A lidar module and its accuracy detection method

Abstract

The present invention relates to the technical field of lidar, and discloses a lidar module and an accuracy detection method therefor. The method includes: setting detection plates at a number of known positions, placing the lidar module to be tested at a detection point at a known position, comparing the detected distance value with the actual distance value, and adjusting the data processing period; when the data processing period meets the conditions, obtaining detection data and establishing a time series data set, and determining whether the lidar module is accurate; if there are out-of-threshold data points, determining that the module is inaccurate, otherwise determining that it is qualified; when the module is inaccurate, extracting inaccurate features and comparing them with a historical data set to determine the cause of inaccuracy; calculating a tolerance value according to the cause of inaccuracy, and sorting the modules according to this value; the inaccurate features include inaccurate data and inaccurate types. This method effectively improves the detection accuracy and production reliability of the lidar module by monitoring in real time and accurately identifying the source of errors.

Description

Technical Field

[0001] The present invention relates to the technical field of lidar, and in particular, to a lidar module and an accuracy detection method thereof. Background Art

[0002] As a high-precision environmental perception technology, lidar is widely used in fields such as autonomous driving, UAV mapping, and security monitoring. The core optical structure of the lidar module usually includes a barrel, a front shell, a lens group, and related electronic components, and its assembly process has a direct impact on the performance, stability, and yield of the product. In addition, the existing lidar accuracy detection methods mainly rely on static calibration and laboratory measurements, such as using a standard reflecting surface or a target object with a known size for distance measurement comparison. Although these methods can provide certain detection capabilities in an ideal experimental environment, in a complex actual application environment, the measurement accuracy of the lidar may be affected by various dynamic factors, making it difficult for traditional methods to accurately evaluate its true working state. In addition, some detection methods only focus on a single factor (such as distance error), while ignoring the comprehensive impact of various factors such as signal intensity and environmental temperature on the accuracy of the lidar, resulting in insufficient reliability of the detection results.

[0003] Therefore, there is an urgent need for a more comprehensive and intelligent lidar accuracy detection method to improve the accuracy, robustness, and adaptability of the detection, so as to provide a more scientific basis for the optimization and screening of lidar modules. Summary of the Invention

[0004] In view of this, the present invention proposes a lidar module and an accuracy detection method thereof, aiming to solve the problem that the accuracy of the lidar in the current technology is difficult to guarantee.

[0005] On the one hand, an accuracy detection method for a lidar module proposed by the present invention includes: setting detection plates at a plurality of known positions, placing the lidar module to be tested at a detection point with a known position, comparing the detected distance value of the lidar module to be tested with the actual distance value calculated according to the detection plate and the detection point, and adjusting the data processing period of the detected distance value according to the distance difference between the detected distance value and the actual distance value.

[0006] When the data processing period is satisfied, the detection data within the data processing period is acquired, a detection data set is established according to the time series, the detection data set is processed, and it is determined whether the lidar module to be tested is accurate; when there is at least one over-threshold data point, it is determined that the lidar module to be tested is inaccurate, and when there is no such over-threshold data point, it is determined that the lidar module to be tested is qualified.

[0007] When it is determined that the lidar module to be measured is inaccurate, extract the inaccurate features, compare the inaccurate features with the historical detection data set, and determine the reason for inaccuracy.

[0008] Calculate the tolerance value according to the reason for inaccuracy and the inaccurate features, and sort the lidar modules to be measured according to the size of the tolerance value.

[0009] Furthermore, the inaccurate features include: inaccurate data and inaccurate types. The inaccurate data is the detection data corresponding when it is determined that the lidar module to be measured is inaccurate, and the inaccurate type is the type of the corresponding detection data.

[0010] Furthermore, when determining whether the lidar module to be measured is accurate, it includes: obtaining the detection data set. The detection data in the detection data set is divided into four categories, including: distance difference, temperature, the received signal strength of the lidar module to be measured, and detection time. Establish data groups for the detection data obtained when the detection times in the detection data set are the same, and obtain N data groups. Each data group includes distance difference, temperature, the received signal strength of the lidar module to be measured, and detection time. Arrange according to the data groups to obtain an N*M matrix, where M = 4, and M is the number of types of the detection data. The N*M matrix is represented by A i，j representing any element in the matrix, and A i，j is the element of the jth category at the detection time of i.

