Real-time vehicle slip perception methods and systems
By combining forward-looking binocular vision sensors and wheel speed sensors to calculate wheel slip, the problem of insufficient accuracy and reliability of traditional methods under multi-wheel slip conditions is solved, achieving more accurate and reliable slip perception and improving the safety and stability of vehicles under extreme conditions.
Patent Information
- Application Number
- CN202611627105.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-21
- Publication Date
- 2026-11-17
AI Technical Summary
In existing technologies, under complex conditions such as simultaneous slippage of multiple wheels, the accuracy and reliability of estimating slippage using wheel speed sensors decrease, leading to erroneous decisions by the vehicle stability control system and affecting driving safety.
By combining a forward-looking binocular vision sensor and a wheel speed sensor, the slippage is quantitatively calculated by calculating the difference between the actual and nominal displacement of the wheel, thus eliminating the reliance on a single sensor.
It improves the accuracy and reliability of slip perception, enhances adaptability to complex environments, and improves the vehicle's handling stability and safety under extreme conditions.
Smart Images

Figure CN123104667A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of vehicle state perception technology, specifically relating to a method and system for real-time vehicle slippage perception. Background Technology
[0002] Vehicle safety and stability largely depend on advanced electronic control systems. The effective operation of these systems relies on the accurate and real-time acquisition of key dynamic parameters of the vehicle. Among these, wheel slip ratio or slip amount is crucial. Precise slip amount information is the core basis for determining the adhesion between the wheels and the road surface and for corresponding braking or driving torque adjustments.
[0003] Currently, the most common method for estimating slip in commercial vehicles relies primarily on wheel speed sensors. The basic principle is to compare the rotational speeds of each wheel, or to convert the wheel speeds into linear velocities and compare them with a baseline vehicle speed. However, this method has inherent limitations. The baseline vehicle speed itself is usually estimated indirectly by filtering or logically selecting the wheel speeds of non-driving wheels. During emergency braking, rapid acceleration, or driving on low-traction surfaces such as ice, snow, or wet surfaces, all wheels may simultaneously experience varying degrees of slippage (skid or drag). In such cases, the baseline vehicle speed estimated based on wheel speed data becomes severely distorted and can no longer serve as a reliable reference.
[0004] Therefore, the traditional method that relies solely on wheel speed sensors will significantly reduce the accuracy and reliability of slip estimation under the aforementioned extreme or complex conditions, and may even output incorrect information, which may lead to incorrect decisions by the vehicle stability control system and affect driving safety.
[0005] In summary, there is an urgent need for a slip sensing method that can break free from the cyclical dependence on its own motion state. Summary of the Invention
[0006] The purpose of this application is to provide a method and system for real-time vehicle slip perception, which aims to solve the technical problem that existing technologies rely solely on internal information sources such as wheel speed sensors, making it difficult to accurately and reliably perceive wheel slip in complex working conditions such as simultaneous slip of multiple wheels.
[0007] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for real-time vehicle slip perception, including: Acquire left and right image sequences characterizing the vehicle's driving environment, collected by a forward-looking binocular vision sensor; Acquire wheel angular velocity data, which characterizes the motion state of the wheel, collected by the wheel speed sensor; Based on the left and right image sequences, calculate the actual displacement of the wheel; Based on the wheel angular velocity data, calculate the nominal displacement of the wheel; The slippage of the wheel is quantitatively calculated based on the difference between the actual displacement and the nominal displacement.
[0008] Secondly, this application provides a real-time vehicle slippage sensing system, including: a forward-looking binocular vision sensor, a wheel speed sensor, and a wheel slippage sensing device; wherein, the wheel slippage sensing device includes: The data acquisition module is used to acquire left and right image sequences representing the vehicle's driving environment collected by the forward-looking binocular vision sensor, and to acquire wheel angular velocity data representing the wheel's motion state collected by the wheel speed sensor. The first calculation module is used to calculate the actual displacement of the wheel based on the left and right image sequences; The second calculation module is used to calculate the nominal displacement of the wheel based on the wheel angular velocity data; The third calculation module is used to quantitatively calculate the slippage of the wheel based on the difference between the actual motion displacement and the nominal motion displacement.
[0009] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the vehicle slip real-time perception method described in any one of the above.
[0010] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the vehicle slip real-time perception method described above.
[0011] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the real-time vehicle slip perception method described above.
