Data cleaning method and device, intelligent driving vehicle and storage medium

Abstract

The present disclosure provides a data cleaning method and device, an intelligent driving vehicle and a storage medium, and relates to the technical field of computers. The data cleaning method comprises the following steps: calculating the number of adjacent points of each target point in point cloud data; determining a threshold value of the number of adjacent points of each target point according to the type of each target point and the distance between each target point and the intelligent driving vehicle; and in the case that the number of adjacent points of the i-th target point is less than the threshold value of the number of adjacent points of the i-th target point, deleting the i-th target point from the point cloud data, wherein N is the total number of target points in the point cloud data, so as to obtain cleaned point cloud data.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular to a data cleaning method and device, an intelligent driving vehicle and a storage medium. BACKGROUND

[0002] The active safety module is an important safety guarantee module in the intelligent driving system, which usually contains two parts of the perception module and the decision module. In the process of vehicle driving, the perception module processes the sensor data in real time to obtain the motion state estimation value (such as speed) of the target point, and the decision module makes a decision based on the output of the perception module to determine whether the vehicle is in a dangerous state and whether there is a conflict risk with the target object, so as to execute emergency braking at the necessary moment to ensure the safety of the vehicle. SUMMARY

[0003] In the prior art, the perception module perceives the motion state of the point-level target. Since the same object usually contains hundreds of discrete target points, this will make the model lack strict consistency constraints on the output values of the target points contained in the same object, especially in the part adjacent to the background at the edge of the object, which often contains a large number of outliers. These outliers will seriously interfere with the downstream decision control and affect the performance of the active safety module.

[0004] Accordingly, the present disclosure provides a data cleaning method which can effectively clean outliers to control the driving of the intelligent driving vehicle safely.

[0005] According to a first aspect of the embodiments of the present disclosure, a data cleaning method is provided, which is executed by a data cleaning device in an intelligent driving vehicle, comprising: calculating the number of adjacent points of each target point in point cloud data; determining the number of adjacent points threshold of each target point according to the type of each target point and the distance between each target point and the intelligent driving vehicle; in the case that the number of adjacent points of the i th target point is less than the number of adjacent points threshold of the i th target point, deleting the i th target point from the point cloud data, , N is the total number of target points in the point cloud data, to obtain cleaned point cloud data.

[0006] In some embodiments, the calculation of the number of adjacent points of each target point in the point cloud data comprises: judging whether the i th target point and the j th target point belong to the same target according to the position, speed and type of the i th target point and the position, speed and type of the j th target point, , and ; in the case that the i th target point and the j th target point belong to the same target, adding 1 to the number of adjacent points of the i th target point and adding 1 to the number of adjacent points of the j th target point.

[0007] In some embodiments, the determining whether the ith target point and the jth target point belong to the same target comprises: in a case that a type of the ith target point and a type of the jth target point are the same, a position error of the ith target point and the jth target point is less than a position error threshold, a velocity error or a velocity relative error rate of the ith target point and the jth target point is less than a predetermined threshold, a velocity direction cosine error of the ith target point and the jth target point is less than a cosine error threshold, and a product of the velocity relative error rate and the velocity direction cosine error is less than a critical threshold, determining that the ith target point and the jth target point belong to the same target.

[0008] In some embodiments, the determining whether the ith target point and the jth target point belong to the same target comprises: in a case that the type of the ith target point and the type of the jth target point are the same, calculating the position error of the ith target point and the jth target point; in a case that the position error is less than the position error threshold, calculating the velocity error and the velocity relative error rate of the ith target point and the jth target point; in a case that the velocity error is less than a velocity error threshold, or the velocity relative error rate is less than a velocity relative error rate threshold, calculating the velocity direction cosine error of the ith target point and the jth target point; in a case that the velocity direction cosine error is less than the cosine error threshold, calculating the product of the velocity relative error rate and the velocity direction cosine error; and in a case that the product of the velocity relative error rate and the velocity direction cosine error is less than a critical threshold, determining that the ith target point and the jth target point belong to the same target.

[0009] In some embodiments, the position error threshold is associated with the type of the jth target point.