[0011] Select any element in the N*M matrix as the center point, set the data judgment range, the radius of the data judgment range is R, select all elements of the same type within the data judgment range, and calculate the average value.

[0012] Set an error threshold. When the difference between A i，j and the average value is greater than or equal to the error threshold, determine that A i，j is the out-of-threshold data point. When the difference between A i，j and the average value is less than the error threshold, determine that A i，j is the normal data point.

[0013] When there is at least one out-of-threshold data point, determine that the lidar module to be measured is inaccurate.

[0014] Furthermore, when determining the reason for inaccuracy, it includes: extracting the inaccurate features, and judging the reason for inaccuracy according to the overlap degree between the inaccurate features and the historical detection data set.

[0015] The overlap degree satisfies the following relationship:

[0016] ;

[0017] Among them, S is the overlap degree, and x i is the i-th data point in the inaccurate data, is the mean value of all inaccurate data, and y i is the i-th data point in the historical dataset, is the mean value of all data points in the historical dataset, and n is the number of all inaccurate data.

[0018] Match the inaccurate data with the highest similarity with the historical dataset, and determine the inaccurate reason of the inaccurate data according to the inaccurate reason in the historical dataset.

[0019] Furthermore, when calculating the tolerance value according to the inaccurate reason and inaccurate feature, it includes:

[0020] ;

[0021] Among them, H is the tolerance value, d is the maximum distance difference in the inaccurate data, and d max is the maximum distance difference in the historical dataset, R is the minimum received signal strength in the inaccurate data, and R min is the minimum received signal strength in the historical dataset, K is the average temperature in the inaccurate data, is the average temperature in the historical dataset, α is the distance difference influence coefficient, β is the signal strength influence coefficient, and η is the temperature influence coefficient.

[0022] Furthermore, the distance difference influence coefficient α is obtained through the following relationship:

[0023] ;

[0024] The signal strength influence coefficient β is obtained through the following relationship:

[0025] ;

[0026] The temperature influence coefficient η is obtained through the following relationship:

[0027] ;

[0028] ;

[0029] Among them, f diff is the variance of all distance differences in the inaccurate data, f R is the variance of all signal strength values in the inaccurate data, f K is the variance of all temperature values in the inaccurate data, and f SUM is the total variance.

[0030] On the other hand, a lidar module proposed by the present invention is obtained by using the accuracy detection method of the above lidar module, and the lidar module includes.

[0031] A front shell, which is connected to a light-transmitting cover plate above it. The front shell is provided with a plurality of lens barrel holes, and the lower surface of the front shell is provided with lens barrel fixing posts and first threaded holes.

[0032] A lens barrel assembly, including a lens base and a lens barrel. The lens base and the lens barrel are of an integral structure. The lens base is provided with a second threaded hole and a lens barrel fixing hole, and a PIN fixing groove is provided on the lower surface of the lens base; the lens barrel fixing post is fixed to the lens barrel assembly by being inserted into the lens barrel fixing hole.

[0033] A fixing screw, which passes through the first threaded hole and the second threaded hole and is fixed to the front shell.

[0034] A PCBA board and PIN pins. One end of the PIN pin is fixedly connected to the PCBA board, and the other end of the PIN pin is connected to the PIN fixing groove.

[0035] A heat-conducting plate and a rear shell. The heat-conducting plate is located in the sandwich between the rear shell and the PCBA board. The rear shell is connected to the front shell, and a connector port is provided on the lower surface of the rear shell.

[0036] Further, the front shell and the rear shell are connected by laser welding, and the PIN pin and the PIN fixing groove are in interference fit.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows: By dynamically adjusting the data processing period and combining time series and multi-dimensional data, such as distance difference, temperature, signal strength, etc., the present invention establishes a detection data set, which can monitor the accuracy of the lidar module in real time. Through the identification of out-of-threshold data points, it can quickly judge whether there is an accuracy problem with the radar module, avoiding the lag of the traditional static detection method and making the detection process more efficient.