[0012] According to the specific embodiments provided in this application, the following technical effects are disclosed: 1. Improved accuracy and reliability of slip perception. This invention provides a "ground truth" reference for slip calculation, independent of wheel rolling state, by fusing the actual displacement of the vehicle relative to the external environment obtained from binocular vision sensors. This fundamentally solves the problem of estimation failure caused by distortion of the reference vehicle speed in traditional methods under conditions such as simultaneous multi-wheel slippage, significantly improving the accuracy and reliability of perception.
[0013] 2. Enhanced adaptability to complex environments. By prioritizing the selection of static roadside objects as tracking features and eliminating feature points in dynamic target areas, this invention effectively resists interference from other road users. Combined with image enhancement, filtering, and other preprocessing techniques, this invention maintains robust performance under varying lighting, weather, and traffic conditions.
[0014] 3. Enhanced vehicle active safety. This invention can output quantitative wheel slip information in real time and with high precision, which can serve as a precise input for vehicle stability control systems (such as ABS, TCS, and ESP). This enables the control system to make more timely control responses based on more reliable vehicle state judgments, thereby effectively improving the vehicle's handling stability and driving safety under extreme conditions, which is of particular value for advanced driver assistance and autonomous driving systems. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 This is a flowchart illustrating a real-time vehicle slip perception method in one embodiment of this application.
[0017] Figure 2 This is a flowchart illustrating a real-time vehicle slip perception method in one embodiment of this application.
[0018] Figure 3 This is a schematic diagram of an embodiment of this application before polar line correction.
[0019] Figure 4 This is a schematic diagram of polar line correction in one embodiment of this application.
[0020] Figure 5 This is the principle of real-time wheel position calculation based on a binocular vision sensor in one embodiment of this application.
[0021] Figure 6 This is a schematic diagram illustrating the positioning coordinate relationship of a binocular vision sensor in one embodiment of this application. Detailed Implementation
[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0024] Reference Figure 1 and Figure 2 The real-time vehicle slip perception method provided in this embodiment of the invention is as follows: First, the left and right image sequences representing the vehicle's driving environment are acquired simultaneously by the forward-looking binocular vision sensor within a preset time period (usually a real-time window, including at least two consecutive timestamps), and wheel angular velocity data representing the wheel's motion state are acquired by the wheel speed sensor within the same preset time period.
[0025] Subsequently, based on the acquired left and right image sequences, the actual displacement of the wheel within the preset time period is calculated. Simultaneously, based on the acquired wheel angular velocity data, the nominal displacement of the wheel within the preset time period is calculated. Then, based on the difference between the actual and nominal displacements, the wheel slippage is quantitatively calculated.
[0026] Finally, the slip amount is output, for example, provided to the vehicle stability control system. This process achieves precise quantification of wheel slip state by comparing displacement information from two different sources, overcoming the limitations of traditional methods that rely on only a single sensor.
[0027] Before performing actual motion displacement calculations, a series of preprocessing operations need to be performed on the left and right image sequences to ensure the quality of the input image data. These operations include: The binocular vision sensor is calibrated for intrinsic and extrinsic parameters and distortion is corrected. The left and right image sequences after distortion correction are then subjected to epipolar correction. Finally, the left and right image sequences after epipolar correction are subjected to image enhancement and / or filtering.
[0028] First, precise endo- and extrinsic parameter calibration of the forward-looking binocular vision sensor is required.
[0029] For example, a series of images can be captured from no fewer than 14 different angles and positions using a printed checkerboard calibration board. The pixel coordinates of the checkerboard corners in the images can be extracted using a corner detection algorithm (such as functions implemented in the Matlab toolbox or OpenCV library), and combined with their known world coordinates, the camera's intrinsic and extrinsic parameters can be solved.
[0030] Specifically, the camera imaging model involves four basic coordinate systems: the vehicle coordinate system (world coordinate system), the camera coordinate system, the pixel coordinate system, and the imaging plane coordinate system. A point in space is transformed from the vehicle coordinate system to the camera coordinate system, then projected onto the imaging plane coordinate system, and finally the coordinates on the imaging plane coordinate system are transformed back to the pixel coordinate system.
[0031] Without considering distortion, the transformation model from the vehicle coordinate system to the camera coordinate system is as follows: ; The transformation model from the camera coordinate system to the imaging plane coordinate system is as follows: ; The transformation model from the imaging plane coordinate system to the pixel coordinate system is as follows: ; The transformation model from the vehicle coordinate system to the pixel coordinate system is as follows: ; in, The coordinates of the point in the pixel coordinate system Let be the coordinates of the camera lens optical center in the imaging plane coordinate system. and The actual physical size corresponding to the pixels of the image sensor. Focal length , , , For camera internal parameters, and The rotation and translation matrices are the extrinsic parameters, and together they constitute the camera's extrinsic parameters. These are the coordinates of the points in the vehicle coordinate system.