[0010] In some embodiments, the calculating the velocity relative error rate of the ith target point and the jth target point comprises: determining the velocity relative error rate according to a maximum value of a velocity of the ith target point and a velocity of the jth target point, and the velocity error of the ith target point and the jth target point.

[0011] In some embodiments, the velocity relative error rate is positively correlated with the velocity error, and inversely correlated with the maximum value.

[0012] In some embodiments, the determining the velocity relative error rate comprises: determining the velocity relative error rate according to a ratio of the velocity error and the maximum value.

[0013] In some embodiments, calculating the velocity direction cosine error of the i-th target point and the j-th target point includes: determining the velocity direction cosine error based on the velocity of the i-th target point, the component of the velocity of the i-th target point in a first direction, the component of the velocity of the i-th target point in a second direction, the velocity of the j-th target point, the component of the velocity of the j-th target point in the first direction, and the component of the velocity of the j-th target point in the second direction, wherein the first direction is perpendicular to the second direction.

[0014] In some embodiments, determining the velocity direction cosine error includes: multiplying the velocity component of the i-th target point in a first direction and the velocity component of the j-th target point in a first direction to obtain a first intermediate quantity; multiplying the velocity component of the i-th target point in a second direction and the velocity component of the j-th target point in a second direction to obtain a second intermediate quantity; adding the first intermediate quantity and the second intermediate quantity to obtain a third intermediate quantity; multiplying the velocity of the i-th target point and the velocity of the j-th target point to obtain a fourth intermediate quantity; dividing the third intermediate quantity by the fourth intermediate quantity to obtain a fifth intermediate quantity; and determining the velocity direction cosine error based on the difference between a predetermined parameter and the fifth intermediate quantity.

[0015] In some embodiments, the threshold number of adjacent points for each target point is positively correlated with the distance between each target point and the intelligent driving vehicle.

[0016] In some embodiments, determining the adjacency threshold for each target point includes: obtaining a first category threshold and a second category threshold corresponding to the type of each target point, wherein the first category threshold is greater than the second category threshold; subtracting the first category threshold from the second category threshold to obtain a sixth intermediate value; obtaining a seventh intermediate value based on the product of the sixth intermediate value and the distance between each target point and the intelligent driving vehicle; and adding the first category threshold and the seventh intermediate value to obtain the adjacency threshold for each target point.

[0017] In some embodiments, the cleaned point cloud data is sent to the vehicle control device in the intelligent driving vehicle.

[0018] According to a second aspect of the present disclosure, a data cleaning apparatus is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to execute instructions stored in the memory to implement the data cleaning method as described in any of the above embodiments.

[0019] According to a third aspect of the present disclosure, an intelligent driving vehicle is provided, comprising: a data cleaning apparatus as described in any of the above embodiments; a lidar configured to collect point cloud data of targets around the intelligent driving vehicle and send the point cloud data to the data cleaning apparatus; and a vehicle control device configured to control the intelligent driving vehicle based on the cleaned point cloud data sent by the data cleaning apparatus.

[0020] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method as described in any of the above embodiments.

[0021] According to a fifth aspect of the present disclosure, a computer program product is provided, including computer instructions, wherein the computer instructions, when executed by a processor, implement the method as described in any of the above embodiments.

[0022] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a data cleaning method according to an embodiment of the present disclosure;

[0025] Figure 2 This is a flowchart illustrating a method for determining whether two target points belong to the same target according to an embodiment of this disclosure;

[0026] Figure 3 This is a schematic diagram of the structure of a data cleaning apparatus according to an embodiment of the present disclosure;

[0027] Figure 4 This is a schematic diagram of the structure of an intelligent driving vehicle according to an embodiment of the present disclosure;

[0028] Figure 5 Speed ​​prediction graph for data that has not been cleaned;

[0029] Figure 6 A speed prediction chart for data cleaning. Detailed Implementation

[0030] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0031] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of this disclosure.

[0032] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0033] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0034] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0035] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0036] Figure 1 This is a schematic flowchart of a data cleaning method according to an embodiment of the present disclosure. In some embodiments, the following data cleaning method is performed by a data cleaning device in an intelligent driving vehicle, including steps 11-13.

[0037] In step 11, the number of adjacent points for each target point in the point cloud data is calculated.

