Method and System for Removing Noise Points from Raw Point Cloud of Vehicle-mounted Millimeter Wave Radar

Abstract

The present invention relates to the field of radar technology, and provides a method and system for excluding noise points from the original point cloud of a vehicle-mounted millimeter-wave radar. The method includes: obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar; calculating a stationary radar cross-section threshold and a moving radar cross-section threshold based on the stationary point data and the moving point data in the original point cloud data; excluding the stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and the moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain the point cloud data after noise point exclusion. On the one hand, the present invention directly performs noise point exclusion operations on the point cloud data, providing a more accurate data basis for subsequent different perception tasks. On the other hand, by separately calculating and judging the stationary and moving point data, and then excluding the point data smaller than the threshold, it can avoid the problem of insufficient accuracy of the radar cross-section (RCS) caused by radar performance, and achieve more refined noise exclusion.

Description

Technical Field

[0001] The present invention relates to the field of radar technology, and in particular, to a method and system for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar. Background Art

[0002] Due to its frequency band, waveform, and the usage characteristics of mobile platforms and high-speed scenarios, the point cloud distribution of vehicle-mounted millimeter-wave radars has deficiencies such as sparse distribution, many false detections, strong interference, range-velocity ambiguity, etc., resulting in problems of target false detection and missed detection in autonomous driving and assisted driving scenarios.

[0003] To solve this problem, existing technologies usually optimize for false detection and missed detection in the signal processing stage. The universality of this optimization is poor, and specific optimization methods are often only effective for a single perception task. As a result, multiple different parallel tasks that utilize the original point cloud data often need to independently process the point cloud data separately.

[0004] Currently, there is no systematic and effective method for preprocessing the radar original points in the point cloud stage. Therefore, how to provide a method for eliminating noise points in radar original point cloud data has become a technical problem that urgently needs to be solved in the industry. Summary of the Invention

[0005] The present invention provides a method and system for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar to solve the defect of target false detection in the prior art and achieve noise point elimination in the radar original point cloud processing stage.

[0006] The present invention provides a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar, including:

[0007] Obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar;

[0008] Calculating a stationary radar cross-section threshold and a moving radar cross-section threshold based on the stationary point data and moving point data in the original point cloud data;

[0009] Excluding stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain point cloud data after noise point elimination.

[0010] According to the method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, the stationary radar cross-section threshold T1 is calculated by the first formula:

[0011] T1 = k c +e ka+R1 +k m +rcs angle offset+rcs aeb offset

[0012] where k c , k a , k m are all constants related to the millimeter-wave radar; R1 is the distance of the stationary point data relative to the radar; e is the base of the natural logarithm; rcs aeb offset is the automatic emergency braking AEB compensation constant for the forward stationary point data of the vehicle; rcs angle offset is the angular offset of the radar cross section.

[0013] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, the moving radar cross section threshold T2 is calculated by the second formula:

[0014] T2 = k c + e ka+R2 + k m + rcs angle offset

[0015] where k c , k a , k m are all constants related to the millimeter-wave radar; R2 is the distance of the moving point data relative to the radar; e is the base of the natural logarithm; rcs angle offset is the angular offset of the radar cross section.

[0016] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, for a vehicle-mounted forward radar, the angular offset rcs angle offset of the radar cross section satisfies the third formula:

[0017] rcs angle offset = max(k1 × |azimuth|, k2)

[0018] where azimuth is the azimuth angle of the point data; k1 and k2 are both empirical constants; max is a function for taking the maximum value;

[0019] For a vehicle-mounted angular radar, the angular offset rcs angle offset of the radar cross section satisfies the fourth formula:

[0020]

[0021] Wherein, azimuth is the azimuth angle of the point data; k3, k4, and k5 are all empirical constants; R is the distance of the point data relative to the vehicle-mounted angular radar; e is the base of the natural logarithm; max is a function for taking the maximum value.

[0022] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, the stationary point data and the moving point data are determined according to the absolute speed of the point data;

[0023] The absolute speed is the speed relative to the ground obtained by converting the speed of the point data relative to the vehicle body in the vehicle body coordinate system.