[0032] By accurately calibrating the intrinsic and extrinsic parameters of the forward-looking binocular vision sensor, the coordinates of feature points can be transformed in different coordinate systems, providing a basis for subsequent motion displacement calculations.
[0033] In practical applications, due to lens manufacturing and assembly errors, images will exhibit radial and tangential distortion, i.e., distortion of feature points in the image. An offset will occur. Therefore, distortion parameters, including the radial distortion coefficient, will be introduced during the calibration process. , , and tangential distortion coefficient , All parameters are solved by nonlinear maximum likelihood estimation.
[0034] For stereo cameras, in addition to their individual intrinsic and extrinsic parameters, it is also necessary to calibrate the relative pose relationship between the two cameras, i.e., the rotation matrix from one camera coordinate system to the other. Translation vector It can be accessed through and The calculation yielded the following result. and These represent the rotation matrices of the right and left cameras relative to the world coordinate system, respectively. for The transpose of the matrix, and These represent the translation vectors of the right and left cameras relative to the world coordinate system, respectively. After calibration, the obtained distortion coefficients can be used to correct distortion in subsequently acquired images, restoring distortion-free images and laying the foundation for subsequent accurate calculations.
[0035] After distortion correction, epipolar correction is needed to improve the efficiency and accuracy of subsequent stereo matching. This embodiment uses the Bouguet algorithm for epipolar correction. The goal of epipolar correction is to reproject the left and right images so that the optical axes of the two cameras are parallel and their imaging planes are coplanar, thus ensuring that the projection points of the same feature point in space lie on the same horizontal line in the left and right images.
[0036] Specifically, before correction, the optical centers of the left and right cameras are not parallel. The line connecting the two optical centers is called the baseline, the intersection of the image plane and the baseline is the pole, the line connecting the image point and the pole is the epipolar line, and the plane formed by the left and right epipolar lines and the baseline is the epipolar plane corresponding to the spatial point. After correction, the pole is at infinity, and the optical axes of the two cameras are parallel. The image point is at the same height on both left and right images. This is the goal of epipolar correction. After correction, subsequent stereo matching only requires searching for matching points on the left and right image planes in the same row, greatly improving efficiency.
[0037] For example, refer to Figure 3 Before correction, the matching point pairs between the left and right images may not be on the same horizontal line. (Refer to...) Figure 4 After correction, the matching point pairs between the left and right images are strictly located on the same horizontal epipolar line. Thus, when performing feature point matching, it is not necessary to perform a two-dimensional search across the entire image; instead, a one-dimensional search is required on the corresponding epipolar line. This significantly reduces the computational complexity and time consumption of matching, improving the real-time performance of the system.
[0038] Furthermore, to address the complex and varied lighting conditions in real-world driving environments, image enhancement and / or filtering can be performed on the corrected image before feature extraction. For example, in low-light scenarios such as nighttime or entering / exiting tunnels, image contrast is low and features are not readily apparent. In such cases, the contrast-limited adaptive histogram equalization (CLAHE) algorithm or adaptive contrast enhancement (ACE) algorithm can be used to improve the local contrast of the image, making details in both dark and bright areas clearer. Simultaneously, to suppress the interference of image noise (such as sensor thermal noise or noise generated by high ISO) on feature detection, algorithms such as median filtering or Gaussian filtering can be selected to smooth the image based on the noise characteristics. These preprocessing steps can significantly improve the stability and reliability of subsequent feature point detection.
[0039] After preprocessing, the system begins to calculate the vehicle's actual motion displacement based on the high-quality image data.
[0040] In this embodiment, the actual motion displacement of the wheel is calculated based on the left and right image sequences. Specifically, this includes: extracting and matching static feature points from the left and right image sequences to obtain static feature points; calculating the pose change of the wheel between consecutive time stamps using the static feature points as a reference, and determining the pose change as the actual motion displacement.