[0038] When it is detected that the module is inaccurate, the present invention can extract non-accurate features and compare them with historical detection data to analyze the error source, helping developers accurately identify the root cause of the performance degradation of the module. This diagnostic process can provide data support for subsequent optimization and improve the stability and reliability of the product.

[0039] By combining environmental parameters such as temperature influence coefficient and signal strength factor, the accuracy detection method of the present invention can effectively adapt to the influence of different environmental conditions on the performance of the lidar module. Especially in an environment with large temperature fluctuations, it can still maintain a high detection accuracy. This flexibility makes the method have stronger adaptability in practical applications.

[0040] By optimizing the module structure design, such as adopting an integrated lens barrel assembly and an optimized screw fixing structure, the number of components and complex assembly steps are reduced, thus simplifying the production process and reducing the production cost. At the same time, the need for debugging and repair is reduced, and the production efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0042] Figure 1 is a flowchart of a method for detecting the accuracy of a lidar module provided by an embodiment of the present invention.

[0043] Figure 2 is an exploded view of a lidar module at a certain angle provided by an embodiment of the present invention.

[0044] Figure 3 is an exploded view of a lidar module at another angle provided by an embodiment of the present invention.

[0045] In the figure, 10 is a light-transmitting cover plate; 20 is a front shell; 21 is a lens barrel hole; 22 is a lens barrel fixing post; 23 is a first threaded hole; 30 is a lens barrel assembly; 31 is a lens holder; 32 is a lens barrel; 33 is a lens barrel fixing hole; 34 is a second threaded hole; 35 is a PIN fixing groove; 36 is a fixing screw; 37 is a PIN needle; 38 is a PCBA board; 40 is a rear shell; 41 is a connector port. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] The exemplary embodiments disclosed in the present application will be described in more detail below with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully communicated to those skilled in the art. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0047] Refer to Figure 1As shown in the figure, an embodiment of the present invention provides a method for detecting the accuracy of a lidar module, including: S1, setting detection plates at several known positions, placing the lidar module to be tested at the detection points at the known positions, comparing the detected distance value of the lidar module to be tested with the actual distance value calculated based on the detection plate and the detection point, and adjusting the data processing cycle of the detected distance value according to the distance difference between the detected distance value and the actual distance value.

[0048] S2, when the data processing cycle is satisfied, obtaining the detection data within the data processing cycle, establishing a detection data set according to the time series, processing the detection data set, and determining whether the lidar module to be tested is accurate; when there is at least one out-of-threshold data point, determining that the lidar module to be tested is inaccurate, and when there is no out-of-threshold data point, determining that the lidar module to be tested is qualified.

[0049] S3, when it is determined that the lidar module to be tested is inaccurate, extracting the inaccurate features, comparing the inaccurate features with the historical detection data set, and determining the reason for inaccuracy.

[0050] S4, calculating the tolerance value according to the reason for inaccuracy and the inaccurate features, and sorting the lidar modules to be tested according to the size of the tolerance value.

[0051] The inaccurate features include: inaccurate data and inaccurate types. The inaccurate data is the detection data corresponding when it is determined that the lidar module to be tested is inaccurate, and the inaccurate type is the type of the corresponding detection data.

[0052] First, by comparing the lidar module to be tested with the detection plates at known positions, automatically calculating the difference between the actual distance value and the detected distance value, the data processing cycle can be accurately adjusted to ensure the accuracy and stability of the data. This real-time and automated adjustment mechanism can significantly improve the detection efficiency.

[0053] Second, when an out-of-threshold data point is detected, it can be timely determined whether the module is inaccurate, avoiding the errors caused by manual determination. In the case of detecting inaccuracy, the system can extract the inaccurate features and analyze them in combination with the historical detection data set, so as to accurately determine the reason for inaccuracy. This process not only helps to locate the problem, but also can take targeted solutions for specific problems.

[0054] In addition, through the calculation and sorting mechanism of the tolerance value, sorting the modules according to the inaccurate features and reasons can further optimize the production process, reduce the defective product rate, and improve the quality and reliability of the lidar module. Generally speaking, this method greatly improves the detection accuracy and production efficiency, reduces the production cost, can effectively ensure the accuracy of the lidar module under different environmental conditions, and has a wide application prospect.

