Method and apparatus for detecting relative pose of trailer

Abstract

Embodiments of the present application provide a method and apparatus for detecting the relative pose of a trailer. The method comprises: acquiring first angle values of a trailer relative to a tractor in the current frame determined by means of a plurality of preset modes, and first timestamps corresponding to the current frame under the plurality of first angle values; performing anomaly detection on at least two preset modes on the basis of the plurality of first angle values and the plurality of first timestamps, to obtain a target angle value; and determining relative pose information of the trailer on the basis of the target angle value, dimension information of the trailer, and distance information between the trailer and a towing coupling point and between the tractor and the towing coupling point. The effect of improving the accuracy in detecting the relative pose of the trailer is achieved.

Description

Methods and devices for detecting the relative position and orientation of trailers

[0001] This application claims priority to Chinese Patent Application No. 202411815162.0, filed on December 10, 2024, entitled “Method and Apparatus for Detecting Relative Position of Trailer”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of loading and unloading technology, and in particular to a method and device for detecting the relative position of a trailer. Background Technology

[0003] With the rapid development of logistics and transportation, the demand for high-precision detection of the relative position and attitude of trailers is increasing. For example, in automated cargo loading and unloading operations, it is necessary to accurately know the position and attitude of the trailer relative to the tractor in order to achieve efficient and safe cargo transportation.

[0004] Traditional trailer relative pose detection methods mainly rely on position sensors to estimate the vehicle's position and attitude in order to detect the trailer's pose relative to the tractor.

[0005] However, existing methods for detecting the relative position of trailers suffer from low accuracy. Summary of the Invention

[0006] This application provides a method and apparatus for detecting the relative pose of a trailer, thereby improving the accuracy of trailer relative pose detection.

[0007] In a first aspect, embodiments of this application provide a method for detecting the relative pose of a trailer, comprising:

[0008] The first angle value of the trailer relative to the tractor in the current frame, determined by multiple preset methods, and the first timestamp corresponding to the current frame under the multiple first angle values ​​are obtained. The multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the contact surface of the trailer.

[0009] Based on the plurality of first angle values ​​and the plurality of first timestamps, anomaly detection is performed on the at least two preset methods to obtain the target angle value;

[0010] The relative pose information of the trailer is determined based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

[0011] In one possible implementation, the step of performing anomaly detection on the at least two preset methods based on the plurality of first angle values ​​and the plurality of first timestamps to obtain the target angle value includes:

[0012] Obtain the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp of the previous frame corresponding to the second angle value;

[0013] For any given first angle value, determine a first difference between the first angle value and the second angle value, and a second difference between the first timestamp and the second timestamp;

[0014] If the ratio of the first difference to the second difference is less than a first preset threshold, then the first angle value is determined to be a non-abnormal third angle value. The first preset threshold is determined based on the operating information of the tractor.

[0015] The target angle value is determined based on at least one third angle value.

[0016] In one possible implementation, determining the target angle value based on at least one third angle value includes:

[0017] If the number of the at least one third angle value is an integer greater than 1, then the two fourth angle values ​​with the smallest angle difference are determined from the at least one third angle value.

[0018] If the angle difference between the two fourth angle values ​​is greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the first priority sequence. The priority order in the first priority sequence from largest to smallest is: crossbeam or baffle, lane line, motion information, and first posture information of the tractor.

[0019] If the angle difference between the two fourth angle values ​​is not greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the second priority sequence. The priority order in the second priority sequence from largest to smallest is: the first posture information of the tractor, the motion information, the crossbeam or baffle, and the lane line.

[0020] In one possible implementation, determining the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the traction coupling point, respectively, includes:

[0021] Based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the traction coupling point, the relative position and relative angle of the center of the trailer relative to the tractor are determined.

[0022] The relative position and the relative angle are determined as the relative pose information of the trailer.

[0023] In one possible implementation, the preset method is determined based on the pose information of the tractor in the previous frame;

[0024] Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes:

[0025] Obtain the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame. The pose information includes: timestamp, speed, and absolute orientation.

[0026] The first angle value is determined based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

[0027] In one possible implementation, the preset method is determined based on the first point cloud data of the trailer's crossbeam;

[0028] Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes:

[0029] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the trailer's crossbeam is determined;

[0030] Based on the region of interest of the crossbeam, filter the third point cloud data corresponding to the trailer to obtain the first point cloud data of the crossbeam;

[0031] Based on the first point cloud data, the equation of the first straight line of the beam on the horizontal plane is determined;

[0032] The first angle value is determined based on the equation of the first straight line.

[0033] In one possible implementation, the preset method is determined based on the second point cloud data of the trailer's fender;

[0034] Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes:

[0035] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the baffle is determined.

[0036] Based on the region of interest of the baffle, the fourth point cloud data corresponding to the trailer is filtered to obtain the second point cloud data of the baffle;

[0037] Based on the second point cloud data, the equation of the second straight line of the baffle on the horizontal plane is determined;

[0038] The first angle value is determined based on the second straight line equation.

[0039] In one possible implementation, the preset method is determined based on image data including lane lines behind the trailer;

[0040] Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes:

[0041] Image data including lane lines behind the trailer is input into the lane line detection model to obtain the position information of the lane lines;

[0042] Based on the position information of the lane line, determine the fifth angle value of the lane line relative to the trailer;

[0043] The first angle value is determined based on the fifth angle value and the absolute orientation of the lane line.

[0044] In one possible implementation, the preset method is determined based on the motion information of the metering wheel on the trailer contact surface;

[0045] Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes:

[0046] The motion information of the metering wheel collected by the rotary encoder on the coupling base of the tractor is obtained;

[0047] Based on the motion information, determine the change in angle of the trailer relative to the tractor.

[0048] The first angle value is determined based on the second angle value of the trailer relative to the tractor and the angle change amount in the previous frame.

[0049] Secondly, embodiments of this application provide a device for detecting the relative pose of a trailer, comprising:

[0050] The acquisition module is used to acquire the first angle value of the trailer relative to the tractor in the current frame and the first timestamp corresponding to the current frame under the multiple first angle values, determined by multiple preset methods. The multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the contact surface of the trailer.