[0041] Specifically, the extraction and matching of static feature points from the left and right image sequences to obtain matched static feature point pairs includes: removing feature points located in dynamic target areas from the left and right image sequences and / or performing geometric consistency screening on the left and right image sequences to preferentially retain static feature points located on roadside static objects or road markings in the left and right image sequences; and detecting static feature points in a preset region of interest and uniformly selecting static feature points using a gridding strategy so that the uniformly selected static feature points can be used for subsequent matching.
[0042] Furthermore, the method of calculating the pose change of the wheel with reference to static feature points specifically includes: determining the coordinates of the static feature points in the vehicle coordinate system; establishing a feature point coordinate system with the static feature points as the origin, and determining the transformation relationship between the feature point coordinate system and the vehicle coordinate system based on the coordinates of the static feature points in the vehicle coordinate system; transforming the coordinates of the wheel in the vehicle coordinate system to the feature point coordinate system; and determining the pose change of the wheel based on the coordinate change of the wheel in the feature point coordinate system between two consecutive time stamps.
[0043] The basic principle of calculating the actual displacement is to inversely calculate the motion of the wheels by identifying and tracking static feature points in the vehicle's external environment.
[0044] This embodiment discloses a specific calculation path: by establishing a feature point coordinate system with the external static feature point as the origin, and calculating the displacement of the wheel over time in the feature point coordinate system.
[0045] Specifically, in the step of extracting and matching static feature points, this embodiment employs the ORB (Oriented FAST and Rotated BRIEF) algorithm. The matching process of this algorithm clearly demonstrates its workflow. The ORB algorithm combines the high speed of FAST corner detection with the low computational complexity of the BRIEF descriptor, and introduces orientation information to achieve rotation invariance, making it very suitable for automotive applications with high real-time requirements.
[0046] To avoid static feature points clustering in locally textured areas of the image, which could lead to overfitting of pose change calculations to these areas and reduce overall stability, this embodiment employs a meshing strategy. The image is divided into several meshes, and static feature points are detected independently within each mesh. The strongest static feature points are then selected based on their response values. This design ensures that the feature points used for pose change calculations are evenly distributed across the entire image, thereby improving the robustness and accuracy of the calculations.
[0047] In complex road scenes, there are numerous dynamic objects, such as other vehicles and pedestrians. Tracking points on these dynamic objects as feature points would lead to calculated vehicle displacements that include the relative motion of these objects, resulting in serious errors. To address this issue, this embodiment performs rigorous screening before feature point matching. On one hand, deep learning methods such as semantic segmentation can be used to identify dynamic target regions (such as vehicles and pedestrians) in the image and directly eliminate feature points located within these regions. On the other hand, the RANSAC (Random Sample Consensus) algorithm is used to estimate the fundamental or essential matrix between consecutive frames, and outliers that do not conform to the overall motion model are eliminated based on geometric consistency constraints. Through these two levels of screening, the system can prioritize the retention of stable feature points located on roadside static objects (such as road signs, utility poles, tree trunks, and guardrails) and road markings (such as lane lines and arrow markings), treating them as static feature points. This ensures that the calculation is based on the relative motion between the vehicle and the static world, which is crucial for obtaining accurate actual motion displacement.
[0048] During the left-right image matching stage, the epipolar constraint after binocular correction is used to limit the search range to the same row. The effective parallax range is set near the pixel and in conjunction with the relationship between parallax and distance. (equivalent to depth range) To reduce depth noise and mismatch probability caused by small parallax at long distances, the system employs Hamming distance as the matching metric, combined with cross-validation and a ratio test for initial screening. RANSAC is then used to further eliminate mismatches, resulting in stable matching point pairs. Each matching point pair corresponds to one static feature point. After obtaining stable and reliable static feature points, the system reconstructs the 3D spatial coordinates and calculates the actual displacement of the wheels.
[0049] First, for each pair of feature points that are successfully matched on the left and right images... and Its corresponding static feature points are . Reference Figure 5 and Figure 6 Static feature points can be deduced using pinhole imaging and the principle of similar triangles. The corresponding coordinates in the camera coordinate system, where It's the camera's focal length. This is the viewing distance of the stereo camera, i.e., the difference in the x-coordinates of the matching points on the left and right images. Then, the static feature points... coordinates in the camera coordinate system It can be represented as: ; ; in, Feature points The horizontal and vertical coordinates on the left image, For feature points The horizontal and vertical coordinates on the right image.
[0050] This allows us to determine the coordinates of static feature points in the camera coordinate system. Furthermore, based on the rotation and translation relationship between the camera and vehicle coordinate systems, we can determine the coordinates of the static feature points in the vehicle coordinate system. The key to solving these coordinate problems is obtaining the feature matching point pairs from the images captured by the left and right cameras. and In other words, the key to localization lies in the accuracy of matching feature points.