[0055] In some embodiments of the present application, when determining whether a lidar module to be measured is accurate, it includes: obtaining a detection data set. The detection data in the detection data set is divided into four categories, including: distance difference, temperature, received signal strength of the lidar module to be measured, and detection time. Establish data groups for the detection data obtained when the detection times in the detection data set are the same, and obtain N data groups. Each data group includes distance difference, temperature, received signal strength of the lidar module to be measured, and detection time. Arrange the data groups to obtain an N*M matrix, where M = 4, and M is the number of types of detection data. The N*M matrix is represented by A i，j represents any element in the matrix, and A i，j is the element of the j-th category at the detection time i; select any element in the N*M matrix as the center point, set a data judgment range, the radius of the data judgment range is R, select all elements of the same category within the data judgment range, and calculate the average value; set an error threshold. When the difference between A i，j and the average value is greater than or equal to the error threshold, determine that A i，j is the out-of-threshold data point. When the difference between A i，j and the average value is less than the error threshold, determine that A i，j is the normal data point; when there is at least one out-of-threshold data point, determine that the lidar module to be measured is inaccurate.

[0056] Specifically, first, obtain a detection data set from the lidar module. This data set contains four types of data, and each type of data represents a specific measurement characteristic:

[0057] Distance difference: Represents the difference between the actually measured distance and the expected distance, reflecting the measurement accuracy.

[0058] Temperature: The operating temperature of the lidar module. Temperature has a significant impact on the radar performance and accuracy.

[0059] Received signal strength: The signal strength received by the radar. The signal strength directly affects the measurement accuracy.

[0060] Detection time: The time stamp of each detection, used to arrange the data according to time.

[0061] Divide the detection data with the same time stamp into a data group. Each data group contains these four types of data. Through these data groups, an N×M matrix can be obtained, where N represents the number of data groups, and M represents the number of types of data (4 in this example). The elements of the matrix correspond to specific data at a certain moment.

[0062] Determine the data judgment range and calculate the average value: Next, select any element in the N×M matrix as the center point, and set a data judgment range with a radius of R. Select the same type of data within this range (for example, all distance differences or all temperature values), and calculate the average value of these data. In this way, the stability and consistency of the data at that moment can be evaluated.

[0063] Set the error threshold and judge the out-of-threshold data points: According to the calculated average value, set an error threshold. When the difference between the detected data and the average value is greater than or equal to the error threshold, it is determined as an out-of-threshold data point, that is, there is a large abnormality or error in this data point. When the difference from the average value is less than the error threshold, it is determined as a normal data point.

[0064] Finally, any out-of-threshold data points in the detected data set are considered abnormal, resulting in the judgment that the laser radar module to be tested is "inaccurate". If there are no out-of-threshold data points in the detected data set, it is considered that the module is qualified under the test conditions and the accuracy meets the requirements.

[0065] This method can automatically judge the accuracy of the laser radar module, reduce manual intervention, and automatically detect whether the system meets the accuracy requirements. This helps to improve the detection efficiency and accuracy, and reduce human errors. By setting the error threshold and judging the out-of-threshold data points, abnormal data can be effectively identified, especially in the case of weak signal strength or large temperature fluctuations. In this way, the detection system can discover potential problems in advance and avoid inaccurate detection results from affecting the performance of the system. By using multiple data dimensions (such as distance difference, temperature, signal strength, and time) for comprehensive analysis, the accuracy of the laser radar module can be more comprehensively evaluated, and the reliability and stability of the system can be improved. Through the processing of time series data and matrix analysis, not only can the accuracy of the detected data be evaluated in real time, but also the detection parameters (such as error threshold and judgment range) can be optimized according to historical data, thereby improving the long-term detection accuracy and efficiency. By setting the data judgment range and calculating the average value, the system can effectively reduce the influence of single data fluctuations or instantaneous errors on the judgment accuracy, enhance the fault tolerance of the overall test, and ensure the stability of the accuracy detection. Since this method takes into account the influence of temperature and signal strength, it can automatically adjust the detection standard according to environmental changes to ensure that the laser radar module can maintain high accuracy in various different working environments.