[0051] The detection module is used to perform anomaly detection on the at least two preset methods based on the plurality of first angle values ​​and the plurality of first timestamps to obtain the target angle value;

[0052] The determination module is used to determine the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

[0053] In one possible implementation, the detection module is specifically used for:

[0054] Obtain the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp of the previous frame corresponding to the second angle value;

[0055] For any given first angle value, determine a first difference between the first angle value and the second angle value, and a second difference between the first timestamp and the second timestamp;

[0056] If the ratio of the first difference to the second difference is less than a first preset threshold, then the first angle value is determined to be a non-abnormal third angle value. The first preset threshold is determined based on the operating information of the tractor.

[0057] The target angle value is determined based on at least one third angle value.

[0058] In one possible implementation, the detection module determines the target angle value based on at least one third angle value, specifically for:

[0059] If the number of the at least one third angle value is an integer greater than 1, then the two fourth angle values ​​with the smallest angle difference are determined from the at least one third angle value.

[0060] If the angle difference between the two fourth angle values ​​is greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the first priority sequence. The priority order in the first priority sequence from largest to smallest is: crossbeam or baffle, lane line, motion information, and first posture information of the tractor.

[0061] If the angle difference between the two fourth angle values ​​is not greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the second priority sequence. The priority order in the second priority sequence from largest to smallest is: the first posture information of the tractor, the motion information, the crossbeam or baffle, and the lane line.

[0062] In one possible implementation, the determining module is specifically used for:

[0063] Based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the traction coupling point, the relative position and relative angle of the center of the trailer relative to the tractor are determined.

[0064] The relative position and the relative angle are determined as the relative pose information of the trailer.

[0065] In one possible implementation, the preset method is determined based on the pose information of the tractor in the previous frame;

[0066] Accordingly, the acquisition module acquires the first angle value between the trailer and the tractor in the current frame, determined by the preset method, specifically for:

[0067] Obtain the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame. The pose information includes: timestamp, speed, and absolute orientation.

[0068] The first angle value is determined based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

[0069] In one possible implementation, the preset method is determined based on the first point cloud data of the trailer's crossbeam;

[0070] Accordingly, the acquisition module acquires the first angle value between the trailer and the tractor in the current frame, determined by the preset method, specifically for:

[0071] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the trailer's crossbeam is determined;

[0072] Based on the region of interest of the crossbeam, filter the third point cloud data corresponding to the trailer to obtain the first point cloud data of the crossbeam;

[0073] Based on the first point cloud data, the equation of the first straight line of the beam on the horizontal plane is determined;

[0074] The first angle value is determined based on the equation of the first straight line.

[0075] In one possible implementation, the preset method is determined based on the second point cloud data of the trailer's fender;

[0076] Accordingly, the acquisition module acquires the first angle value between the trailer and the tractor in the current frame, determined by the preset method, specifically for:

[0077] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the crossbeam of the baffle is determined;

[0078] Based on the region of interest of the crossbeam, the fourth point cloud data corresponding to the trailer is filtered to obtain the second point cloud data of the baffle;

[0079] Based on the second point cloud data, the equation of the second straight line of the baffle on the horizontal plane is determined;

[0080] The first angle value is determined based on the second straight line equation.

[0081] In one possible implementation, the preset method is determined based on image data including lane lines behind the trailer;

[0082] Accordingly, the acquisition module acquires the first angle value between the trailer and the tractor in the current frame, determined by the preset method, specifically for:

[0083] Image data including lane lines behind the trailer is input into the lane line detection model to obtain the position information of the lane lines;

[0084] Based on the position information of the lane line, determine the fifth angle value of the lane line relative to the trailer;

[0085] The first angle value is determined based on the fifth angle value and the absolute orientation of the lane line.

[0086] In one possible implementation, the preset method is determined based on the motion information of the metering wheel on the trailer contact surface;

[0087] Accordingly, the acquisition module acquires the first angle value between the trailer and the tractor in the current frame, determined by the preset method, specifically for:

[0088] The motion information of the metering wheel collected by the rotary encoder on the coupling base of the tractor is obtained;

[0089] Based on the motion information, determine the change in angle of the trailer relative to the tractor.

[0090] The first angle value is determined based on the second angle value of the trailer relative to the tractor and the angle change amount in the previous frame.

[0091] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0092] The memory stores computer-executed instructions;

[0093] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0094] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0095] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0096] The trailer relative pose detection method and apparatus provided in this application obtain a first angle value of the trailer relative to the tractor in the current frame and a first timestamp corresponding to the current frame for the first angle value, determined by multiple preset methods. These multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the trailer's contact surface. Based on the multiple first angle values ​​and multiple first timestamps, anomaly detection is performed on at least two preset methods to obtain a target angle value. Based on the target angle value, the trailer's size information, and the distance information between the trailer and the tractor and their respective traction coupling points, the relative pose information of the trailer is determined, thereby improving the accuracy of trailer relative pose detection. Attached Figure Description

[0097] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0098] Figure 1 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0099] Figure 2 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0100] Figure 3 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0101] Figure 4 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0102] Figure 5 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0103] Figure 6 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0104] Figure 7 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application.

[0105] Figure 8 is a schematic diagram of the structure of the trailer relative pose detection device provided in this application;

[0106] Figure 9 is a schematic diagram of the structure of the electronic device provided in this application.

[0107] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0108] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0109] First, let me explain the terms used in this application:

[0110] Lane markings: These are the lines used to divide different lanes on a road surface.

[0111] Point cloud data refers to a data structure used to represent the geometric shape of a three-dimensional object or environment. It consists of a set of points in three-dimensional space, each point typically containing spatial coordinates and possibly other information such as color and reflectance intensity.

[0112] Secondly, a description of the technical background involved in this application:

[0113] With the rapid development of logistics and transportation, the demand for high-precision detection of the relative position and attitude of trailers is increasing. For example, in automated cargo loading and unloading operations, it is necessary to accurately know the position and attitude of the trailer relative to the tractor in order to achieve efficient and safe cargo transportation.

[0114] Traditional methods for detecting the relative position and orientation of trailers mainly rely on sensors to estimate the vehicle's position and attitude. Then, based on the vehicle's position and attitude, the angle between the trailer and the tractor is estimated as the final angle.