[0038] In some embodiments, based on the position, velocity, and type of the i-th target point and the position, velocity, and type of the j-th target point, it is determined whether the i-th target point and the j-th target point belong to the same target. N is the total number of target points in the point cloud data, and If the i-th target point and the j-th target point belong to the same target, increment the number of adjacent points of the i-th target point by 1, and increment the number of adjacent points of the j-th target point by 1.

[0039] In some embodiments, if the types of the i-th target point and the j-th target point are the same, the position error between the i-th target point and the j-th target point is less than the position error threshold, the velocity error or velocity relative error rate between the i-th target point and the j-th target point is less than a predetermined threshold, the velocity direction cosine error between the i-th target point and the j-th target point is less than the cosine error threshold, and the product of the velocity relative error rate and the velocity direction cosine error is less than a critical threshold, then the i-th target point and the j-th target point are determined to belong to the same target.

[0040] In some embodiments, the step of determining whether the i-th target point and the j-th target point belong to the same target is as follows: Figure 2 As shown.

[0041] Figure 2 This is a flowchart illustrating a method for determining whether two target points belong to the same target according to an embodiment of this disclosure. In some embodiments, the following method for determining whether two target points belong to the same target is executed by a data cleaning device in an intelligent driving vehicle, including steps 21-27.

[0042] In step 21, it is determined whether the types of the i-th target point and the j-th target point are the same, where, ,and N is the total number of target points.

[0043] If the type of the i-th target point is the same as the type of the j-th target point, proceed to step 22. Otherwise, proceed to step 27.

[0044] It should be noted that if two target points are of different types, it means that the two target points do not belong to the same target, and in this case, no further processing is required for these two target points.

[0045] In step 22, the position error between the i-th target point and the j-th target point is calculated, and it is determined whether the position error between the i-th target point and the j-th target point is greater than or equal to the position error threshold.

[0046] If the position error is greater than or equal to the position error threshold, proceed to step 27. Otherwise, if the position error is less than the position error threshold, proceed to step 23.

[0047] For example, if the position of the i-th target point is (Xi, Yi) and the position of the j-th target point is (Xj, Yj), then the position error loc_error between the i-th target point and the j-th target point is as shown in formula (1).

[0048] loc_error =||(Xi-Xj, Yi-Yj)|| (1)

[0049] It should be noted that if the positional error between two target points is greater than or equal to the positional error threshold, it indicates that the two target points do not belong to the same target, and in this case, no further processing is required for these two target points.

[0050] In some embodiments, the position error threshold is associated with the type of the j-th target point.

[0051] For example, the position error threshold can be represented as loc_thresh[j_objType].

[0052] For example, the position error thresholds are shown in Table 1.

[0053] Table 1.

[0054] In step 23, calculate the velocity error and relative velocity error rate between the i-th target point and the j-th target point, and determine whether the velocity error is greater than or equal to the velocity error threshold and whether the relative velocity error rate is greater than or equal to the relative velocity error rate threshold.

[0055] If the speed error is greater than or equal to the speed error threshold, and the speed relative error rate is greater than or equal to the speed relative error rate threshold, proceed to step 27. Otherwise, if the speed error is less than the speed error threshold, or the speed relative error rate is less than the speed relative error rate threshold, proceed to step 24.

[0056] For example, if the velocity of the i-th target point is Vi and the velocity of the j-th target point is Vj, then the velocity error V_diff between the i-th target point and the j-th target point is as shown in formula (2).

[0057] V_diff=abs(Vi-Vj) (2)

[0058] In formula (2), abs is an absolute value function.

[0059] In some embodiments, the relative speed error rate is determined based on the maximum value of the speed of the i-th target point and the speed of the j-th target point, and the speed error between the i-th target point and the j-th target point.

[0060] In some embodiments, the relative speed error rate is positively correlated with the speed error and negatively correlated with the maximum value.

[0061] For example, the relative speed error rate can be determined based on the ratio of the speed error to the maximum value.

[0062] For example, the velocity error between the i-th target point and the j-th target point is shown in formula (3).