[0066] In some embodiments of the present application, when determining the reason for inaccuracy, it includes: extracting inaccurate features and judging the reason for inaccuracy according to the overlap degree between the inaccurate features and the historical detection data set.

[0067] The overlap degree satisfies the following relationship:

[0068] ;

[0069] where S is the overlap degree, x i is the i-th data point in the inaccurate data, is the mean value of all inaccurate data, y i is the i-th data point in the historical dataset, is the mean value of all data points in the historical dataset, and n is the number of all inaccurate data.

[0070] Match the inaccurate data with the highest similarity with the historical dataset, and determine the inaccurate reason of the inaccurate data according to the inaccurate reason in the historical dataset.

[0071] It should be noted that the historical detection dataset contains several known reasons for abnormal performance of lidar modules and their characteristic data.

[0072] By calculating the overlap degree between the inaccurate data and the historical data, the similarity between the two can be accurately quantified, so as to match the closest historical inaccurate data. This data-driven method is more objective and stable compared with simple rule matching or empirical judgment. This method dynamically adjusts the matching rules based on historical data. Even if the distribution of inaccurate data changes, it can still adaptively identify different types of inaccurate reasons through the continuously updated historical detection dataset, improving the generalization ability and robustness of the system. By calculating the similarity, quickly locate the most similar inaccurate data in the historical dataset, and infer the reason of the current inaccurate data based on the historical record, avoiding the inefficient process of manual analysis or trial and error required by traditional methods, and greatly improving the efficiency of fault diagnosis and problem tracing. Using mathematical formulas to quantify the similarity between inaccurate data and historical data can effectively reduce misjudgments caused by subjective judgment, noise interference or random errors, and improve the consistency and reliability of the system's identification of inaccurate reasons. This method can be widely applied to scenarios such as equipment fault detection, data quality analysis, and automatic control systems, and is applicable to any occasion that requires analyzing data accuracy and finding out the reasons for anomalies, with good versatility.

[0073] In some embodiments of the present application, when calculating the tolerance value according to the inaccurate reason and inaccurate feature, it includes:

[0074] ;

[0075] where H is the tolerance value, d is the maximum distance difference in the inaccurate data, d max is the maximum distance difference in the historical dataset, R is the minimum received signal strength in the inaccurate data, R min is the minimum received signal strength in the historical dataset, K is the average temperature in the inaccurate data, is the average temperature in the historical dataset, α is the distance difference influence coefficient, β is the signal strength influence coefficient, and η is the temperature influence coefficient.

[0076] It should be noted that the formula takes into account factors such as distance difference, signal strength, and temperature, and assigns corresponding influence coefficients to each factor, making the calculation results more comprehensive and accurate. By comparing the inaccurate data with the historical dataset (maximum / minimum or mean), the degree to which the inaccurate data deviates from the normal situation can be effectively measured. Since the formula uses normalization processing, data of different scales can be compared uniformly, avoiding biases caused by different numerical sizes of certain factors.

[0077] The influence coefficients α, β, and η can be dynamically adjusted according to the specific scenario, enabling the system to adapt to different application environments and improve the generalization ability. The formula has a simple form and small computational complexity, and is suitable for real-time analysis and rapid detection. Especially in the case of a large amount of data, it can effectively reduce the computational overhead. By comparing the deviation degrees of inaccurate data and historical data, it can effectively identify which data are tolerable and which are abnormal, thereby reducing false alarms and missed detections. Combining historical data, the existing experience can be used to optimize the detection logic and improve the intelligence level of the system. This calculation method can be widely used in multiple fields such as wireless communication, environmental monitoring, and sensor data analysis, helping to optimize device performance, improve data quality and stability.

[0078] In some embodiments of the present application, the distance difference influence coefficient α is obtained through the following relationship:

[0079] ;

[0080] The signal strength influence coefficient β is obtained through the following relationship:

[0081] ;

[0082] The temperature influence coefficient η is obtained through the following relationship:

[0083] ;

[0084] ;

[0085] where f diff is the variance of all distance differences in the inaccurate data, f R is the variance of all signal strength values in the inaccurate data, f K is the variance of all temperature values in the inaccurate data, f SUM is the total variance.