[0115] However, existing methods for detecting the relative position of trailers have low accuracy and are difficult to represent the actual relative position of trailers.

[0116] The method and apparatus for detecting the relative pose of a trailer provided in this application aim to solve the above-mentioned technical problems of the prior art. The inventive concept of this application is as follows: The influencing factors of the relative pose information of the trailer include the angle value of the trailer relative to the tractor, the size information of the trailer, and the distance information of the trailer and the tractor respectively to the traction coupling point. At this time, multiple angle values ​​of the trailer relative to the tractor and corresponding timestamps can be obtained from different data sources. The data sources can be radar, cameras, motion simulation, rotary encoders, etc. Then, anomaly detection can be performed on the angle values ​​obtained from different sources to remove unreliable angle values ​​and obtain a more accurate angle value of the trailer relative to the tractor. Thus, based on the reliable angle value, a more accurate relative pose of the trailer can be determined.

[0117] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0118] Figure 1 is a flowchart illustrating the method for detecting the relative pose of a trailer provided in this application. As shown in Figure 1, the method includes:

[0119] The electronic device can be the vehicle itself or a control device installed on the vehicle.

[0120] Step 11: Obtain the first angle value of the trailer relative to the tractor in the current frame determined by multiple preset methods, and the first timestamp corresponding to the current frame under multiple first angle values;

[0121] Among them, the preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the crossbeam of the trailer, determination based on the second point cloud data of the trailer baffle, determination based on the image data of the trailer rear including lane lines, and determination based on the motion information of the metering wheel on the trailer contact surface.

[0122] In this step, there are multiple preset methods to determine the first angle value of the trailer relative to the tractor in the current frame. At least two preset methods are selected, and then at least two first angle values ​​are obtained according to the specific content of the selected at least two preset methods. The first timestamps corresponding to the current frame under the at least two first angle values ​​are obtained respectively.

[0123] In one possible implementation, the first point cloud data refers to the three-dimensional spatial data describing the crossbeams of the trailer; the second point cloud data refers to the three-dimensional spatial data describing the baffles of the trailer.

[0124] In one possible implementation, preset method 1 can be determined based on the first pose information of the tractor in the previous frame, preset method 2 can be determined based on the first point cloud data of the trailer's crossbeam, preset method 3 can be determined based on the second point cloud data of the trailer's baffle, preset method 4 can be determined based on image data including lane lines behind the trailer, and preset method 5 can be determined based on the motion information of the metering wheel on the trailer's contact surface.

[0125] The first angle value of the trailer relative to the tractor in the current frame can be determined by two preset methods, such as preset method 1 and preset method 2, preset method 3 and preset method 4, etc.; it can be determined by three preset methods, such as preset method 1, preset method 2 and preset method 3, or preset method 2, preset method 3 and preset method 4, etc.; or it can be determined by four preset methods, such as preset method 1, preset method 2, preset method 3 and preset method 4.

[0126] In addition, the first timestamp corresponding to the current frame under multiple first angle values ​​can be obtained from log data or timestamp data contained in relevant information and data in multiple preset methods.

[0127] Step 12: Based on multiple first angle values ​​and multiple first timestamps, perform anomaly detection on at least two preset methods to obtain the target angle value.

[0128] In this step, based on multiple first angle values ​​and multiple first timestamps, anomaly detection can be performed on at least two preset methods for determining multiple first angle values ​​and multiple first timestamps. The detected abnormal values ​​are processed, and at least one processed first angle value is used as the target angle value.

[0129] For example, anomaly detection refers to anomaly processing of multiple first angle values ​​and multiple first timestamps determined by at least two preset methods; target angle values ​​refer to multiple first angle values ​​obtained after anomaly detection.

[0130] In one possible implementation, anomaly detection can be performed on multiple first angle values ​​of the current frame based on multiple historical first angle values ​​and multiple first timestamps determined by at least two preset methods. Depending on different actual application scenarios, different thresholds for trailer angle change rates can be set. When the trailer angle change rate is greater than the threshold corresponding to the application scenario, it indicates that an abnormal trailer angle change value has occurred, and the abnormal trailer angle change value is deleted.

[0131] For example, practical application scenarios could include reversing, turning, stable driving, hill start, or hill climbing. Different thresholds can be set for the tractor's operating information in different scenarios.

[0132] Step 13: Determine the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point respectively.

[0133] In this step, the distance information between the trailer and the traction coupling point can be obtained based on the position of the traction coupling point and the position of the trailer. At the same time, the distance information between the tractor and the traction coupling point can be obtained based on the position of the traction coupling point and the position of the tractor. Then, based on the target angle value, the angle of the trailer relative to the tractor is determined. Based on the length and width information in the trailer's size information, the positions of the four sides of the trailer relative to the tractor are determined. Finally, the relative pose information of the trailer is determined.

[0134] For example, the traction coupling point refers to the connection center point between the trailer and the tractor, and the relative pose information of the trailer refers to the information representing the position and attitude of the trailer relative to the tractor.

[0135] In one possible implementation, the distance information between the trailer and the traction coupling point can be the straight-line distance between the center point of the rear axle of the trailer and the traction coupling point, and the distance information between the tractor and the traction coupling point can be the straight-line distance between the center point of the rear axle of the tractor and the traction coupling point.

[0136] Optionally, step 13 may include the following: determining the relative position and relative angle of the trailer's center relative to the tractor based on the target angle value, the trailer's size information, and the distance information between the trailer and the tractor and the traction coupling point respectively; and determining the relative position and relative angle as the trailer's relative pose information.

[0137] In this implementation, the center of the trailer can be determined based on its size information (such as length, width, and height). Then, based on the distance information between the center of the trailer and the center of the tractor and the towing coupling point,

[0138] Furthermore, using the traction coupling point as the origin of the reference coordinate system, the distance between the trailer and the traction coupling point and the coordinates of the trailer, as well as the distance between the tractor and the traction coupling point and the coordinates of the tractor, are obtained. Coordinate calculations are performed to obtain the relative position of the trailer's center relative to the tractor. Based on the coordinate relationship, trigonometric calculations are performed to obtain the relative angle of the trailer's center relative to the tractor. The relative position and relative angle are then determined as the relative pose information of the trailer.