[0024] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, it further includes:

[0025] Eliminate the multiple reflection point data to obtain the point cloud data after multiple reflection elimination;

[0026] The multiple reflection point data refers to the point data whose distances to the radar are in a multiple relationship, speeds are in a multiple relationship, and the azimuth angle difference is within a set range relative to any point data.

[0027] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, it further includes:

[0028] Determine the field of view angle range according to the scanning state of the radar when the point data is obtained;

[0029] Eliminate the point data whose azimuth angle is not within the field of view angle range to obtain the point cloud data after field of view angle elimination.

[0030] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, it further includes:

[0031] Determine the road edge point data and the drivable area formed by enclosing the road edge according to the distribution of the original point cloud data;

[0032] Eliminate the point data outside the road edge point data and the drivable area to obtain the point cloud data after road edge elimination.

[0033] According to a method for eliminating noise points in the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention, before the step of calculating the stationary radar cross-section threshold and the moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data, any one or any combination of the following steps is further included:

[0034] Determine that the original point cloud data is non-empty data;

[0035] Determine that there is at least one point data in the original point cloud data whose distance from the radar is greater than a set threshold;

[0036] Determine the values of the point data coordinates, speed, and Doppler value with the highest confidence in the original point cloud data.

[0037] The present invention also provides a noise removal system for the original point cloud of a vehicle-mounted millimeter-wave radar, including:

[0038] An acquisition module for acquiring the original point cloud data of the vehicle-mounted millimeter-wave radar;

[0039] A threshold module for calculating a stationary radar cross-section threshold and a moving radar cross-section threshold according to the stationary point data and moving point data in the original point cloud data;

[0040] An exclusion module for excluding stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain point cloud data after noise removal.

[0041] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the noise removal method for the original point cloud of the vehicle-mounted millimeter-wave radar as described in any one of the above are implemented.

[0042] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the noise removal method for the original point cloud of the vehicle-mounted millimeter-wave radar as described in any one of the above are implemented.

[0043] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the noise removal method for the original point cloud of the vehicle-mounted millimeter-wave radar as described in any one of the above are implemented.

[0044] The noise removal method and system for the original point cloud of the vehicle-mounted millimeter-wave radar provided by the present invention directly perform noise removal operations on the point cloud data, providing a more accurate data basis for subsequent different perception tasks;

[0045] At the same time, by separately calculating and judging the radar cross-section thresholds of the stationary point data and the moving point data, and then excluding the point data with a radar cross-section smaller than the threshold, the problem of insufficient accuracy of the radar cross-section RCS caused by radar performance can be avoided, and more refined noise removal can be achieved. Description of the Drawings

[0046] To more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0047] Figure 1 It is a schematic flowchart of a method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention;

[0048] Figure 2 It is a schematic diagram of curb elimination in the method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar provided by an embodiment of the present invention;

[0049] Figure 3 It is a schematic structural diagram of a system for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar provided by the present invention;

[0050] Figure 4 It is a schematic structural diagram of an electronic device provided by the present invention.

[0051] Reference numerals:

[0052] 1: Acquisition module; 2: Threshold module; 3: Exclusion module;

[0053] 410: Processor; 420: Communication interface; 430: Memory;

[0054] 440: Communication bus. Detailed implementation manners

[0055] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0056] The following will describe Figure 1 , Figure 2 the method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar of the present invention.

[0057] As Figure 1 shown, an embodiment of the present invention provides a method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar, including:

[0058] Step 101, obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar;

[0059] Step 103: Calculate the stationary radar cross-section threshold and the moving radar cross-section threshold based on the stationary point data and the moving point data in the original point cloud data;

[0060] Step 105: Exclude the stationary points with a radar cross-section less than the stationary radar cross-section threshold and the moving points with a radar cross-section less than the moving radar cross-section threshold to obtain the point cloud data after noise point exclusion.