[0086] It should be noted that traditional methods usually require manual setting of the weights of various influencing factors, while this method automatically determines the weights through a data statistical method (based on variance calculation), reducing human intervention and improving the level of intelligence. This data-driven approach can better adapt to different scenarios, such as wireless communication environments, sensor data analysis, industrial measurements, etc. Since the variance between non-exact data and exact data directly affects the calculation of each influence coefficient, it can more accurately identify which factors have the greatest impact on errors, thus being more precise in anomaly detection. For example, when the variance of the distance difference is significantly greater than the variance of the signal strength or temperature, the system will automatically increase the weight of α to ensure that this factor receives greater attention and reduce misjudgments caused by improper weight settings. The calculation method is based on variance normalization and only involves simple addition, division, and variance calculations, with low computational complexity, suitable for real-time data analysis and large-scale data processing, and can improve the real-time performance and response speed of the system. This method can be widely applied in fields such as wireless communication, environmental monitoring, intelligent sensing, data quality control, etc., especially in scenarios involving error analysis and data accuracy optimization, where it can significantly improve the accuracy and stability of data processing

[0087] Refer to Figures 2 - 3 As shown, an embodiment of the present invention further provides a lidar module, which includes: a front shell 20, which is connected to a light-transmitting cover plate 10 above it. The front shell 20 is provided with a plurality of lens barrel holes 21, and the lower surface of the front shell 20 is provided with lens barrel fixing posts 22 and first threaded holes 23.

[0088] A lens barrel assembly 30, including a lens base 31 and a lens barrel 32. The lens base 31 and the lens barrel 32 are of an integral structure. The lens base 31 is provided with a second threaded hole 34 and a lens barrel fixing hole 33, and a PIN fixing groove 35 is provided on the lower surface of the lens base 31; the lens barrel fixing post 22 is inserted into the lens barrel fixing hole 33 to fix the lens barrel assembly 30.

[0089] A fixing screw 36, which passes through the first threaded hole 23 and the second threaded hole 34 and is fixed to the front shell 20.

[0090] A PCBA board 38 and a PIN needle 37. One end of the PIN needle 37 is fixedly connected to the PCBA board 38, and the other end of the PIN needle 37 is connected to the PIN fixing groove 35.

[0091] A heat conduction plate and a rear shell 40. The heat conduction plate is located in the sandwich between the rear shell 40 and the PCBA board 38. The rear shell 40 is connected to the front shell 20, and a connector port 41 is provided on the lower surface of the rear shell 40.

[0092] In some embodiments of the present application, the front shell 20 and the rear shell 40 are connected by laser welding, and the PIN needle 37 and the PIN fixing groove 35 are in interference fit.

[0093] It should be noted that the lens barrel assembly 30 can be two or three. When there are two lens barrel assemblies 30, one lens is used for receiving and one for transmitting; when there are three lens barrel assemblies 30, one lens for receiving and two lenses for transmitting are provided. The heat conducting plate is not shown in Figure 2 and Figure 3 . In this embodiment, the heat conducting plate is preferably a graphene thin plate, and its function is to dissipate heat. Other plate-shaped material structures for heat dissipation purposes can also be used.