[0139] In one possible implementation, the center of the trailer can be the center of the front axle or the center of the rear axle of the trailer, and the center of the tractor can be the center of the front axle or the center of the rear axle of the tractor.

[0140] The trailer relative pose detection method provided in this application obtains a first angle value of the trailer relative to the tractor in the current frame and a first timestamp corresponding to the current frame for the first angle value, determined by multiple preset methods. These multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the trailer's contact surface. Based on the multiple first angle values ​​and multiple first timestamps, anomaly detection is performed on at least two preset methods to obtain a target angle value. Based on the target angle value, the trailer's size information, and the distance information between the trailer and the tractor and their respective traction coupling points, the relative pose information of the trailer is determined, thereby improving the accuracy of trailer relative pose detection.

[0141] Based on the above embodiments, Figure 2 is a second flowchart of the trailer relative pose detection method provided in this application. As shown in Figure 2, step 12 above may include the following implementation:

[0142] Step 21: Obtain the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp of the previous frame corresponding to the second angle value.

[0143] In this step, relevant information from the previous frame can be detected, and the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp corresponding to the second angle value in the previous frame can be obtained.

[0144] In practice, depending on the preset method, the second angle value and the second timestamp can be the angle value and timestamp of the previous frame under the preset method; or they can be the standard angle and standard timestamp of the previous frame under actual conditions.

[0145] For example, it could be based on historical records or logs to obtain the second angle value of the trailer relative to the tractor in the previous frame and the record of the second timestamp corresponding to the second angle value in the previous frame.

[0146] Step 22: For any first angle value, determine the first difference between the first angle value and the second angle value, and the second difference between the first timestamp and the second timestamp.

[0147] In this step, for each first angle value, the difference between the first angle value and the second angle value can be calculated and recorded as the first difference; the difference between the first timestamp and the second timestamp can be calculated and recorded as the second difference.

[0148] Optionally, before step 22, the following can also be performed: filtering any of the first angle value and the second angle value mentioned above. Kalman filtering can be used to predict the first angle value based on the second angle value, compare the predicted first angle value with the existing first angle value, and smooth the first angle value containing noise to eliminate noise fluctuations.

[0149] Step 23: If the ratio of the first difference to the second difference is less than the first preset threshold, then the first angle value is determined to be the third angle value that is not abnormal.

[0150] The first preset threshold is determined based on the operating information of the tractor.

[0151] In this step, the first difference is divided by the second difference to obtain a ratio. Based on the tractor's operating information, a first preset threshold for the current operating state is determined. The ratio is compared with the first preset threshold. If the ratio is less than the first preset threshold, the first angle value is taken as the non-abnormal third angle value.

[0152] In one possible implementation, different application scenarios could be: reversing, turning, climbing, and going straight. The operating information of the tractor could be the vehicle's speed, gear information, steering wheel angle, etc., in the corresponding scenario.

[0153] A reasonable first preset threshold can be determined based on the tractor's operating information. The angle between the tractor and trailer changes differently under different operating conditions. When the operating information indicates that the tractor is traveling at a constant speed in a straight line, the relative angle between the tractor and trailer changes slightly, and the first preset threshold is relatively small. When the tractor is determined to be turning, its turning angle will affect the trailer's path and angle, causing a change in the relative angle between the tractor and trailer, and the first preset threshold will be relatively large.

[0154] Step 24: Determine the target angle value based on at least one third angle value.

[0155] In this step, at least one third angle value can be filtered based on a preset rule to obtain the filtered third angle value, which can then be used as the target angle value.

[0156] Optionally, step 24 above may include the following implementation:

[0157] Step 1: If the number of at least one third angle value is an integer greater than 1, then determine the two fourth angle values ​​with the smallest angle difference among at least one third angle value.

[0158] For example, if the number of at least one third angle value is 2, then these two third angle values ​​are determined as two fourth angle values; if the number of at least one third angle value is 3, then calculate the angle difference 1 between third angle value 1 and third angle value 2, the angle difference 2 between third angle value 1 and third angle value 3, and the angle difference 3 between third angle value 2 and third angle value 3, and compare the size of angle difference 1, angle difference 2, and angle difference 3, and take the two third angle values ​​with the smallest angle difference as two fourth angle values.

[0159] In one possible implementation, if the number of at least one third angle value is 1, then at least one third angle value is determined to be the target angle value.

[0160] Step 2: If the angle difference between two fourth angle values ​​is greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the first priority sequence. The priority order in the first priority sequence from largest to smallest is: crossbeam or baffle, lane line, motion information, and first position information of the tractor.

[0161] In this implementation, the angle difference between two fourth angle values ​​is compared with a second preset threshold. If it is greater than the second preset threshold, the fourth angle value with higher priority is taken as the target angle value according to the priority order in the first priority sequence.

[0162] The first priority sequence refers to the priority of various preset methods, that is, the priority order is: radar (crossbeam or baffle), camera (lane line), rotary encoder (motion information), motion deduction (first attitude information of the tractor).

[0163] Step 3: If the angle difference between the two fourth angle values ​​is not greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the second priority sequence. The priority order in the second priority sequence from largest to smallest is: the first posture information of the tractor, motion information, crossbeam or baffle, and lane line.

[0164] In this implementation, the angle difference between two fourth angle values ​​is compared with a second preset threshold. If it is not greater than the second preset threshold, the fourth angle value with higher priority is taken as the target angle value according to the priority order in the second priority sequence.

[0165] For example, the second priority sequence refers to the priority of multiple preset methods, that is, the priority order is: motion deduction (first posture information of the tractor), rotary encoder (motion information), radar (crossbeam or baffle), camera (lane line).

[0166] It should be understood that the examples of quantities involved in this application are only for illustrating the principle of the scheme and do not limit the actual quantities.

