Method and system for checking plausibility of echo data of lidar, vehicle and computer program product
Through vehicle-mounted sensors, the target object position size information is sensed, the target object is simulated, and the predicted echo data is calculated, which solves the problem that the reliability of lidar echo data is difficult to ensure, and the efficient credibility evaluation of lidar echo data is achieved , improves the robustness of the detection results.
Patent Information
- Application Number
- CN202411948228.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-02-18
AI Technical Summary
Due to the angular resolution limitation in lidar detection, the credibility of echo data is difficult to ensure.
The position size information of the target is sensed by the vehicle-mounted camera and/or millimeter wave radar, and the simulated target is constructed and the predicted echo data is calculated based on this, and then compared with the actual echo data to conduct a confidence test.
Use the sensing data of other on-board sensors to efficiently and accurately evaluate the credibility of lidar echo data, improve the robustness of detection results, and provide a foundation for the reliable implementation of autonomous driving and assisted driving functions.
Smart Images
Figure CN119471648A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of vehicle-mounted sensors, and in particular to a method for verifying the credibility of echo data of a laser radar, a system for verifying the credibility of echo data of a laser radar, a vehicle comprising the system, and a computer program product for at least assisting in implementing the steps of the method according to the present application. Background Art
[0002] With the development of vehicle technology, more and more vehicles are equipped with laser radar, which detects the position, size, speed and other information of the target by emitting a laser beam to the target and receiving the echo data reflected by the target. However, the detection ability of the laser radar for the target is limited by the angular resolution of the laser radar. The larger the angular resolution of the laser radar, the fewer the echo points distributed within a unit spatial angle, and the lower the detection ability of the laser radar, which makes it difficult to guarantee the credibility of the laser radar's echo data about the target.
[0003] Therefore, how to verify the credibility of the echo data received by the lidar has become a technical problem that needs to be solved. Summary of the invention
[0004] The purpose of the present application is to provide a method for credibility verification of laser radar echo data, a system for credibility verification of laser radar echo data, a vehicle comprising the system, and a computer program product, so as to at least partially solve the problems in the prior art.
[0005] According to a first aspect of the present application, a method for credibility verification of laser radar echo data is provided, the method comprising the following steps:
[0006] - sensing position and size information about a target object at a predetermined distance in front of the vehicle by means of a first on-board sensor, wherein the first on-board sensor comprises an on-board camera and / or a millimeter-wave radar;
[0007] - obtaining predicted echo data of the target by the laser radar based on at least the position and size information of the sensed target and the configuration parameters of the laser radar; and
[0008] - Performing a credibility check on actual echo data about the target object acquired by the laser radar based on the predicted echo data.
[0009] The core concept of this application is to sense the position and size information of the target object through the vehicle-mounted camera and / or millimeter-wave radar, construct a simulated target object based on the sensed position and size information, and obtain the predicted echo data about the simulated target object. By comparing the predicted echo data about the target object with the actual echo data, the credibility of the laser radar echo data can be verified. In this way, the sensing data of other vehicle-mounted sensors can be used to efficiently and accurately evaluate the credibility of the laser radar echo data, improve the robustness of the laser radar detection results, and lay the foundation for the reliable implementation of the vehicle's automatic driving function and / or assisted driving function.
[0010] According to an optional embodiment of the present application, the location and size information of the target object may include a width b of the target object, a height h of the target object, and a distance d of the target object relative to the vehicle.
[0011] According to another optional embodiment of the present application, when the first vehicle-mounted sensor is configured as a binocular vehicle-mounted camera, the position and size information of the target object can be determined with the help of the stereoscopic imaging effect of two camera lenses.
[0012] According to another optional embodiment of the present application, when the first vehicle-mounted sensor is configured as a monocular vehicle-mounted camera, the position and size information of the target object can be determined, for example, with the help of a motion recovery structure algorithm.
[0013] According to another optional embodiment of the present application, when the first vehicle-mounted sensor is configured as a millimeter wave radar, the distance d of the target object relative to the vehicle can be determined by means of the Doppler effect.
[0014] According to another optional embodiment of the present application, the configuration parameters of the laser radar may include a lateral angular resolution α h and the longitudinal angular resolution α v .