[0061] In this embodiment, the stationary point data and the moving point data are determined according to the absolute velocity of the point data;

[0062] The absolute velocity is the velocity relative to the ground obtained by converting the velocity of the point data relative to the vehicle body in the vehicle body coordinate system.

[0063] Specifically, this embodiment provides a method for determining stationary point data and moving point data, including:

[0064] Determine the acquisition radar of the data points in the original point cloud data;

[0065] Determine the relative velocity direction of the point data according to the position of the acquisition radar and the coordinates of the point data;

[0066] Obtain the absolute velocity of the point data according to the velocity of the vehicle body (i.e., this vehicle) and the relative velocity information of the point data (i.e., the velocity of the point data relative to the vehicle body);

[0067] Determine the point data with an absolute velocity of zero as the stationary point data;

[0068] Determine the point data with a non-zero absolute velocity as the moving point data;

[0069] It should be noted that in step 103, the calculation of the stationary radar cross-section threshold and the moving radar cross-section threshold is also based on the vehicle body coordinates of the data points.

[0070] The steps for obtaining the vehicle body coordinates include:

[0071] Determine the polar coordinate system of the acquisition radar and the polar coordinates of the data points according to the acquisition radar and the data points;

[0072] Convert the polar coordinates of the data points into the vehicle body coordinates in the vehicle body coordinate system according to the position and calibration information of the acquisition radar;

[0073] The beneficial effect of this embodiment is that:

[0074] By directly performing noise point exclusion operations on the point cloud data, a more accurate data basis is provided for subsequent different perception tasks;

[0075] Meanwhile, by separately calculating and judging the radar cross-section thresholds for the stationary point data and the moving point data, and then excluding the point data with a radar cross-section smaller than the threshold, it is possible to avoid the problem of insufficient accuracy of the radar cross-section RCS caused by radar performance, and achieve more refined noise exclusion. The radar cross-section RCS represents the physical quantity of the echo intensity generated under the illumination of radar waves. Millimeter-wave radar has a high frequency and a short wavelength. Autonomous driving targets such as cars and trucks are made of metal and have complex external structures, which can form strong reflections on their surfaces. At this wavelength band, they have a relatively large RCS. The sources of noise and clutter include thermal noise within the system, clutter caused by rain, fog, the ground, etc., and interference caused by multipath reflection, etc., and their RCS is often small. Traditional algorithms often use RCS for filtering at the clustering target stage after DBSCAN. This algorithm performs filtering at the point preprocessing stage, which can greatly reduce the possibility of noise points entering the next stage, and has great advantages in improving the accuracy of DBSCAN and reducing the clustering calculation amount.

[0076] According to the above embodiments, in this embodiment:

[0077] The stationary radar cross-section threshold T1 is calculated by the first formula:

[0078]

[0079] In the formula, k c 、k a 、k m are all constants related to the millimeter-wave radar; R1 is the distance of the stationary point data relative to the radar; e is the base of the natural logarithm; rcs aeb offset is the automatic emergency braking AEB compensation constant for the stationary point data in front of the vehicle (where the vehicle itself refers to the vehicle where the millimeter-wave radar is located). In a preferred embodiment, rcs aeb offset =-3; rcs angle offset is the angular offset of the radar cross-section.

[0080] The moving radar cross-section threshold T2 is calculated by the second formula:

[0081]

[0082] In the formula, k c 、k a 、k m are all constants related to the millimeter-wave radar; R2 is the distance of the moving point data relative to the radar; e is the base of the natural logarithm; rcs angle offsetis the angular offset of the radar cross section.

[0083] For the vehicle-mounted forward radar, the angular offset rcs of the radar cross section angle offset satisfies the third formula:

[0084] rcs angle offset = max(k1×|azimuth|, k2)

[0085] where azimuth is the azimuth angle of the point data; k1 and k2 are both empirical constants. In a preferred embodiment, k1 = -4 and k2 = -3.5; max is a function to take the maximum value;

[0086] For the vehicle-mounted angular radar, the angular offset rcs of the radar cross section angle offset satisfies the fourth formula:

[0087]

[0088] where azimuth is the azimuth angle of the point data; k3, k4, and k5 are all empirical constants. In a preferred embodiment, k3 = -4.5, k4 = -0.2, and k5 = -5.5; R is the distance of the point data relative to the vehicle-mounted angular radar; e is the base of the natural logarithm; max is a function to take the maximum value.