[0167] The trailer relative pose detection method provided in this application obtains a second angle value of the trailer relative to the tractor in the previous frame and a second timestamp corresponding to the second angle value in the previous frame. Then, for any first angle value, it determines a first difference between the first and second angle values, and a second difference between the first and second timestamps. If the ratio of the first difference to the second difference is less than a first preset threshold, the first angle value is determined to be a non-abnormal third angle value. The first preset threshold is determined based on the tractor's operating information. Finally, a target angle value is determined based on at least one third angle value. This technical solution uses the first preset threshold to filter out angle values ​​that do not conform to vehicle operating conditions, thereby obtaining an accurate target angle value corresponding to the trailer relative to the tractor, ensuring the accuracy of subsequent trailer relative pose determination.

[0168] Based on the above embodiments, the preset method is determined based on the pose information of the tractor in the previous frame; correspondingly, Figure 3 is a flowchart of the trailer relative pose detection method provided in this application. The first angle value of the trailer relative to the tractor in the current frame determined by the preset method in step 11 above can be implemented as follows:

[0169] Step 31: Obtain the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame;

[0170] The first pose information and the second pose information can both include the timestamp, velocity, and absolute orientation of the corresponding frame.

[0171] In this step, the current frame information is first determined, and the second pose information of the tractor in the current frame is obtained. Based on the current frame information, the previous frame information is determined, and then the first pose information of the tractor in the previous frame is obtained.

[0172] In one possible implementation, the timestamp could be a standard time format for recording trailer pose data acquisition; the speed could be the trailer's linear or angular velocity, which helps to dynamically track changes in trailer motion; and the absolute orientation could be an orientation determined based on latitude, longitude, and altitude information.

[0173] Step 32: Determine the first angle value based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point respectively.

[0174] In this step, the distance information between the trailer and the tractor and the traction coupling point (i.e., the distance information between the trailer and the traction coupling point and the distance information between the tractor and the traction coupling point) can be obtained based on the structural data (dimensional data) of the trailer and the tractor. Then, based on the first pose information, the second pose information, and the distance information, they are input into the kinematic deduction model to obtain the first angle value of the trailer relative to the tractor.

[0175] The trailer-to-tractor relative pose detection method provided in this application obtains the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame. This pose information includes a timestamp, speed, and absolute orientation. Then, based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point, a first angle value is determined. In this technical solution, the first angle value of the trailer relative to the tractor can be obtained based on the pose information of the previous and current frames, as well as the vehicle's structural data.

[0176] Based on the above embodiments, the preset method is determined based on the first point cloud data of the trailer's crossbeam; correspondingly, Figure 4 is a flowchart of the trailer relative pose detection method provided in this application. The step 11 above, obtaining the first angle value of the trailer relative to the tractor in the current frame determined by the preset method, can include the following implementation:

[0177] Among them, a sensor lidar is installed on the tractor, and the installation position and angle can cover the entire trailer. The point cloud data of the entire trailer is further obtained based on the sensor lidar, and the sensor data is converted to a coordinate system fixed on the tractor (such as the center of the tractor) to obtain the third point cloud data.

[0178] Step 41: Determine the region of interest of the trailer's crossbeam based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame.

[0179] In this step, the distance information between the trailer and the tractor and the traction coupling point, as well as the angle information of the trailer in the previous frame, can be analyzed to obtain position data and angle data related to the region of interest of the crossbeam. Furthermore, based on the length and width of the trailer crossbeam, the boundary of the region of interest of the crossbeam is determined, thereby determining the region of interest of the trailer crossbeam.

[0180] In one possible implementation, the specific position of the trailer relative to the tractor can be determined based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame. This leads to the specific position of the trailer crossbeam relative to the tractor. Through coordinate transformation, the coordinate position of the crossbeam is transformed from the trailer's local coordinate system to one of the aforementioned coordinate systems of the tractor and trailer. Simultaneously, considering the width and length of the trailer crossbeam, the region of interest for the trailer crossbeam is determined.

[0181] Step 42: Based on the region of interest of the crossbeam, filter the third point cloud data corresponding to the trailer to obtain the first point cloud data of the crossbeam.

[0182] In this step, the third point cloud data mentioned above is obtained. Based on the relative area of ​​the trailer crossbeam to the overall area of ​​the trailer, the point cloud data corresponding to the relative area in the third point cloud data is extracted. The extracted point cloud data is used as the first point cloud data of the crossbeam.

[0183] Step 43: Based on the first point cloud data, determine the equation of the first straight line of the beam on the horizontal plane.

[0184] In this step, the random sampling consistency method can be used to fit the first point cloud data to obtain the first straight line equation of the beam on the horizontal plane.

[0185] In one possible implementation, the normal vector (n) of each point is first estimated from the first point cloud data of the region of interest of the beam. x n y n z ), where n x *n x +n y *n y +n z *n z =1. Project the normal vectors of each point onto the xy plane (i.e., the horizontal plane) to obtain a two-dimensional vector (n). x n y According to relation n x *n x +n y *n y >T filtering obtains points in the point cloud whose normal vectors are approximately parallel to the xy plane, i.e., points that fall on the vertical plane of the trailer crossbeam, where T is a set threshold, such as 0.99.

[0186] The random sampling consistency method is used to perform plane fitting on the points falling on the vertical plane of the trailer crossbeam. The resulting equation of the vertical plane of the trailer crossbeam is ax + by + cz + d = 0. At the same time, the interior points in the fitting process are obtained, which are the points used to determine the equation of the vertical plane of the trailer crossbeam in the final fitting process.

[0187] Based on the points within the vertical plane equation of the trailer crossbeam, all z-values ​​of the point cloud are assigned to 0, effectively compressing the point cloud entirely onto the xy-plane, resulting in a two-dimensional point cloud. A random sampling consistency method is then used to fit a straight line to the obtained two-dimensional point cloud, yielding the first straight line equation of the trailer crossbeam in the xy-plane: ax + by + c = 0.

[0188] Step 44: Determine the first angle value based on the equation of the first straight line.

[0189] In one possible implementation, the angle value obtained by calculating arctan(b / a) based on the coefficients a and b in the first linear equation is used as the first angle value.

[0190] The trailer relative pose detection method provided in this application first determines the region of interest (ROI) of the trailer's crossbeam based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame. Then, based on the ROI, the third point cloud data corresponding to the trailer is filtered to obtain the first point cloud data of the crossbeam. Next, based on the first point cloud data, the first straight line equation of the crossbeam on the horizontal plane is determined. Finally, based on the first straight line equation, the first angle value is determined. By using the point cloud data of the trailer's crossbeam region, the angle value of the trailer relative to the tractor can be accurately mapped.