[0015] According to another optional embodiment of the present application, the target object may be at least based on its height h and longitudinal angular resolution α. v Calculate the number of rows p of the laser radar's predicted echo data about the target r , whose formula is, for example:
[0016] p r =rounddown(h / (d*tanα v )+x v ),
[0017] Among them, rounddown() represents the function of rounding down, x v Indicates the longitudinal deviation reserved value.
[0018] According to another optional embodiment of the present application, the width b and the lateral angular resolution α of the target object may be at least h Calculate the number of columns q of the laser radar's predicted echo data about the target r , whose formula is, for example:
[0019] q r =rounddown(b / (d*tanα h )+x h ),
[0020] Among them, rounddown() represents the function of rounding down, x h Indicates the lateral deviation reserve value.
[0021] According to another optional embodiment of the present application, the number of rows p of the predicted echo data may be calculated as r Relative to the number of rows p of the actual echo data m The first deviation, and the number of columns q of the predicted echo data r Relative to the column number q of the actual echo data m The second deviation, and the number of echo points n of the predicted echo data r A third deviation of the number of echo points of the actual echo data, wherein the number of echo points of the predicted echo data is n r Equal to the number of rows p of the predicted echo data r The number of columns of the predicted echo data q r Next, the average deviation value δ of the first deviation, the second deviation and the third deviation, and the average deviation value δ relative to the number of echo points n of the predicted echo data can be calculated. r The relative average deviation δ r .
[0022] According to another optional embodiment of the present application, the calculated relative average deviation value δ r The reliability of the actual echo data is tested by comparing it with the preset deviation threshold, wherein the calculated relative average deviation value δ r When the deviation is greater than a preset threshold, the actual echo data has an unreliable confidence level.
[0023] According to a second aspect of the present application, a system for credibility check of laser radar echo data is provided, wherein the system may include the following components:
[0024] - a first vehicle-mounted sensor, which is configured to sense position and size information about a target object at a predetermined distance in front of the vehicle, wherein the first vehicle-mounted sensor includes a vehicle-mounted camera and / or a millimeter-wave radar;
[0025] - a laser radar configured to acquire actual echo data about the target object; and
[0026] - A control unit configured to perform the method according to the application.
[0027] According to a third aspect of the present application, a vehicle is provided, comprising a system according to the present application.
[0028] According to a fourth aspect of the present application, a computer program product, such as a computer-readable program carrier, is provided, which contains or stores computer program instructions, and when the computer program instructions are executed by a processor, they at least assist in implementing the steps of the method described in the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] The principles, features and advantages of the present application may be better understood by describing the present application in more detail below with reference to the accompanying drawings. The accompanying drawings show:
[0030] Figure 1 A schematic diagram showing a scene of a laser radar detecting a target object according to an exemplary embodiment of the present application;
[0031] Figure 2 A schematic diagram showing a scene of a laser radar detecting a target object according to another exemplary embodiment of the present application;
[0032] Figure 3 A flowchart showing a method for credibility verification of laser radar echo data according to an exemplary embodiment of the present application is shown;
[0033] Figures 4 to 7 an exemplary schematic diagram showing actual echo data regarding a target object disposed at different distances; and
[0034] Figure 8 A schematic block diagram of a system for credibility verification of laser radar echo data according to an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION
[0035] In order to make the technical problems, technical solutions and beneficial technical effects to be solved by the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and multiple exemplary embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the scope of protection of the present application.
[0036] The width of the target that the laser radar can effectively detect is limited by the angular resolution of the laser radar, which represents the angular interval between two adjacent laser scanning points, including the lateral angular resolution α in the horizontal direction. hand the longitudinal angular resolution α in the vertical direction v .like Figure 1 The schematic diagram of an exemplary scene of a laser radar detecting a target object is shown in FIG. 1 , and the longitudinal angular resolution α of the laser radar 12 of the vehicle 1 is v For example, when it detects a target object 2 200m in front of the vehicle 1, the distance between the two longitudinal laser beams is 200m*tan0.4°≈1.4m, which means that at 200m in front of the vehicle, the laser radar 12 can only effectively detect a target object 2 higher than 1.4m.
