021] S403: When |F-F_i|<=200ps, pulse width compensation ends at this temperature, and proceeds to S5.

[0022] In step S5, a frame of raw data is acquired as temperature compensation data at that temperature. After acquisition, the process returns to step S3, increments the temperature step once, and continues with steps S3 and S4. When Tmin reaches Tmax, the process proceeds to step S6.

[0023] In step S6, after the temperature polling process from Tmin to Tmax ends, temperature compensation calculation is performed.

[0024] S601: Calculate the centroid of the histogram of the entire pixel based on the raw data collected at different temperatures in S5 above.

[0025] S602: Using 10x10 pixels as a pixel block, take the average value and list the centroid scatter plots at different temperatures.

[0026] S603: Perform linear fitting on the scatter plot in S602, and output the temperature compensation file based on the fitted curve to complete the temperature compensation of the lidar. Attached Figure Description

[0027] Figure 1 This is the overall flowchart of the present invention;

[0028] Figure 2 This is a comparison chart of pulse width compensation results;

[0029] Figure 3 The graph shows the results of the linear fitting for temperature compensation. Detailed Implementation

[0030] The invention will now be further described with reference to the accompanying drawings.

[0031] This invention provides a method for the joint automatic calibration of pulse width and temperature in solid-state lidar production lines, overcoming the problems of long calibration times, repetitive operations, and poor compensation consistency across different batches. In this invention, the principle of "prioritizing pulse width compensation, followed by temperature compensation" is adopted. During pulse width compensation, a dynamic compensation method is used. Temperature compensation data is collected after a single pulse width compensation cycle. After a single temperature cycle, when pulse width compensation is complete, temperature compensation calculation begins. Temperature compensation employs a linear fitting method, calculating the centroid of the imaging point to compensate for the centroid shift relative to room temperature at different temperatures. This generates a temperature compensation parameter file for subsequent algorithms, achieving joint automatic calibration of pulse width and temperature for solid-state lidar.

[0032] To achieve the above objectives, please refer to the appendix. Figure 1 The specific execution steps of this invention are as follows:

[0033] S1: Set the calibration temperature range and variation step size;

[0034] S2: Collect and calculate the full width at half maximum (FWHM) of the laser output at room temperature;

[0035] S3: Set the temperature of the incubator, collect the half-peak width F_i at different temperatures, and calculate |F-F_i|;

[0036] S4: Perform pulse width compensation based on the result of |F-F_i|;

[0037] S5: Temperature compensation data acquisition;

[0038] S6: Temperature compensation calculation.

[0039] In S1, the minimum temperature Tmin and the maximum temperature Tmax of the calibration temperature are set, and the temperature change step size of the temperature chamber is set to ΔT. Then, the number of temperature points collected in a single cycle of temperature collection is K = (Tmax - Tmin) / ΔT.

[0040] In step S2, a data acquisition environment is deployed in a temperature chamber at room temperature to acquire one frame of raw data and calculate the half-peak width F.

[0041] In step S3, the temperature of the incubator is set to Tmin. After the module stabilizes, a frame of raw data is collected, the half-peak width F_i is calculated, and |F-F_i| is calculated.

[0042] See attached document Figure 2 In S4, when |F-F_i|>200ps, the pulse width is compensated.

[0043] S401: When F<F_i, perform negative compensation on the LiDAR, re-acquire raw data after compensation, repeatedly execute S401 until |F-F_i|<=200ps, then the pulse width compensation at this temperature ends, and the process proceeds to S5.

[0044] S402: When F>F_i, perform positive compensation on the LiDAR, re-acquire raw data after compensation, repeatedly execute S401 until |F-F_i|<=200ps, then the pulse width compensation at this temperature ends, and the process proceeds to S5.

[0045] S403: When |F-F_i|<=200ps, the pulse width compensation at this temperature ends, and the process proceeds to S5.

[0046] In S5 described above, one frame of raw data is collected as the temperature compensation data at this temperature. After the collection is completed, the process returns to S3, the temperature is stepped once, and S3 and S4 are continuously executed. After Tmin is stepped to Tmax, the process proceeds to S6.

[0047] Referring to the accompanying Figure 3 , in S6 described above, after the temperature polling from Tmin to Tmax is completed, temperature compensation calculation is performed.

[0048] S601: Calculate the histogram centroid of all pixels based on the raw data at different temperatures collected in S5 described above.

[0049] S602: Take every 10x10 pixels as one pixel block, calculate the average value, and list the centroid scatter plots at different temperatures.

[0050] S603: Perform linear fitting on the scatter plot in S602, output a temperature compensation file according to the fitting curve, and complete the temperature compensation of the LiDAR.

[0051] The above description is merely the preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any variations or replacements that can be easily conceived by a person skilled in the art within the technical scope disclosed by the present invention shall be covered within the protection scope of the present invention.

Claims

1. A method for automatic calibration of pulse width and temperature in a solid-state lidar production line, characterized in that, Following the principle of "pulse width compensation first, with temperature compensation following", a dynamic compensation method is adopted in the pulse width compensation process. Temperature compensation data collection starts after a single pulse width compensation is completed. After a single round of temperature cycling, when the pulse width compensation has been finished, temperature compensation calculation is initiated. The temperature compensation adopts a linear fitting method, which compensates for the centroid deviation relative to the normal temperature at different temperatures by calculating the centroid of the imaging point, generates a temperature compensation parameter file applied to subsequent algorithms, and realizes the joint automatic calibration of pulse width and temperature for solid-state LiDAR; the specific implementation steps of the present invention are as follows: S1: Set the calibration temperature range and variation step size; S2: Collect and calculate the full width at half maximum F of the laser output at normal temperature; S3: Set the temperature of the temperature chamber, collect the full width at half maximum F_i at different temperatures, and calculate |F-F_i|; S4: Perform pulse width compensation according to the result of |F-F_i|; S5: Collect temperature compensation data; S6: Perform temperature compensation calculation.

2. The method for automatic calibration of pulse width and temperature in a solid-state lidar production line as described in claim 1, characterized in that, In S2, at normal temperature, a data collection environment is deployed in the temperature chamber, a frame of raw data is collected, and the full width at half maximum F is calculated; in S3, the temperature of the temperature chamber is set to Tmin, after the module is stabilized, a frame of raw data is collected, the full width at half maximum F_i is calculated, and |F-F_i| is calculated; in S4, when |F-F_i|>200ps, if F<F_i, negative compensation is performed on the LiDAR, raw data collection is re-performed after the compensation, S401 is executed cyclically until |F-F_i|<=200ps, then the pulse width compensation at this temperature ends, and the process proceeds to S5; if F>F_i, positive compensation is performed on the LiDAR, raw data collection is re-performed after the compensation, S401 is executed cyclically until |F-F_i|<=200ps, then the pulse width compensation at this temperature ends, and the process proceeds to S5; when |F-F_i|<=200ps, the pulse width compensation at this temperature ends, and the process proceeds to S5; in S5, a frame of raw data is collected as the temperature compensation data at this temperature, after the collection is completed, the process returns to S3, the temperature is stepped once, and S3 and S4 are continued to be executed; when Tmin is stepped to Tmax, the process proceeds to S6; in S6, after the temperature polling from Tmin to Tmax is completed, the temperature compensation calculation is performed; according to the raw data at different temperatures collected in S5 above, the histogram centroid of all pixels is calculated; Taking every 10x10 pixels as a pixel block, taking the average value, and listing the centroid scatter plots at different temperatures; performing linear fitting according to the above scatter plots, outputting a temperature compensation file according to the fitting curve, and completing the temperature compensation of the LiDAR.

Images