[0191] Based on the above embodiments, the preset method is determined based on the second point cloud data of the trailer's baffle; correspondingly, Figure 5 is a flowchart of the trailer relative pose detection method provided in this application. The step 11 above, obtaining the first angle value of the trailer relative to the tractor in the current frame determined by the preset method, can include the following implementation:

[0192] Among them, the tractor is equipped with a sensor lidar, and the installation position and angle can cover the entire trailer. The sensor data is converted to a coordinate system fixed on the tractor (such as the center of the tractor) to obtain the fourth point cloud data.

[0193] Step 51: Determine the region of interest for the baffle based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame.

[0194] In this step, based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the boundaries of the regions of interest for the left and right side panels of the trailer are determined, and the regions of interest for the panels include the regions of interest for the panels corresponding to the left and right side panels of the trailer.

[0195] Step 52: Based on the region of interest of the baffle, filter the fourth point cloud data corresponding to the trailer to obtain the second point cloud data of the baffle.

[0196] In this step, based on the region of interest of the trailer's baffle, the point cloud data corresponding to the region of interest of the baffle in the fourth point cloud data is extracted, and the extracted point cloud data is used as the second point cloud data of the baffle.

[0197] Step 53: Based on the second point cloud data, determine the equation of the second straight line of the baffle on the horizontal plane.

[0198] In this step, the random sampling consistency method can be used to fit the second point cloud data to obtain the second straight line equation of the baffle on the horizontal plane.

[0199] In one possible implementation, the normal vectors of each point in the region of interest of the baffle are estimated, and the normal vectors are projected onto the xy ground to obtain a two-dimensional vector. The vectors are then filtered to obtain points that fall on the vertical plane of the trailer side baffle. At the same time, the point clouds on the vertical planes of the left and right trailer side baffles are merged into a point cloud of the vertical plane of the trailer side baffle. Then, the point cloud is fitted to obtain the equation of the vertical plane of the trailer side baffle, mx + ny + pz + q = 0. The in-plane points are obtained during the fitting process, and the z-values ​​of all in-plane points are set to 0. The second straight line equation of the trailer side baffle in the xy plane, mx + ny + q = 0, is obtained by further fitting.

[0200] Step 54: Determine the first angle value based on the equation of the second straight line.

[0201] In one possible implementation, the angle value obtained by calculating arctan(n / m) based on the coefficients m and n in the second linear equation is used as the first angle value.

[0202] The trailer relative pose detection method provided in this application first determines the region of interest (ROI) for the baffle based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame. Then, based on the ROI, the fourth point cloud data corresponding to the trailer is filtered to obtain the second point cloud data of the baffle. Next, based on the second point cloud data, the second straight line equation of the baffle on the horizontal plane is determined. Finally, based on the second straight line equation, the first angle value is determined. In this technical solution, the angle value of the trailer relative to the tractor can be accurately mapped using the point cloud data of the trailer's baffle region.

[0203] Based on the above embodiments, the preset method is determined based on image data including lane lines behind the trailer; correspondingly, Figure 6 is a flowchart of the trailer relative pose detection method provided in this application. The step 11 above, obtaining the first angle value of the trailer relative to the tractor in the current frame determined by the preset method, can include the following implementation:

[0204] The sensor camera installed on the trailer must be positioned and angled to cover the ground area directly behind the trailer.

[0205] Step 61: Input the image data of the area behind the trailer, including the lane lines, into the lane line detection model to obtain the position information of the lane lines.

[0206] In this step, image data including lane lines behind the trailer is acquired based on the camera, and the image data is analyzed and processed by the lane line detection model to detect and identify lane lines in the image data, extract the coordinate data of the lane, and obtain the position information of the lane lines.

[0207] For example, a lane line detection model refers to a model that detects and identifies lane lines in an image.

[0208] In one possible implementation, the image data can be obtained by transforming the data captured by the camera into a coordinate system with the center point of the tractor as the origin, the front of the tractor as the positive x-axis, the left side of the tractor as the positive y-axis, and the top of the tractor as the positive z-axis.

[0209] Step 62: Determine the fifth angle value of the lane line relative to the trailer based on the lane line position information.

[0210] In one possible implementation, the position information of the lane line may include the coordinates of the center points at both ends of the lane line. The fifth angle value of the lane line relative to the trailer can be calculated based on the coordinates of the trailer in the coordinate system of the tractor and the coordinates of the lane line in the coordinate system of the tractor.

[0211] Step 63: Determine the first angle value based on the fifth angle value and the absolute orientation of the lane line.

[0212] In this step, the first angle value can be calculated by taking into account the absolute position of the trailer, the absolute orientation of the lane line, and the fifth angle value of the lane line relative to the trailer, and all the relevant angle values.

[0213] In one possible implementation, the absolute position of the trailer can be calculated based on the absolute position of the tractor in the previous frame and the relative position of the trailer to the tractor. Based on the absolute position of the trailer and standard map information, the absolute orientation of the lane lines in the area behind the trailer can be obtained. Then, the fifth angle value and the absolute orientation of the lane lines are added together to obtain the absolute orientation of the trailer. Finally, based on the absolute orientation of the trailer and the absolute orientation of the tractor, the first angle value of the trailer relative to the tractor is determined.

[0214] The trailer relative pose detection method provided in this application involves inputting image data of the area behind the trailer, including lane lines, into a lane line detection model to obtain the position information of the lane lines. Then, based on the position information of the lane lines, a fifth angle value relative to the trailer is determined. Finally, a first angle value is determined based on the fifth angle value and the absolute orientation of the lane lines. In this technical solution, the angle value of the trailer relative to the tractor can be accurately mapped by the orientation of the trailer's lane lines at the center point of the tractor and the absolute orientation of the lane lines.

[0215] Based on the above embodiments, the preset method is determined based on the motion information of the metering wheel on the trailer contact surface; correspondingly, Figure 7 is a flowchart of the trailer relative pose detection method provided in this application. The first angle value of the trailer relative to the tractor determined by the preset method in the current frame in step 11 above can be implemented as follows:

[0216] Step 71: Obtain the motion information of the metering wheel collected by the rotary encoder on the coupling base of the tractor vehicle.