[0037] Figure 2 Another exemplary scene schematic diagram of laser radar detecting a target object is shown. For the sake of clarity, Figure 2 The first onboard sensor 11 of the vehicle 1 is no longer shown in the top view of the vehicle 1. The lateral angular resolution α of the laser radar 12 of the vehicle 1 is h For example, when the laser radar 12 detects a target 2 200m in front of the vehicle 1, the distance between the two transverse laser beams is 200m*tan0.2°≈0.7m, which means that at 200m in front of the vehicle, the laser radar 12 can only effectively detect a target 2 with a width greater than 1.4m. When the laser radar 12 detects the target 2, a certain amount of echo data is required. The farther the target is from the vehicle and / or the smaller the size of the target, the less echo data the laser radar receives, and the higher the possibility of receiving interference echo data. The echo data of the laser radar 12 can be tested for credibility by the method according to the present application.
[0038] Figure 3 A flowchart of a method for credibility check of laser radar echo data according to an exemplary embodiment of the present application is shown. The following exemplary embodiments describe the method according to the present application in more detail.
[0039] like Figure 3 As shown, the method may include steps S1 to S3. In step S1, the position and size information of the target object 2 at a predetermined distance in front of the vehicle 1 is sensed by means of the first onboard sensor 11. In the current embodiment of the present application, the vehicle 1 is equipped with a first onboard sensor 11 different from the laser radar 12, which includes not only an onboard camera, such as a binocular onboard camera and / or a monocular onboard camera, but also a millimeter wave radar.
[0040] In the case where the first vehicle-mounted sensor 11 is configured as a binocular vehicle-mounted camera, the position and size information of the target object 2 can be determined by means of the stereoscopic imaging effect of the two camera lenses. The two camera lenses will form two two-dimensional images when shooting the same target object 2. Due to the baseline distance between the two cameras, the position of the same target object 2 in the two two-dimensional images is offset, and this offset is the parallax. The position and size information of the target object 2 can be determined by calculating the offset of the corresponding pixel points of the target object 2 in the two two-dimensional images, which may include the width b of the target object 2, the height h of the target object 2, and the distance d of the target object 2 relative to the vehicle 1, etc.
[0041] In the case where the first vehicle-mounted sensor 11 is configured as a monocular vehicle-mounted camera, the position and size information of the target object 2 can be determined, for example, by means of a structure-from-motion algorithm. During the movement of the vehicle 1 relative to the target object 2, a plurality of two-dimensional images of the target object 2 can be captured by a monocular camera, and the position and size information of the target object 2 can be determined by feature matching and geometric relationships between these two-dimensional images, which can also include the width b of the target object 2, the height h of the target object 2, and the distance d of the target object 2 relative to the vehicle 1, etc.
[0042] In the case where the first vehicle-mounted sensor 11 is configured as a millimeter-wave radar, during the movement of the vehicle 1 relative to the target 2, the distance d of the target 2 relative to the vehicle 1 can be determined by means of the Doppler effect. However, in the case where the target 2 is a stationary object, such as a lying pedestrian or motorcyclist, the accuracy of the distance d determined by the millimeter-wave radar is low. In addition, the detection accuracy of the millimeter-wave radar for the size information perpendicular to the movement direction of the target 2 is also low. For example, when the target 2 moves toward or away from the vehicle 1, the detection accuracy of the millimeter-wave radar for the width b and height h of the target 2 is much lower than that of the vehicle-mounted camera.
[0043] In step S2, the predicted echo data of the laser radar 12 about the target 2 can be obtained based at least on the position and size information of the sensed target 2 and the configuration parameters of the laser radar 12. In the meaning of the present application, "predicted echo data" can be understood as the laser radar 12 with the current parameter configuration predicting the echo data that may be received when detecting a simulated target constructed based on the sensed position and size information, and the echo data is composed of a certain number of rows and columns of radar echo points, wherein the number of rows and columns of the radar echo points depends on the position and size information of the target. The configuration parameters of the laser radar 12 may include the lateral angular resolution α in the horizontal direction. h and the longitudinal angular resolution α in the vertical direction v .