[0063] velocity_error= V_diff / max(Vi,Vj) / a (3)

[0064] In formula (3), max is the maximum value function and a is a parameter, for example, a is 0.1.

[0065] For example, the speed error threshold is 0.5 m / s, and the speed relative error rate threshold is 1.

[0066] It should be noted that V_diff is used to represent the difference in magnitude between two velocity vectors. If V_diff is greater than or equal to the velocity error threshold and velocity_error is greater than or equal to the velocity relative error rate threshold, it indicates that the error in the velocity magnitude is large, and the i-th target point and the j-th target point should not belong to the same object.

[0067] In step 24, calculate the velocity direction cosine error between the i-th target point and the j-th target point, and determine whether the velocity direction cosine error is greater than or equal to the cosine error threshold.

[0068] If the cosine error in the velocity direction is greater than or equal to the cosine error threshold, proceed to step 27. Otherwise, if the cosine error in the velocity direction is less than the cosine error threshold, proceed to step 25.

[0069] In some embodiments, a velocity direction cosine error is determined based on the velocity of the i-th target point, the component of the velocity of the i-th target point in a first direction, the component of the velocity of the i-th target point in a second direction, the velocity of the j-th target point, the component of the velocity of the j-th target point in a first direction, and the component of the velocity of the j-th target point in a second direction, wherein the first direction is perpendicular to the second direction.

[0070] For example, the steps to determine the cosine error of the velocity direction include the following.

[0071] 1) Multiply the velocity component of the i-th target point in the first direction and the velocity component of the j-th target point in the first direction together to obtain the first intermediate quantity.

[0072] 2) Multiply the velocity component of the i-th target point in the second direction and the velocity component of the j-th target point in the second direction to obtain the second intermediate quantity.

[0073] 3) Add the first intermediate quantity and the second intermediate quantity to obtain the third intermediate quantity.

[0074] 4) Multiply the velocity of the i-th target point and the velocity of the j-th target point to obtain the fourth intermediate quantity.

[0075] 5) Divide the third intermediate by the fourth intermediate to obtain the fifth intermediate.

[0076] 6) Determine the velocity direction cosine error based on the difference between the predetermined parameters and the fifth intermediate quantity.

[0077] For example, the cosine error of the velocity direction of the i-th target point and the j-th target point is shown in formula (4).

[0078] cosine_error=(b-(Vxi×Vxj+Vyi×Vyj) / (Vi×Vj)) / c (4)

[0079] In formula (4), Vxi is the component of the velocity Vi of the i-th target point in the X direction, Vyi is the component of the velocity Vi of the i-th target point in the y direction, Vxj is the component of the velocity Vj of the j-th target point in the X direction, and Vyj is the component of the velocity Vj of the j-th target point in the y direction. For example, parameter b is 1 and parameter c is 0.016.

[0080] For example, the velocity Vi of the i-th target point is shown in formula (5), and the velocity Vj of the j-th target point is shown in formula (6).

[0081] Vi = ||(Vxi,Vyi)|| (5)

[0082] Vj=||(Vxj,Vyj)|| (6)

[0083] For example, the cosine error threshold is 1.

[0084] It should be noted that if the cosine error of the velocity direction between two target points is greater than or equal to the cosine error threshold, it indicates that there is a large deviation in the direction of the velocity vectors of the two target points, that is, the i-th target point and the j-th target point should not belong to the same object.

[0085] In step 25, the product of the relative speed error rate and the cosine error of the speed direction is calculated, and it is determined whether the product of the relative speed error rate and the cosine error of the speed direction is greater than or equal to the critical threshold.

[0086] If the product of the relative velocity error rate and the velocity direction cosine error is greater than or equal to the critical threshold, proceed to step 27. Otherwise, if the product of the relative velocity error rate and the velocity direction cosine error is less than the critical threshold, proceed to step 26.

[0087] For example, the critical threshold is 0.6.

[0088] It's important to note that even if the parameters `velocity_error` and `cosine_error` pass the checks, the target point may still be in a critical state. In such cases, the product of `velocity_error` and `cosine_error` is used as a constraint to comprehensively consider the magnitude and direction errors of the velocity vector. This effectively removes target points in a critical state.

[0089] In step 26, it is determined that the i-th target point and the j-th target point belong to the same target.