[0089] The beneficial effect of this embodiment is that:

[0090] By adding an automatic emergency braking AEB compensation constant to the threshold of the stationary point data, the subsequent tasks can be more sensitive to the recognition of stationary obstacles, thereby reducing the probability of missed detection in the autonomous driving or assisted driving scenarios.

[0091] According to any of the above embodiments, this embodiment further includes:

[0092] Excluding the multiple reflection point data to obtain the point cloud data after multiple reflection exclusion;

[0093] The multiple reflection point data refers to the point data whose distance to the radar, speed, and azimuth difference are in a multiple relationship within a set range relative to any point data.

[0094] It should be noted that due to the existence of the Doppler effect, the data returned by the millimeter-wave radar naturally includes the velocity information of the point, and for the data of the millimeter-wave reflection points with multiple reflections, the distance and velocity are in a multiple relationship and the direction angles are the same. However, in actual use, due to the existence of radar errors, an error threshold can be set so that the distance and velocity do not have to strictly conform to the multiple relationship, and the azimuth angles do not have to be strictly the same, thereby more effectively excluding the multiple reflection points.

[0095] Determine the field of view angle range according to the scanning state of the radar when the point data is obtained;

[0096] Exclude the point data whose azimuth angles are not within the field of view angle range to obtain the point cloud data after field of view angle exclusion.

[0097] Determine the road edge point data and the drivable area formed by the enclosure of the road edge according to the distribution of the original point cloud data;

[0098] Exclude the point data outside the road edge point data and the drivable area to obtain the point cloud data after road edge exclusion.

[0099] Before the step of calculating the stationary radar cross-section threshold and the moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data, any one or any combination of the following steps is also included:

[0100] Determine that the original point cloud data is non-empty data;

[0101] Determine that in the original point cloud data, at least one point data has a distance from the radar greater than the set threshold;

[0102] That is to say, if the distances of all point data from the radar are less than or equal to the set threshold, it means that the point cloud is concentrated in a relatively small range. Such point cloud data is usually misidentified data and therefore does not need to be considered and excluded.

[0103] Determine the values of the point data coordinates, velocity, and Doppler value in the original point cloud data with the highest confidence.

[0104] In the data returned by the radar, for the same point data, there are often at least two sets of data values with different confidences, which is caused by the settings of the radar itself. The method provided in this embodiment can be based on the data values with higher confidence.

[0105] The beneficial effects of this embodiment are as follows:

[0106] Based on the above embodiments, this embodiment further incorporates technical means such as multiple reflection exclusion, field of view exclusion, and curb exclusion, further effectively excluding noise from multiple angles to obtain more accurate and effective point cloud data.

[0107] At the same time, this embodiment adds a judgment step to the original point cloud data, avoiding meaningless operations on invalid data and further improving the efficiency of noise exclusion.

[0108] According to the above embodiments, an embodiment including a complete process will be provided below.

[0109] The purpose of this embodiment is to eliminate various interferences and clutters in the radar point cloud through radar point cloud features, and obtain the reflection points of real targets and obstacles. The specific process is as follows:

[0110] 1. Divide all the original points into two categories: stationary and moving points, and perform subsequent steps separately;

[0111] 2. According to the working state and working mode of the radar, perform a primary filtering on the points outside the effective area;

[0112] 3. For points from different radars, set different rcs thresholds according to the working characteristics for secondary filtering;

[0113] 4. Perform the judgment and filtering of multi-path (i.e., the points detected by the radar after the millimeter wave is reflected by different objects);

[0114] 5. Perform the judgment and filtering of multi-bounce (i.e., the points detected by the radar after the millimeter wave is reflected by the vehicle and the same object multiple times);

[0115] 6. Perform velocity ambiguity resolution of the points.