[0044] Here, the height h and the longitudinal angular resolution α of the target object 2 may be used to determine the target object 2. v Calculate the number of rows p of the predicted echo data of the laser radar 12 about the target 2 r , whose formula is, for example:
[0045] p r =rounddown(h / (d*tanα v )+x v ),
[0046] Among them, rounddown() represents the function of rounding down; x v represents the longitudinal deviation reserve value, which is equal to 0.1, for example. By setting x v The number of rows of predicted echo data obtained can be increased r Robustness; d*tanα v Indicates the predicted distance between two rows of radar echo points.
[0047] Similarly, the width b and the lateral angular resolution α of the target object 2 may be used to determine the h Calculate the number of columns q of the predicted echo data of the laser radar 12 about the target 2 r , whose formula is, for example:
[0048] q r =rounddown(b / (d*tanα h )+x h ),
[0049] Among them, rounddown() represents the function of rounding down; x h represents the lateral deviation reserve value, which is equal to 0.1, for example. By setting x h The number of columns of predicted echo data obtained can be increased r Robustness; d*tanα h Represents the predicted distance between two columns of radar echo points.
[0050] For simplicity, it is assumed that the target object 2 has a rectangular outline, and the number of echo points n of the predicted echo data is r It can be calculated as the number of rows p of the predicted echo data r The number of columns of the predicted echo data q r The product of .
[0051] In step S3, the actual echo data about the target object 2 acquired by the laser radar 12 is subjected to a credibility check based on the predicted echo data. Here, the number of rows p of the predicted echo data can be calculated. r Relative to the number of rows p of the actual echo data mThe first deviation, and the number of columns q of the predicted echo data r Relative to the column number q of the actual echo data m The second deviation, and the number of echo points n of the predicted echo data r The number of echo points n relative to the actual echo data m The third deviation.
[0052] Next, the average deviation value δ of the first deviation, the second deviation and the third deviation is calculated by the following formula:
[0053]
[0054] The average deviation value δ is further calculated with respect to the number of echo points n of the predicted echo data by the following formula: r The relative average deviation δ r :
[0055]
[0056] Here, the relative mean deviation value δ r It can be used to evaluate the concentration of data distribution, that is, the closeness between the predicted echo data and the actual echo data, where the relative average deviation value δ r The smaller the value, the closer the predicted echo data is to the actual echo data, and the higher the credibility of the actual echo data. r The comparison with the preset deviation threshold can verify the credibility of the actual echo data, wherein the calculated relative average deviation value δ r When the deviation is greater than a preset threshold, the actual echo data has an unreliable confidence level.
[0057] In the test scenario according to the present application, a target object 2, such as a stationary dummy model, may be placed at different positions away from the vehicle 1, and the target object 2 may be detected by the laser radar 12 to obtain actual echo data about the target object 2. In order to improve the accuracy of the on-board camera sensing the target object 2, the ambient light intensity may be adjusted based on the distance of the target object 2 from the vehicle 1.
[0058] Figure 4 FIG1 shows an exemplary schematic diagram of actual echo data received by the laser radar 12 for a target object 2 located 100 m away from the vehicle. The actual echo data includes a total of 21 echo points (marked with hollow circles) and arranged in 5 rows and 5 columns. As shown in Table 1, the number of rows p of the predicted echo data for the target object 2 is r Equal to 4, the number of columns is q r =5, and the number of echo points n of the predicted echo data is calculatedr = 20. Furthermore, the average deviation δ of the actual echo data relative to the predicted echo data is calculated to be 0.67, and the relative average deviation δ r Equal to 0.03.