[0051] After obtaining the coordinates of the static feature point in the vehicle coordinate system, a new feature point coordinate system can be established within the vehicle coordinate system, with the static feature point as the origin, and the transformation relationship between the vehicle coordinate system and the feature point coordinate system is known. Since the coordinates of the wheel in the vehicle coordinate system are known, the coordinates of the wheel in the vehicle coordinate system can be further transformed to the feature point coordinate system. Then, by calculating the change in the wheel's coordinates in the feature point coordinate system between two consecutive time points, the actual displacement of the wheel can be obtained.
[0052] For example, in At any given moment, the coordinates of any static feature point in front of the vehicle's direction of travel can be obtained in the vehicle coordinate system using a forward-looking binocular vision sensor. The coordinates of the four wheels can be calculated by transforming the vehicle coordinate system and the feature point coordinate system. Coordinates of the feature point in the coordinate system at any given time ,in .
[0053] exist At that moment, the coordinates of the static feature point in the vehicle coordinate system are: Based on this, it can be calculated The coordinates of the four wheels in the feature point coordinate system at time are . arrive The actual displacement of the four wheels can be expressed as follows: .
[0054] It should be noted that this invention does not consider the height undulation of the road surface, that is, it is assumed that the vehicle is traveling on an ideal horizontal surface. Therefore, the coordinates of static feature points and wheels only consider the longitudinal and transverse dimensions.
[0055] In addition to calculating the actual displacement of the wheels, the wheel speed sensors of all four wheels can also be used to obtain... arrive The change in wheel speed of each wheel within a given time interval allows for the calculation of the nominal displacement of all four wheels. Through with By performing a differential calculation, it is possible to determine whether the wheel is slipping and to calculate the amount of wheel slippage.
[0056] The wheel slip calculated by the method of this invention has significant application value. In practical applications, the calculated slip amount or slip ratio can serve as a crucial real-time input signal, transmitted to the vehicle's stability control system via CAN bus or Ethernet. For example, when the traction control system (TCS) detects that the slip ratio of the drive wheels exceeds a safety threshold, it can immediately take measures, such as reducing engine output torque or applying slight braking to the slipping wheel, to restore traction. Similarly, when the anti-lock braking system (ABS) detects that the wheel slip ratio is too high during braking and tends to lock up, it can quickly perform pulse adjustment of the braking pressure. Because the slip amount provided by this invention is based on an "external" measurement that combines vision and wheel speed, its accuracy and reliability are far superior to methods that rely solely on wheel speed sensors for estimation. Therefore, it enables these advanced driver assistance systems to make more precise and timely responses, thereby significantly improving vehicle handling stability and safety under extreme conditions.
[0057] The present invention also provides a real-time vehicle slip perception system, comprising: a forward-looking binocular vision sensor, a wheel speed sensor, and a wheel slip perception device; wherein the wheel slip perception device comprises: The data acquisition module is used to acquire left and right image sequences representing the vehicle's driving environment collected by the forward-looking binocular vision sensor, and to acquire wheel angular velocity data representing the wheel's motion state collected by the wheel speed sensor. The first calculation module is used to calculate the actual displacement of the wheel based on the left and right image sequences; The second calculation module is used to calculate the nominal displacement of the wheel based on the wheel angular velocity data; The third calculation module is used to quantitatively calculate the wheel slip based on the difference between the actual displacement and the nominal displacement.
[0058] Specifically, the calculation of the actual displacement of the wheel based on the left and right image sequences includes: extracting and matching static feature points from the left and right image sequences to obtain static feature points; and calculating the actual displacement of the wheel between consecutive time stamps using the static feature points as a reference.
[0059] The calculation of the actual displacement of the wheel using static feature points as a reference includes: determining the coordinates of the static feature points in the vehicle coordinate system; establishing a feature point coordinate system with the static feature points as the origin and determining the transformation relationship between the feature point coordinate system and the vehicle coordinate system; transforming the wheel's coordinates in the vehicle coordinate system to the feature point coordinate system; and determining the actual displacement of the wheel based on the change in the wheel's coordinates in the feature point coordinate system between two consecutive time stamps.
[0060] It should be noted that this system is the physical implementation carrier of the real-time vehicle slip perception method in this invention. Specifically, the wheel slip perception device, as a core computing device, is the main body executing the real-time vehicle slip perception method, while the forward-looking binocular vision sensor and wheel speed sensor are used to collect the necessary data required by the method.