[0090] In step 27, it is determined that the i-th target point and the j-th target point do not belong to the same target.

[0091] return Figure 1 In step 12, based on the type of each target point and the distance between each target point and the autonomous driving vehicle, a threshold for the number of adjacent points of each target point is determined.

[0092] In some embodiments, the threshold number of adjacent points for each target point is positively correlated with the distance between each target point and the intelligent driving vehicle.

[0093] For example, the steps to determine the threshold number of adjacent points for each target point include the following.

[0094] 1) Obtain the first category threshold and the second category threshold corresponding to the type of each target point, where the first category threshold is greater than the second category threshold.

[0095] 2) Subtract the first category threshold from the second category threshold to obtain the sixth intermediate value.

[0096] 3) The seventh intermediate quantity is obtained by multiplying the sixth intermediate quantity by the distance between each target point and the intelligent driving vehicle.

[0097] 4) Add the first category threshold to the seventh intermediate value to obtain the threshold for the number of adjacent points of each target point.

[0098] For example, the first category threshold for each target point's type is Th1, the second category threshold is Th2, and Th1 is greater than Th2.

[0099] For example, for target points with different types, the values ​​of Th1 and Th2 are shown in Table 2.

[0100] Table 2.

[0101] It should be noted that, because different targets have different surface areas, the distance between the target and the autonomous vehicle will vary depending on whether the target is close to or far away from the vehicle.

[0102] For example, the threshold for the number of adjacent points of each target point, adj_count_thresh, is shown in formula (7).

[0103] adj_count_thresh=Th1+(Th2-Th1)×(||(x,y)|| / d) (7)

[0104] In formula (7), ||(x,y)|| represents the distance between the target point and the autonomous driving vehicle. For example, the parameter d is 40.

[0105] In step 13, if the number of adjacent points of the i-th target point is less than the threshold for the number of adjacent points of the i-th target point, the i-th target point is deleted from the point cloud data. N is the total number of target points in the point cloud data, in order to obtain cleaned point cloud data.

[0106] In some embodiments, the cleaned point cloud data is sent to the vehicle control unit in the intelligent driving vehicle so that the vehicle control unit can perform safe control of the driving of the intelligent driving vehicle.

[0107] In the data cleaning method provided in the above embodiments of this disclosure, the number of neighboring points of each target point in the point cloud data is calculated; a threshold for the number of neighboring points of each target point is determined according to the type of each target point and the distance between each target point and the intelligent driving vehicle; if the number of neighboring points of the i-th target point is less than the threshold for the number of neighboring points of the i-th target point, the i-th target point is deleted from the point cloud data, thereby effectively cleaning outliers and enabling safe control of the driving of the intelligent driving vehicle.

[0108] Figure 3 This is a schematic diagram of the structure of a data cleaning apparatus according to an embodiment of the present disclosure.

[0109] like Figure 3 As shown, the data cleaning apparatus 30 is presented in the form of a general-purpose computing device. The data cleaning apparatus 30 includes a memory 31, a processor 32, and a bus 33 connecting different system components.

[0110] The memory 31 may include, for example, system memory, non-volatile storage media, etc. System memory may store, for example, an operating system, application programs, a boot loader, and other programs. System memory may include volatile storage media, such as random access memory (RAM) and / or cache memory. Non-volatile storage media may store, for example, instructions for a corresponding embodiment of at least one data cleaning method being executed. Non-volatile storage media include, but are not limited to, disk storage, optical storage, flash memory, etc.

[0111] Processor 32 can be implemented using a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA) or other programmable logic devices, discrete hardware components such as discrete gates or transistors. Accordingly, each module, such as the acquisition module, calculation module, and adjustment module, can be implemented by executing instructions in the central processing unit (CPU) running memory to perform the corresponding steps, or by implementing dedicated circuitry to perform the corresponding steps.

[0112] For example, processor 32 is configured for memory-based instruction execution implementation such as Figure 1 and Figure 2 The method involved in any of the embodiments.

[0113] Bus 33 can use any of the various bus architectures. For example, bus architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MCA) bus, and the Peripheral Component Interconnect (PCI) bus.