[0217] In this step, a rotary encoder is installed on the tractor coupling base, and the meter wheel is placed on the trailer contact surface.

[0218] Then, the acquired encoder data is transformed into a coordinate system with the center point of the tractor as the origin. The front of the tractor is the positive x-axis of this coordinate system, the left side of the tractor is the positive y-axis, and the top of the tractor is the positive z-axis, so as to obtain the motion information of the metering wheel.

[0219] Step 72: Based on the motion information, determine the change in angle between the trailer and the tractor.

[0220] In this step, the motion information of the measuring wheel processed by the encoder can be converted into the change in angle between the trailer and the tractor.

[0221] Step 73: Determine the first angle value based on the second angle value and angle change of the trailer relative to the tractor in the previous frame.

[0222] In this step, the angle between the trailer and the tractor in the current frame is calculated based on the angle value of the trailer relative to the tractor in the previous frame (i.e., the second angle value) and the change in the angle value of the trailer relative to the tractor in step 72. This first angle value is the first angle value.

[0223] In one possible implementation, the first angle value is the sum of the second angle value and the angle change.

[0224] The trailer relative pose detection method provided in this application embodiment acquires the motion information of the measuring wheel collected by the rotary encoder of the tractor on the coupling base; then, based on the motion information, it determines the angular change of the trailer relative to the tractor; finally, based on the second angle value and the angular change of the trailer relative to the tractor in the previous frame, it determines the first angle value. In this technical solution, the angle of the trailer relative to the tractor can be accurately determined by the rotary encoder.

[0225] Figure 8 is a structural schematic diagram of the trailer relative pose detection device provided in this application. As shown in Figure 8, the trailer relative pose detection device provided in this embodiment includes:

[0226] The acquisition module 81 is used to acquire the first angle value of the trailer relative to the tractor in the current frame and the first timestamp corresponding to the current frame under the multiple first angle values, determined by multiple preset methods. The multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the trailer's contact surface.

[0227] The detection module 82 is used to perform anomaly detection on at least two preset methods based on multiple first angle values ​​and multiple first timestamps to obtain the target angle value;

[0228] The determination module 83 is used to determine the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point respectively.

[0229] In one possible implementation, the detection module 82 is specifically used for:

[0230] Get the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp of the previous frame corresponding to the second angle value;

[0231] For any given first angle value, determine the first difference between the first angle value and the second angle value, and the second difference between the first timestamp and the second timestamp;

[0232] If the ratio of the first difference to the second difference is less than the first preset threshold, then the first angle value is determined to be the third angle value that is not abnormal. The first preset threshold is determined based on the operating information of the tractor.

[0233] Determine the target angle value based on at least one third angle value.

[0234] In one possible implementation, the detection module 82 determines the target angle value based on at least one third angle value, specifically for:

[0235] If the number of at least one third angle value is an integer greater than 1, then the two fourth angle values ​​with the smallest angle difference are determined from at least one third angle value.

[0236] If the angle difference between two fourth angle values ​​is greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the first priority sequence. The priority order in the first priority sequence from largest to smallest is: crossbeam or baffle, lane line, motion information, and first position information of the tractor.

[0237] If the angle difference between two fourth angle values ​​is not greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the second priority sequence. The priority order in the second priority sequence from largest to smallest is: the first posture information of the tractor, motion information, crossbeam or baffle, and lane line.

[0238] In one possible implementation, the determining module 83 is specifically used for:

[0239] Based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the towing coupling point, the relative position and relative angle of the center of the trailer relative to the tractor are determined.

[0240] The relative position and relative angle are determined as the relative posture information of the trailer.

[0241] In one possible implementation, the preset method is determined based on the pose information of the tractor in the previous frame;

[0242] Correspondingly, module 81 acquires the first angle value of the trailer relative to the tractor in the current frame, determined by a preset method, specifically for:

[0243] Obtain the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame. The pose information includes: timestamp, speed, and absolute orientation.

[0244] The first angle value is determined based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

[0245] In one possible implementation, the preset method is determined based on the first point cloud data of the trailer's crossbeam;

[0246] Correspondingly, module 81 acquires the first angle value of the trailer relative to the tractor in the current frame, determined by a preset method, specifically for:

[0247] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the trailer's crossbeam is determined.

[0248] Based on the region of interest of the crossbeam, filter the third point cloud data corresponding to the trailer to obtain the first point cloud data of the crossbeam;

[0249] Based on the first point cloud data, the equation of the first straight line of the beam on the horizontal plane is determined;

[0250] The first angle value is determined based on the equation of the first straight line.

[0251] In one possible implementation, the preset method is determined based on the second point cloud data of the trailer's fender;

[0252] Correspondingly, module 81 acquires the first angle value of the trailer relative to the tractor in the current frame, determined by a preset method, specifically for:

[0253] Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the baffle is determined;

[0254] Based on the region of interest of the baffle, filter the fourth point cloud data corresponding to the trailer to obtain the second point cloud data of the baffle;

[0255] Based on the second point cloud data, the equation of the second straight line of the baffle on the horizontal plane is determined;

[0256] The first angle value is determined based on the equation of the second straight line.

[0257] In one possible implementation, the preset method is determined based on image data including lane lines behind the trailer;

[0258] Correspondingly, module 81 acquires the first angle value of the trailer relative to the tractor in the current frame, determined by a preset method, specifically for:

[0259] Image data including lane lines behind the trailer is input into the lane line detection model to obtain the position information of the lane lines;

[0260] Based on the position information of the lane lines, determine the fifth angle value of the lane lines relative to the trailer;

[0261] The first angle value is determined based on the fifth angle value and the absolute orientation of the lane lines.

[0262] In one possible implementation, the preset method is determined based on the motion information of the metering wheel on the trailer contact surface;

[0263] Correspondingly, module 81 acquires the first angle value of the trailer relative to the tractor in the current frame, determined by a preset method, specifically for:

[0264] Acquire the motion information of the metering wheel collected by the rotary encoder on the coupling base of the tractor;

[0265] Based on the motion information, determine the change in angle between the trailer and the tractor.