[0059] d[m] 100 160 180 200 <![CDATA[p m ]]> 5 3 3 2 <![CDATA[p r ]]> 4 3 2 2 <![CDATA[q m ]]> 5 3 3 1 <![CDATA[q r ]]> 5 3 3 3 <![CDATA[n m ]]> 21 7 5 2 <![CDATA[n r ]]> 20 9 6 6 δ 0.67 0.67 0.67 2.00 <![CDATA[δ r ]]> 0.03 0.07 0.11 0.33
[0060] Table 1: Statistical results of actual echo data and predicted echo data in different test scenarios
[0061] Figure 5 FIG. 1 shows an exemplary schematic diagram of actual echo data received by the laser radar 12 for a target object 2 located 160 m away from the vehicle. The actual echo data includes a total of 7 echo points arranged in 3 rows and 3 columns. As shown in Table 1, the number of rows p of the predicted echo data for the target object 2 is r Equal to 3, the number of columns is q r =3, thus the number of echo points n of the predicted echo data is calculated r = 9. Furthermore, the average deviation value δ of the actual echo data relative to the predicted echo data is calculated to be 0.67, and the relative average deviation value δ r Equal to 0.07.
[0062] Figure 6 The figure shows an exemplary schematic diagram of the actual echo data received by the laser radar 12 for the target object 2 set at 180m away from the vehicle. The actual echo data includes a total of 5 echo points, which are arranged in 3 rows and 3 columns. As shown in Table 1, the number of rows p of the predicted echo data for the target object 2 is r Equal to 2, the number of columns is q r =3, thus the number of echo points n of the predicted echo data is calculated r = 6. Furthermore, the average deviation value δ of the actual echo data relative to the predicted echo data is calculated to be 0.67, and the relative average deviation value δ r Equal to 0.11.
[0063] Figure 7 The figure shows an exemplary schematic diagram of the actual echo data received by the laser radar 12 for the target object 2 set at 200m away from the vehicle. The actual echo data includes a total of 2 echo points, arranged in 2 rows and 1 column. As shown in Table 1, the number of rows p of the predicted echo data for the target object 2 is r Equal to 2, the number of columns is q r =3, thus the number of echo points n of the predicted echo data is calculated r = 6. Further, the average deviation value δ is calculated to be 2, and the relative average deviation value δ r Equal to 0.33.
[0064] Depend on Figures 4 to 7 The test results show that as the distance between the target object 2 and the vehicle 1 increases, the relative average deviation value δ between the predicted echo data and the actual echo data increases. r As the value increases gradually, the credibility of the actual echo data gradually decreases until the calculated relative average deviation value δ r If the deviation is greater than a preset deviation threshold, the actual echo data has an untrustworthy confidence level.
[0065] According to the present application, the position and size information of the target object is sensed by the vehicle-mounted camera and / or millimeter-wave radar, a simulated target object is constructed based on the sensed position and size information, and the predicted echo data about the simulated target object is obtained. The credibility check of the laser radar echo data can be completed by comparing the predicted echo data about the target object with the actual echo data. In this way, the sensing data of other vehicle-mounted sensors can be used to efficiently and accurately evaluate the credibility of the laser radar echo data, improve the robustness of the laser radar detection results, and lay the foundation for the reliable implementation of the vehicle's automatic driving function and / or assisted driving function.
[0066] In addition, it should be noted that the step numbers described herein do not necessarily represent a chronological order, but are merely a figure mark. The order can be changed according to specific circumstances as long as the technical purpose of the present application can be achieved.
[0067] Figure 8 A schematic block diagram of a system 10 for credibility verification of laser radar echo data according to an exemplary embodiment of the present application is shown.
[0068] like Figure 8 As shown, the system 10 may include the following components:
[0069] - a first vehicle-mounted sensor 11, which is configured to sense position and size information about a target object 2 at a predetermined distance in front of the vehicle 1, wherein the first vehicle-mounted sensor 11 includes, for example, a vehicle-mounted camera and / or a millimeter-wave radar, etc.;
[0070] - a laser radar 12, which is configured to obtain actual echo data about the target object 2; and
[0071] - A control unit 13 configured to carry out the method according to the present application.
[0072] It should be understood that, in this document, the expressions "first", "second", "third", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance, nor should they be understood as implicitly indicating the number of the indicated technical features.
[0073] If an embodiment includes an “and / or” relationship between a first feature and a second feature, it should be interpreted as follows: according to one embodiment, the embodiment has both the first feature and the second feature, and according to another embodiment, the embodiment has either only the first feature or only the second feature.