[0114] The interfaces 34, 35, and 36 of the data cleaning device 30, as well as the memory 31 and processor 32, can be connected via bus 33. Input / output interface 34 provides a connection interface for input / output devices such as monitors, mice, and keyboards. Network interface 35 provides a connection interface for various networked devices. Storage interface 36 provides a connection interface for external storage devices such as floppy disks, USB flash drives, and SD cards.

[0115] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations thereof, can be implemented by computer-readable program instructions.

[0116] These computer-readable program instructions are provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable device to produce a machine, such that execution of the instructions by the processor produces means for implementing the functions specified in one or more boxes of the flowchart and / or block diagram.

[0117] These computer-readable program instructions may also be stored in a computer-readable storage medium. These instructions cause a computer to work in a particular manner to produce an article of manufacture, including instructions that implement the functions specified in one or more boxes in a flowchart and / or block diagram.

[0118] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0119] This disclosure also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement... Figure 1 and Figure 2 The method involved in any of the embodiments.

[0120] This disclosure also provides a computer program product, including computer instructions, wherein the computer instructions, when executed by a processor, implement as follows: Figure 1 and Figure 2 The method involved in any of the embodiments.

[0121] Figure 4 This is a schematic diagram of the structure of an intelligent driving vehicle according to an embodiment of this disclosure. Figure 4 As shown, the intelligent driving vehicle includes a data cleaning device 41, a lidar 42, and a vehicle control device 43. The data cleaning device 41 is... Figure 3 The data cleaning apparatus involved in any of the embodiments.

[0122] The lidar 42 is configured to collect point cloud data of targets around the intelligent driving vehicle and send the point cloud data to the data cleaning device 41.

[0123] The vehicle control unit 43 is configured to control the intelligent driving vehicle based on the cleaned point cloud data sent by the data cleaning unit 41.

[0124] The following specific example illustrates this disclosure.

[0125] Figure 5 This is a speed prediction graph for data that has not been cleaned. (Example:) Figure 5 As shown, the velocity prediction graph clearly includes a large number of outliers.

[0126] Figure 6 A speed prediction chart for data cleaning using the scheme provided in this disclosure. Compared to Figure 5 , Figure 6 Effective cleaning of outlier values ​​in the velocity prediction plot shown.

[0127] In some embodiments, the functional modules described above may be implemented as general-purpose processors, programmable logic controllers (PLCs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or any suitable combination thereof for performing the functions described herein.

[0128] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0129] The description in this disclosure is provided for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit the disclosure to its forms. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of this disclosure and to enable those skilled in the art to understand this disclosure and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A data cleaning method, performed by a data cleaning device in an intelligent driving vehicle, comprising: Calculate the number of adjacent points for each target point in the point cloud data; Based on the type of each target point and the distance between each target point and the intelligent driving vehicle, determine the threshold number of adjacent points for each target point; If the number of adjacent points of the i-th target point is less than the threshold number of adjacent points of the i-th target point, then the i-th target point is deleted from the point cloud data. N is the total number of target points in the point cloud data, so as to obtain cleaned point cloud data; The number of neighboring points of each target point in the calculated point cloud data includes: Based on the position, velocity, and type of the i-th target point, and the position, velocity, and type of the j-th target point, determine whether the i-th target point and the j-th target point belong to the same target. ,and ; If the i-th target point and the j-th target point belong to the same target, increment the number of adjacent points of the i-th target point by 1, and increment the number of adjacent points of the j-th target point by 1; The step of determining whether the i-th target point and the j-th target point belong to the same target includes: If the types of the i-th target point and the j-th target point are the same, the position error between the i-th target point and the j-th target point is less than the position error threshold, the velocity error or relative velocity error rate between the i-th target point and the j-th target point is less than a predetermined threshold, the velocity direction cosine error between the i-th target point and the j-th target point is less than the cosine error threshold, and the product of the relative velocity error rate and the velocity direction cosine error is less than a critical threshold, then the i-th target point and the j-th target point are determined to belong to the same target.