[0266] The first angle value is determined based on the second angle value and angle change of the trailer relative to the tractor in the previous frame.

[0267] The trailer relative pose detection device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0268] Figure 9 is a schematic diagram of the structure of the electronic device provided in this application. As shown in Figure 9, the electronic device 9 provided in this embodiment includes at least one processor 91 and a memory 92.

[0269] Optionally, the device 9 also includes a communication component 93. The processor 91, memory 92, and communication component 93 are connected via a bus 94.

[0270] In a specific implementation, at least one processor 91 executes computer execution instructions stored in memory 92, causing at least one processor 91 to perform the above-described method.

[0271] The specific implementation process of processor 91 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0272] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0273] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0274] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0275] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0276] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0277] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0278] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0279] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0280] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0281] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0282] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0283] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0284] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

A method for detecting the relative pose of a trailer, characterized in that, include: The first angle value of the trailer relative to the tractor in the current frame, determined by multiple preset methods, and the first timestamp corresponding to the current frame under the multiple first angle values ​​are obtained. The multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the contact surface of the trailer. Based on the plurality of first angle values ​​and the plurality of first timestamps, anomaly detection is performed on the at least two preset methods to obtain the target angle value; The relative pose information of the trailer is determined based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point, respectively. The method according to claim 1, characterized in that, The step of performing anomaly detection on the at least two preset methods based on the plurality of first angle values ​​and the plurality of first timestamps to obtain the target angle value includes: Obtain the second angle value of the trailer relative to the tractor in the previous frame and the second timestamp of the previous frame corresponding to the second angle value; For any given first angle value, determine a first difference between the first angle value and the second angle value, and a second difference between the first timestamp and the second timestamp; If the ratio of the first difference to the second difference is less than a first preset threshold, then the first angle value is determined to be a non-abnormal third angle value. The first preset threshold is determined based on the operating information of the tractor. The target angle value is determined based on at least one third angle value. The method according to claim 2, characterized in that, Determining the target angle value based on at least one third angle value includes: If the number of the at least one third angle value is an integer greater than 1, then the two fourth angle values ​​with the smallest angle difference are determined among the at least one third angle value; If the angle difference between the two fourth angle values ​​is greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the first priority sequence. The priority order in the first priority sequence from largest to smallest is: crossbeam or baffle, lane line, motion information, and first posture information of the tractor. If the angle difference between the two fourth angle values ​​is not greater than the second preset threshold, the target angle value is determined from the fourth angle values ​​based on the second priority sequence. The priority order in the second priority sequence from largest to smallest is: the first posture information of the tractor, the motion information, the crossbeam or baffle, and the lane line. The method according to claim 1, characterized in that, The step of determining the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the traction coupling point, respectively, includes: Based on the target angle value, the size information of the trailer, and the distance information of the trailer and the tractor to the traction coupling point, the relative position and relative angle of the center of the trailer relative to the tractor are determined. The relative position and the relative angle are determined as the relative pose information of the trailer. The method according to any one of claims 1-4, characterized in that, The preset method is determined based on the pose information of the tractor in the previous frame; Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes: Obtain the first pose information of the tractor in the previous frame and the second pose information of the tractor in the current frame. The pose information includes: timestamp, speed, and absolute orientation. The first angle value is determined based on the first pose information, the second pose information, and the distance information between the trailer and the tractor and the traction coupling point, respectively. The method according to any one of claims 1-4, characterized in that, The preset method is determined based on the first point cloud data of the crossbeam of the trailer; Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes: Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the trailer's crossbeam is determined; Based on the region of interest of the crossbeam, filter the third point cloud data corresponding to the trailer to obtain the first point cloud data of the crossbeam; Based on the first point cloud data, the equation of the first straight line of the beam on the horizontal plane is determined; The first angle value is determined based on the equation of the first straight line. The method according to any one of claims 1-4, characterized in that, The preset method is determined based on the second point cloud data of the trailer's baffle; Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes: Based on the distance information between the trailer and the tractor and the traction coupling point, and the angle information of the trailer in the previous frame, the region of interest of the baffle is determined. Based on the region of interest of the baffle, the fourth point cloud data corresponding to the trailer is filtered to obtain the second point cloud data of the baffle; Based on the second point cloud data, the equation of the second straight line of the baffle on the horizontal plane is determined; The first angle value is determined based on the second straight line equation. The method according to any one of claims 1-4, characterized in that, The preset method is determined based on image data including lane lines behind the trailer; Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes: Image data including lane lines behind the trailer is input into the lane line detection model to obtain the position information of the lane lines; Based on the position information of the lane line, determine the fifth angle value of the lane line relative to the trailer; The first angle value is determined based on the fifth angle value and the absolute orientation of the lane line. The method according to any one of claims 1-4, characterized in that, The preset method is determined based on the motion information of the meter wheel on the trailer contact surface; Accordingly, obtaining the first angle value of the trailer relative to the tractor in the current frame, determined by the preset method, includes: The motion information of the metering wheel collected by the rotary encoder on the coupling base of the tractor is obtained; Based on the motion information, determine the change in angle of the trailer relative to the tractor. The first angle value is determined based on the second angle value of the trailer relative to the tractor and the angle change amount in the previous frame. A device for detecting the relative position and orientation of a trailer, characterized in that, include: The acquisition module is used to acquire the first angle value of the trailer relative to the tractor in the current frame and the first timestamp corresponding to the current frame under the multiple first angle values, determined by multiple preset methods. The multiple preset methods include at least two of the following: determination based on the first pose information of the tractor in the previous frame, determination based on the first point cloud data of the trailer's crossbeam, determination based on the second point cloud data of the trailer's baffle, determination based on image data including lane lines behind the trailer, and determination based on the motion information of the metering wheel on the contact surface of the trailer. The detection module is used to perform anomaly detection on the at least two preset methods based on the plurality of first angle values ​​and the plurality of first timestamps to obtain the target angle value; The determination module is used to determine the relative pose information of the trailer based on the target angle value, the size information of the trailer, and the distance information between the trailer and the tractor and the traction coupling point, respectively.

Images