[0094] Compared with the traditional screw fixation or gluing methods, laser welding has higher strength, sealing performance and durability, which can effectively improve the structural strength of the module and reduce displacement or loosening caused by vibration, impact or temperature change. This method can also effectively prevent external factors such as dust and water vapor from entering, improve the dust and waterproof performance, and is suitable for various complex environments (such as autonomous driving, industrial inspection, etc.). The integration of the lens seat 31 and the lens barrel 32 reduces the tolerance accumulation during the assembly process, improves the alignment accuracy of the optical system, and thus enhances the measurement accuracy and stability of the lidar. The lens barrel fixing post 22 is fixed by plugging, which improves the seismic performance of the lens barrel assembly 30, prevents the lens barrel 32 from shifting in position due to long-term use, and ensures the long-term stability of the lidar. The lens barrel assembly 30 is fixed to the front shell 20 through the lens barrel fixing post 22, and the fixing screw 36 passes through the front shell 20 and the lens seat 31. To ensure precise alignment, multi-point fixing is carried out using the fixing screw 36 to ensure the tight connection and optical axis consistency between the lens barrel assembly 30 and the front shell 20, reduce optical offset caused by external force or temperature change, and ensure the long-term stable lidar ranging accuracy. The interference fit makes the PIN pin 37 and the PIN fixing groove 35 maintain a tight connection, which can effectively avoid poor contact caused by factors such as vibration, thermal expansion and contraction, and improve the stability of signal transmission. This design reduces the contact resistance of the traditional plug-in connection, improves the electrical reliability of the lidar module, and ensures the integrity of data transmission. The fixed connection method is adopted to avoid the solder joint cracking caused by mechanical fatigue or vibration, improve the reliability of the PCBA board 38, and ensure the stable operation of the circuit. This structure can effectively improve the heat conduction efficiency, quickly conduct the heat on the PCBA board 38 to the rear shell 40, and dissipate it to the external environment through the rear shell 40, avoiding performance degradation or failure of components due to high temperature. This design is particularly suitable for lidars operating under high load for a long time, such as application scenarios of autonomous driving and industrial inspection. The front shell 20 is provided with a plurality of lens barrel holes 21, which can adapt to different types of lens barrel assemblies 30. This design allows the lidar module to be modularly configured in different application scenarios, and the lens barrel assembly 30 can be flexibly replaced to meet the requirements of different focal lengths, field of view (FOV) or detection distances, improving the versatility and scalability of the product. The connector port 41 is located on the lower surface of the rear shell 40, facilitating the docking of external devices. The design of the connector port 41 optimizes the wiring method, reduces the cable bending, and improves the overall installation convenience.

[0095] The traditional lens barrel 32 and lens base 31 need to be assembled with glue. In the present invention, the lens barrel 32 and lens base 31 are directly produced as a whole, saving the glue assembly process and production process. The PIN pin 37 is fixed to the aluminum shell through a crimping process and is used for PCBA positioning; the PIN pin 37 has an interference fit with the bottom hole and will not become loose under high and low temperature environments and long-term vibration. The connection between the front shell 20 and the rear shell 40 is achieved by laser welding. If a sealant is used to encapsulate the front shell 20 and the rear shell 40, there is a risk of aging after long-term use, which affects the lens. The front and rear shells 40 are encapsulated using the laser welding process, permanently achieving the effects of acid resistance, water resistance, fog resistance, and dust resistance.

[0096] After the PCBA board 38 and the optical component AA are focused, only simple soldering fixation is required, and the operation is simple and efficient.

[0097] The soldering process does not require curing and is quickly welded after AA, ensuring the accuracy consistency of the optical axis.

[0098] Through multi-head laser welding, multiple positions are synchronously welded, ensuring the uniformity of heat distribution.

[0099] 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 should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for detecting the accuracy of a laser radar module, characterized in that: include: Setting up several detection boards with known positions, placing the laser radar module to be tested at the detection point at the known position, comparing the detection distance value of the laser radar module to be tested with the actual distance value calculated according to the detection board and the detection point, and adjusting the data processing cycle of the detection distance value according to the distance difference between the detection distance value and the actual distance value; When the data processing cycle is satisfied, the detection data within the data processing cycle is acquired, and a detection data set is established according to the time series, and the detection data set is processed to determine whether the laser radar module to be tested is accurate; when there is at least one over-threshold data point, the laser radar module to be tested is determined to be inaccurate, and when there is no over-threshold data point, the laser radar module to be tested is determined to be qualified; When it is determined that the laser radar module to be tested is inaccurate, extracting inaccurate features, comparing the inaccurate features with historical detection data sets, and determining the cause of the inaccuracy; A tolerable value is calculated according to the inaccurate cause and the inaccurate feature, and the laser radar modules to be tested are sorted according to the size of the tolerable value.

2. The accuracy detection method of the laser radar module according to claim 1 is characterized in that: The inaccurate features include: inaccurate data and inaccurate types. The inaccurate data is the detection data corresponding to when the laser radar module to be tested is judged to be inaccurate, and the inaccurate type is the type of the corresponding detection data.