[0116] Through the above preprocessing steps, clutter and interference can be basically removed, greatly reducing the difficulty of subsequent detection and tracking and improving the stability of the algorithm.

[0117] It should be noted that the execution order of the above preprocessing steps (i.e., the noise exclusion steps) is not fixed. In other words, the order of steps 2-6 can be changed.

[0118] In a preferred implementation, steps 2-6 are executed in sequence. This execution order can execute the steps that are easy to calculate and exclude earlier, thereby excluding the noise that is easy to exclude earlier, and further improving the execution efficiency of the entire method.

[0119] Furthermore, the steps of this embodiment will be described in more detail below.

[0120] There are a large number of clutter, multipath reflections, and "noise" points generated by multiple reflections in the Radar point cloud.

[0121] According to the information of the Radar point cloud attributes x, y, doppler (Doppler value), rcs (radar cross section), etc., through a series of conditional judgments, the points generated by real objects can be identified and the noise points can be filtered out.

[0122] According to the point cloud distribution, calculate the radar barrier as the edge of the road (as Figure 2 shown), and filter out the points that make up the barrier.

[0123] According to the FOV (field of view) of each radar sensor and the autonomous driving perception area we are interested in, filter out the points outside the area of interest.

[0124] For moving points and stationary points, calculate rcs_offset respectively according to their angles and states. If the rcs of a point is lower than the rcs offset, it is considered a suspicious noise point.

[0125] Sort the remaining points according to the distance, from near to far. For the i-th point, traverse all the points after it. If the distances are in a multiple relationship, determine the multi-bounce points according to the n-multi-bounce that needs to be filtered.

[0126] The beneficial effects of this embodiment are as follows:

[0127] By directly performing noise exclusion operations on the point cloud data, it provides a more accurate data basis for subsequent different perception tasks;

[0128] At the same time, by calculating and judging the thresholds of the radar cross section for the stationary point data and the moving point data respectively, and then excluding the point data with a radar cross section smaller than the threshold, it can avoid the problem of insufficient accuracy of the radar cross section RCS caused by radar performance and achieve more refined noise exclusion.

[0129] Next, the noise exclusion device for the original point cloud of the vehicle-mounted millimeter-wave radar provided by the present invention will be described. The noise exclusion device for the original point cloud of the vehicle-mounted millimeter-wave radar described below can be mutually referred to with the noise exclusion method for the original point cloud of the vehicle-mounted millimeter-wave radar described above.

[0130] The embodiment of the present invention also provides a noise exclusion system for the original point cloud of a vehicle-mounted millimeter-wave radar, including:

[0131] An acquisition module 1, configured to acquire the original point cloud data of the vehicle-mounted millimeter-wave radar;

[0132] A threshold module 2, configured to calculate a stationary radar cross-section threshold and a moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data;

[0133] An exclusion module 3, configured to exclude stationary points with a radar cross-section less than the stationary radar cross-section threshold and moving points with a radar cross-section less than the moving radar cross-section threshold, so as to obtain point cloud data after noise point exclusion.

[0134] Furthermore, the system of this embodiment further includes:

[0135] A multiple reflection exclusion module, configured to exclude multiple reflection point data, so as to obtain point cloud data after multiple reflection exclusion;

[0136] The multiple reflection point data refers to point data whose distance to the radar, speed, and azimuth difference are in a multiple relationship and within a set range with respect to any point data.

[0137] A field of view angle determination module, configured to determine a field of view angle range according to the scanning state of the radar when the point data is acquired;

[0138] A field of view angle exclusion module, configured to exclude point data whose azimuth is not within the field of view angle range, so as to obtain point cloud data after field of view angle exclusion.

[0139] A road edge determination module, configured to determine road edge point data and a drivable area formed by enclosing the road edge according to the distribution of the original point cloud data;

[0140] A road edge exclusion module, configured to exclude point data outside the road edge point data and the drivable area, so as to obtain point cloud data after road edge exclusion.