2. The data cleaning method according to claim 1, wherein, The determination of whether the i-th target point and the j-th target point belong to the same target includes: If the types of the i-th target point and the j-th target point are the same, calculate the positional error between the i-th target point and the j-th target point; If the position error is less than the position error threshold, calculate the velocity error and the relative velocity error rate between the i-th target point and the j-th target point; If the speed error is less than the speed error threshold, or the speed relative error rate is less than the speed relative error rate threshold, calculate the speed direction cosine error between the i-th target point and the j-th target point; If the velocity direction cosine error is less than the cosine error threshold, calculate the product of the velocity relative error rate and the velocity direction cosine error; If the product of the relative velocity error rate and the velocity direction cosine error is less than a critical threshold, the i-th target point and the j-th target point are determined to belong to the same target.

3. The data cleaning method according to claim 2, wherein, The position error threshold is associated with the type of the j-th target point.

4. The data cleaning method according to claim 2, wherein, The calculation of the relative velocity error rate between the i-th target point and the j-th target point includes: The relative speed error rate is determined based on the maximum value between the speed of the i-th target point and the speed of the j-th target point, and the speed error between the i-th target point and the j-th target point.

5. The data cleaning method according to claim 4, wherein, The relative speed error rate is positively correlated with the speed error and negatively correlated with the maximum value.

6. The data cleaning method according to claim 5, wherein, Determining the relative speed error rate includes: The relative speed error rate is determined based on the ratio of the speed error to the maximum value.

7. The data cleaning method according to claim 2, wherein, The calculation of the velocity direction cosine error of the i-th target point and the j-th target point includes: The velocity direction cosine error is determined based on the velocity of the i-th target point, the component of the velocity of the i-th target point in the first direction, the component of the velocity of the i-th target point in the second direction, the velocity of the j-th target point, the component of the velocity of the j-th target point in the first direction, and the component of the velocity of the j-th target point in the second direction, wherein the first direction is perpendicular to the second direction.

8. The data cleaning method according to claim 7, wherein, The determination of the velocity direction cosine error includes: Multiply the velocity component of the i-th target point in the first direction by the velocity component of the j-th target point in the first direction to obtain the first intermediate quantity; Multiply the component of the velocity of the i-th target point in the second direction with the component of the velocity of the j-th target point in the second direction to obtain the second intermediate quantity; Add the first intermediate value and the second intermediate value to obtain the third intermediate value; Multiply the velocity of the i-th target point and the velocity of the j-th target point to obtain the fourth intermediate quantity; Divide the third intermediate quantity by the fourth intermediate quantity to obtain the fifth intermediate quantity; The velocity direction cosine error is determined based on the difference between the predetermined parameter and the fifth intermediate quantity.

9. The data cleaning method according to claim 1, wherein, The threshold number of adjacent points for each target point is positively correlated with the distance between each target point and the intelligent driving vehicle.

10. The data cleaning method according to claim 9, wherein, The threshold for determining the number of adjacent points for each target point includes: Obtain a first category threshold and a second category threshold corresponding to the type of each target point, wherein the first category threshold is greater than the second category threshold; Subtract the first category threshold from the second category threshold to obtain the sixth intermediate value; The seventh intermediate quantity is obtained by multiplying the sixth intermediate quantity and the distance between each target point and the intelligent driving vehicle. The first category threshold is added to the seventh intermediate value to obtain the number of neighboring points threshold for each target point.

11. The data cleaning method according to any one of claims 1-10, further comprising: The cleaned point cloud data is sent to the vehicle control device in the intelligent driving vehicle.

12. A data cleaning apparatus, comprising: Memory; A processor, coupled to a memory, is configured to implement the data cleaning method as described in any one of claims 1-11 based on the memory-stored instructions.

13. An intelligent driving vehicle, comprising: The data cleaning apparatus as described in claim 12; A lidar is configured to collect point cloud data of targets around the intelligent driving vehicle and send the point cloud data to the data cleaning device. The vehicle control device is configured to control the intelligent driving vehicle based on the cleaned point cloud data sent by the data cleaning device.

14. A computer-readable storage medium, wherein, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the data cleaning method as described in any one of claims 1-11.

15. A computer program product comprising computer instructions, wherein the computer instructions, when executed by a processor, implement the data cleaning method as described in any one of claims 1-11.

Images