3. The accuracy detection method of the laser radar module according to claim 2 is characterized in that: When judging whether the laser radar module to be tested is accurate, it includes: The detection data set is obtained, and the detection data in the detection data set is divided into four categories, including: distance difference, temperature, received signal strength of the laser radar module to be tested, and detection time. The detection data obtained when the detection time in the detection data set is the same is used to establish a data group to obtain N data groups, each of which includes distance difference, temperature, received signal strength of the laser radar module to be tested, and detection time. According to the data groups, an N*M matrix is ​​obtained by arranging, wherein M=4, M is the number of types of the detection data, and the N*M matrix is ​​obtained by A i，j Represents any element in the matrix, A i，j is the element of the jth class at the detection time i; Select any element in the N*M matrix as the center point, set the data judgment range, the radius of the data judgment range is R, select all elements of the same type within the data judgment range, and calculate the average value; Set the error threshold. i，j When the difference between the average value and the average value is greater than or equal to the error threshold, it is determined that A i，j is the above-threshold data point, when A i，j When the difference between the average value and the average value is less than the error threshold, it is determined that A i，j is a normal data point; When there is at least one over-threshold data point, it is determined that the laser radar module to be tested is inaccurate.

4. The accuracy detection method of the laser radar module according to claim 3 is characterized in that: When an imprecise cause is determined, include: Extracting the inaccurate features, and determining the inaccurate causes according to the overlap between the inaccurate features and the historical detection data set; The overlap satisfies the following relationship: ； Among them, S is the overlap, x i is the i-th data point in the inexact data, is the mean of all inexact data, y i is the i-th data point in the historical data set, is the mean of all data points in the historical data set, and n is the number of all inaccurate data; The inaccurate data with the highest similarity is matched with the historical data set, and the inaccurate cause of the inaccurate data is determined according to the inaccurate cause in the historical data set.

5. The method for detecting the accuracy of a laser radar module according to claim 4, characterized in that: When calculating the tolerable value based on the inaccurate causes and inaccurate characteristics, it includes: ； Among them, H is the tolerable value, d is the maximum distance difference in the inaccurate data, and d max is the maximum distance difference in the historical data set, R is the minimum received signal strength in the inaccurate data, and R min is the minimum received signal strength in the historical data set, K is the average temperature in the inaccurate data, is the average temperature in the historical data set, α is the distance difference influence coefficient, β is the signal strength influence coefficient, and η is the temperature influence coefficient.

6. The method for detecting the accuracy of a laser radar module according to claim 5, characterized in that: The distance difference influence coefficient α is obtained through the following relationship: ； The signal strength influence coefficient β is obtained through the following relationship: ； The temperature influence coefficient η is obtained through the following relationship: ； ； Among them, f diff is the variance of all distance differences in the inexact data, f R is the variance of all signal strength values ​​in the inaccurate data, f K is the variance of all temperature values ​​in the inexact data, f SUM is the total variance.

7. A laser radar module, obtained by using the laser radar module accuracy detection method according to any one of claims 1 to 6, characterized in that: The laser radar module includes: A front shell, the upper part of which is connected to the light-transmitting cover plate, the front shell is provided with a plurality of lens barrel holes, and the lower surface of the front shell is provided with a lens barrel fixing pile and a first threaded hole; The lens barrel assembly comprises a lens seat and a lens barrel, wherein the lens seat and the lens barrel are an integrated structure, the lens seat is provided with a second threaded hole and a lens barrel fixing hole, and a PIN fixing groove is provided on the lower surface of the lens seat; the lens barrel fixing pile is inserted into the lens barrel fixing hole to fix the lens barrel assembly; A fixing screw passing through the first threaded hole and the second threaded hole and fixed on the front shell; A PCBA board and a PIN pin, one end of the PIN pin is fixedly connected to the PCBA board, and the other end of the PIN pin is connected to the PIN fixing groove; A heat conducting plate and a rear shell, wherein the heat conducting plate is located in the interlayer between the rear shell and the PCBA board, the rear shell is connected to the front shell, and a connector port is disposed on the lower surface of the rear shell.

8. The laser radar module according to claim 7, characterized in that: The front shell and the rear shell are connected by laser welding, and the PIN pin and the PIN fixing groove are interference fit.

Images