[0061] Therefore, for a detailed explanation of the principles or operation of this system, please refer to the detailed description of the real-time vehicle slip perception method of this invention.
[0062] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.
[0063] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0064] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.
[0065] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0066] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).
[0067] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0069] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for real-time vehicle slip perception, characterized in that, include: Acquire left and right image sequences characterizing the vehicle's driving environment, collected by a forward-looking binocular vision sensor; Acquire wheel angular velocity data, which characterizes the motion state of the wheel, collected by the wheel speed sensor; Based on the left and right image sequences, calculate the actual displacement of the wheel; Based on the wheel angular velocity data, calculate the nominal displacement of the wheel; The slippage of the wheel is quantitatively calculated based on the difference between the actual displacement and the nominal displacement.
2. The real-time vehicle slip perception method according to claim 1, characterized in that, Based on the left and right image sequences, the actual displacement of the wheel is calculated, specifically including: Static feature points are extracted and matched from the left and right image sequences to obtain static feature points; Between consecutive timestamps, the actual displacement of the wheel is calculated with reference to the static feature point.
3. The real-time vehicle slip perception method according to claim 2, characterized in that, Extraction and matching of static feature points from the left and right image sequences specifically includes: By removing feature points located within the dynamic target area in the left and right image sequences and / or performing geometric consistency screening on the left and right image sequences, static feature points located on roadside static objects or road markings in the left and right image sequences are preferentially retained.
4. The real-time vehicle slip perception method according to claim 2, characterized in that, Extracting and matching static feature points from the left and right image sequences further includes: The static feature points are detected within a preset region of interest, and a gridding strategy is used to uniformly select the static feature points for subsequent matching.
5. The real-time vehicle slip perception method according to claim 2, characterized in that, The actual displacement of the wheel is calculated using the static feature point as a reference, specifically including: Determine the coordinates of the static feature point in the vehicle coordinate system; Establish a feature point coordinate system with the static feature point as the origin and determine the transformation relationship between the feature point coordinate system and the vehicle coordinate system; The coordinates of the wheel in the vehicle coordinate system are transformed to the coordinates of the feature point. Between two consecutive timestamps, the actual displacement of the wheel is determined based on the change in the wheel's coordinates in the feature point coordinate system.
6. The real-time vehicle slip perception method according to claim 1, characterized in that, Before calculating the actual displacement of the wheel, the following steps are also included: The left and right image sequences are sequentially subjected to distortion correction, epipolar correction, image enhancement, and / or filtering.
7. The real-time vehicle slip perception method according to claim 1, characterized in that, Also includes: The amount of slippage of the wheels is provided to the vehicle's stability control system for adjusting the vehicle's braking or driving.
8. A real-time vehicle slip perception system, characterized in that, include: A forward-looking binocular vision sensor, a wheel speed sensor, and a wheel slip detection device; wherein, the wheel slip detection device includes: The data acquisition module is used to acquire left and right image sequences representing the vehicle's driving environment collected by the forward-looking binocular vision sensor, and to acquire wheel angular velocity data representing the wheel's motion state collected by the wheel speed sensor. The first calculation module is used to calculate the actual displacement of the wheel based on the left and right image sequences; The second calculation module is used to calculate the nominal displacement of the wheel based on the wheel angular velocity data; The third calculation module is used to quantitatively calculate the slippage of the wheel based on the difference between the actual motion displacement and the nominal motion displacement.
9. The vehicle skid real-time sensing system according to claim 8, characterized in that, Based on the left and right image sequences, the actual displacement of the wheel is calculated, specifically including: Static feature points are extracted and matched from the left and right image sequences to obtain static feature points; Between consecutive timestamps, the actual displacement of the wheel is calculated with reference to the static feature point.
10. The vehicle skid real-time sensing system according to claim 9, characterized in that, The actual displacement of the wheel is calculated using the static feature point as a reference, specifically including: Determine the coordinates of the static feature point in the vehicle coordinate system; Establish a feature point coordinate system with the static feature point as the origin and determine the transformation relationship between the feature point coordinate system and the vehicle coordinate system; The coordinates of the wheel in the vehicle coordinate system are transformed to the coordinates of the feature point. Between two consecutive timestamps, the actual displacement of the wheel is determined based on the change in the wheel's coordinates in the feature point coordinate system.