[0074] Although specific embodiments have been described above, these embodiments are not intended to limit the scope of the present disclosure, even when only a single embodiment is described with respect to specific features. The feature examples provided in the present disclosure are intended to be illustrative, not restrictive, unless otherwise stated. In specific implementations, multiple features may be combined with each other, where technically feasible, according to actual needs. Various substitutions, changes, and modifications may also be conceived without departing from the spirit and scope of the present application.
Claims
1. A method for verifying the credibility of echo data of a laser radar (12), the method comprising the following steps: Using a first on-board sensor (11), sensing position and size information about a target object (2) at a predetermined distance in front of the vehicle (1), wherein the first on-board sensor (11) comprises an on-board camera and / or a millimeter-wave radar; Obtaining predicted echo data of the laser radar (12) about the target (2) based at least on the position and size information of the sensed target (2) and the configuration parameters of the laser radar (12); and Based on the predicted echo data, the actual echo data about the target object (2) acquired by the laser radar (12) is subjected to a credibility check.
2. The method according to claim 1, wherein: The position and size information of the target object (2) includes a width b of the target object (2), a height h of the target object (2), and a distance d of the target object (2) relative to the vehicle (1).
3. A method according to any one of the preceding claims, wherein: When the first vehicle-mounted sensor (11) is configured as a binocular vehicle-mounted camera, the position and size information of the target object (2) is determined by means of a stereo imaging effect of two camera lenses; and / or When the first vehicle-mounted sensor (11) is configured as a monocular vehicle-mounted camera, determining the position and size information of the target object (2) by means of a structure-from-motion algorithm; and / or When the first onboard sensor (11) is configured as a millimeter wave radar, the distance d of the target object (2) relative to the vehicle (1) is determined by means of the Doppler effect.
4. A method according to any one of the preceding claims, wherein: The configuration parameters of the laser radar (12) include the lateral angular resolution α h and the longitudinal angular resolution α v .
5. A method according to any one of the preceding claims, wherein: At least based on the height h and the longitudinal angular resolution α of the target (2) v Calculate the number of rows p of the predicted echo data of the laser radar (12) about the target (2) r , whose formula is, for example: p r =rounddown(h / (d*tanα v )+x v ), Among them, rounddown() represents the function of rounding down, x v Indicates the longitudinal deviation reserved value.
6. A method according to any one of the preceding claims, wherein: At least based on the width b and the lateral angular resolution α of the target (2) h Calculate the number of columns q of the predicted echo data of the laser radar (12) about the target (2) r , whose formula is, for example: q r =rounddown(b / (d*tanα h )+x h ), Among them, rounddown() represents the function of rounding down, x h Indicates the lateral deviation reserve value.
7. A method according to any one of the preceding claims, wherein: Calculate the number of rows p of the predicted echo data r Relative to the number of rows p of the actual echo data m The first deviation, and the number of columns q of the predicted echo data r Relative to the column number q of the actual echo data m The second deviation, and the number of echo points n of the predicted echo data r A third deviation of the number of echo points of the actual echo data, wherein the number of echo points of the predicted echo data is n r Equal to the number of rows p of the predicted echo data r The number of columns of the predicted echo data q r Calculate the average deviation value δ of the first deviation, the second deviation and the third deviation, and the average deviation value δ relative to the number of echo points n of the predicted echo data r The relative average deviation δ r ; and / or Based on the calculated relative average deviation value δ r The reliability of the actual echo data is tested by comparing it with the preset deviation threshold, wherein the calculated relative average deviation value δ r When the deviation is greater than a preset threshold, the actual echo data has an unreliable confidence level.
8. A system (10) for checking the credibility of echo data of a laser radar (12), wherein: The system (10) comprises the following components: A first vehicle-mounted sensor (11) configured to sense position and size information about a target object (2) at a predetermined distance in front of the vehicle (1), wherein the first vehicle-mounted sensor (11) comprises a vehicle-mounted camera and / or a millimeter-wave radar; A laser radar (12) configured to obtain actual echo data about the target object (2); and A control unit (13) configured to carry out the method according to any one of the preceding claims.
9. A vehicle (1) comprising a system (10) according to claim 8. 10 . A computer program product, such as a computer-readable program carrier, comprising or storing computer program instructions, which, when executed by a processor, at least assist in implementing the steps of the method according to claim 1 .