[0141] A non-empty determination module, configured to determine that the original point cloud data is non-empty data;

[0142] A valid determination module, configured to determine that at least one point data in the original point cloud data has a distance greater than a set threshold with respect to the radar;

[0143] A confidence determination module, configured to determine the values of the point data coordinates, speed, and Doppler value with the highest confidence in the original point cloud data.

[0144] The beneficial effects of this embodiment are as follows:

[0145] By directly performing noise point exclusion operations on the point cloud data, a more accurate data basis is provided for subsequent different perception tasks;

[0146] Meanwhile, by separately calculating and judging the radar cross-section thresholds for the stationary point data and the moving point data, and then excluding the point data with a radar cross-section smaller than the threshold, it is possible to avoid the problem of insufficient accuracy of the radar cross-section (RCS) caused by radar performance, and achieve more refined noise exclusion.

[0147] Figure 4 An example of the physical structure diagram of an electronic device is shown as Figure 4 shown. The electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440. Among them, the processor 410, the communication interface 420, and the memory 430 complete mutual communication through the communication bus 440. The processor 410 can call the logical instructions in the memory 430 to execute the method for excluding noise points from the original point cloud of the vehicle-mounted millimeter-wave radar. The method includes: obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar; calculating the stationary radar cross-section threshold and the moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data; excluding the stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and the moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain the point cloud data after noise exclusion.

[0148] In addition, when the logical instructions in the above-mentioned memory 430 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0149] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar provided by each of the above methods. The method includes: obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar; calculating a stationary radar cross-section threshold and a moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data; excluding stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain the point cloud data after noise point removal.

[0150] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the method for removing noise points from the original point cloud of a vehicle-mounted millimeter-wave radar provided by each of the above methods. The method includes: obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar; calculating a stationary radar cross-section threshold and a moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data; excluding stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain the point cloud data after noise point removal.

[0151] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0152] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for excluding noise points from the original point cloud of a vehicle-mounted millimeter-wave radar, characterized in that Including: Obtaining the original point cloud data of the vehicle-mounted millimeter-wave radar; Calculating a stationary radar cross-section threshold and a moving radar cross-section threshold based on the stationary point data and the moving point data in the original point cloud data; Excluding stationary points with a radar cross-section less than the stationary radar cross-section threshold and moving points with a radar cross-section less than the moving radar cross-section threshold to obtain the point cloud data after noise point exclusion, wherein, the stationary radar cross-section threshold T_1 is calculated by the first formula: T_1 = k_c + e^(k_a + R_1) + k_m + rcs_(angle offset) + rcs_(aeb offset) In the formula, k_c, k_a, and k_m are all constants related to the millimeter-wave radar; R_1 is the distance of the stationary point data relative to the radar; e is the base of the natural logarithm; rcs_(aeb offset) is the automatic emergency braking AEB compensation constant of the stationary point data in front of the vehicle; rcs_(angle offset) is the angular offset of the radar cross-section; The moving radar cross-section threshold T_2 is calculated by the second formula: T_2 = k_c + e^(k_a + R_2) + k_m + rcs_(angle offset) In the formula, k_c, k_a, and k_m are all constants related to the millimeter-wave radar; R_2 is the distance of the moving point data relative to the radar; e is the base of the natural logarithm; rcs_(angle offset) is the angular offset of the radar cross-section; and The stationary point data and the moving point data are determined according to the absolute speed of the point data, The absolute speed is the speed relative to the ground obtained by converting the speed of the point data relative to the vehicle body in the vehicle body coordinate system, wherein the point data with an absolute speed of zero is determined as the stationary point data, and the point data with a non-zero absolute speed is determined as the moving point data.

2. The method for excluding noise points from the original point cloud of an in-vehicle millimeter-wave radar according to claim 1, wherein, For the vehicle-mounted corner radar, the angular offset rcs of the radar cross section angleoffset satisfies the third formula: rcs angleoffset = max(k1 × |azimuth|, k2) In the formula, azimuth is the azimuth angle of the point data; k1 and k2 are both empirical constants; max is the function of taking the maximum value; For an in-vehicle forward radar, the angular offset rcs of the radar cross section angleoffset satisfies the fourth formula: In the formula, azimuth is the azimuth angle of the point data; k3, k4, and k5 are all empirical constants; R is the distance of the point data relative to the vehicle-mounted angular radar; e is the base of the natural logarithm; max is the function of taking the maximum value.

3. The method for excluding noise points from the original point cloud of an in-vehicle millimeter-wave radar according to claim 1, wherein, Also including: Excluding the multiple reflection point data to obtain the point cloud data after multiple reflection exclusion; The multiple reflection point data refers to the point data whose distance to the radar is in a multiple relationship, the speed is in a multiple relationship, and the azimuth difference is within a set range relative to any point data.

4. The method for excluding noise points from the original point cloud of an in-vehicle millimeter-wave radar according to claim 1, wherein, Also including: Determining the field of view angle range according to the scanning state of the radar when the point data is obtained; Excluding the point data whose azimuth is not within the field of view angle range to obtain the point cloud data after field of view angle exclusion.

5. The method for excluding noise points from the original point cloud of an in-vehicle millimeter-wave radar according to claim 1, characterized in that, Also including: Determining the road edge point data and the drivable area formed by the enclosure of the road edge according to the distribution of the original point cloud data; Excluding the point data outside the road edge point data and the drivable area to obtain the point cloud data after road edge exclusion.

6. The method for excluding noise points from the original point cloud of an in-vehicle millimeter-wave radar according to claim 1, characterized in that Before the step of calculating the stationary radar cross-section threshold and the moving radar cross-section threshold based on the stationary point data and the moving point data in the original point cloud data, any one or any combination of the following steps is also included: Determine that the original point cloud data is non-empty data; Determine that in the original point cloud data, at least one point data has a distance greater than a set threshold relative to the radar; Determine that in the original point cloud data, the point data coordinates, speed, and Doppler value are the values with the highest confidence; 7. A noise point elimination system for the original point cloud of a vehicle-mounted millimeter-wave radar, characterized in that, Include: An acquisition module for acquiring the original point cloud data of the vehicle-mounted millimeter-wave radar; A threshold module for calculating the stationary radar cross-section threshold and the moving radar cross-section threshold according to the stationary point data and the moving point data in the original point cloud data; An exclusion module for excluding stationary points with a radar cross-section smaller than the stationary radar cross-section threshold and moving points with a radar cross-section smaller than the moving radar cross-section threshold to obtain the point cloud data after noise point exclusion, wherein, the stationary radar cross-section threshold T_1 is calculated by the first formula: T_1 = k_c + e^(k_a + R_1) + k_m + rcs_(angle offset) + rcs_(aeb offset) In the formula, k_c, k_a, and k_m are all constants related to the millimeter-wave radar; R_1 is the distance of the stationary point data relative to the radar; e is the base of the natural logarithm; rcs_(aeb offset) is the automatic emergency braking AEB compensation constant for the stationary point data in front of the vehicle; rcs_(angle offset) is the angular offset of the radar cross-section; The moving radar cross-section threshold T_2 is calculated by the second formula: T_2 = k_c + e^(k_a + R_2) + k_m + rcs_(angle offset) In the formula, k_c, k_a, and k_m are all constants related to the millimeter-wave radar; R_2 is the distance of the moving point data relative to the radar; e is the base of the natural logarithm; rcs_(angle offset) is the angular offset of the radar cross-section; and The stationary point data and the moving point data are determined according to the absolute speed of the point data, The absolute speed is the speed relative to the ground obtained by converting the speed of the point data relative to the vehicle body in the vehicle body coordinate system, wherein the point data with an absolute speed of zero is determined as the stationary point data, and the point data with a non-zero absolute speed is determined as the moving point data.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the steps of the method for excluding noise points from the original point cloud of the vehicle-mounted millimeter-wave radar according to any one of claims 1 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method for excluding noise points from the original point cloud of the vehicle-mounted millimeter-wave radar according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method for excluding noise points from the original point cloud of the vehicle-mounted millimeter-wave radar according to any one of claims 1 to 6 are implemented.

Images