A vehicle speed detection method based on an epipolar image

Abstract

The present application relates to a kind of vehicle speed detection methods based on outer pole face image, belong to the field of automatic driving.The method includes: S1: periodically capture the image of the road in front of vehicle, obtain time-space distance image;S2: the curve cluster surface of the time-space distance image obtained is divided into several parts;S3: the curve corresponding to each part obtained;S4: by its analytic differential, the speed curve of each part is obtained;S5: the generation of full interval speed curve, according to the displacement of feature point, the speed of vehicle is calculated.The present application is based on the principle of visual kinematics, using the image obtained by vehicle-mounted camera and distance sensor, the speed of vehicle is calculated by observing the motion of feature point in image.Compared with the existing speed detection method based on physical sensor, the method is lower in cost, but the principle is simple, and the accuracy is also higher.The method can be used alone, can also be combined with the data of other sensors, improve detection accuracy and stability.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving and relates to a vehicle speed detection method based on epipolar images. Background Technology

[0002] In the measurement or reconstruction of the real world using sensors, determining the sensor's own velocity is a crucial issue. Macroscopic velocity determination can utilize external sensing devices such as GPS and inertial (gyroscope) sensors, while microscopic velocity determination can rely on velocity or accelerometer sensors; for example, a vehicle speed pulse sensor can be used in automobiles. However, GPS has several limitations. Firstly, its use is heavily restricted by factors such as obstruction from tall buildings and reflections, and satellite loss due to tunnels. Secondly, there are few intermediate sensors connecting macroscopic and microscopic data, and few methods for improving macroscopic and microscopic data. The vehicle speed detection method based on epipolar images proposed in this paper is relatively simple in algorithm implementation and can effectively avoid the leakage of sensitive information such as user location data by GPS, while also remaining unaffected by signal interference in remote areas. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a vehicle speed detection method based on an epipolar image.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A vehicle speed detection method based on epipolar images, the method comprising:

[0006] S1: A distance sensor that performs repeated route scanning is mounted on the vehicle to periodically capture images of the road ahead of the vehicle and obtain spatiotemporal distance images;

[0007] S2: Divide the obtained spatiotemporal distance image into several parts;

[0008] S3: The curves corresponding to each part are obtained, representing how each object changes over time within the sensor's field of view, and representing the vehicle's position and speed changes as detected by the sensor.

[0009] S4: Express the curves corresponding to each part in S3 using analytical expressions to represent the positional changes of each local short interval; obtain the velocity curves of each part by analytical differentiation;

[0010] S5: Generation of the full-range speed curve. The vehicle speed is calculated based on the displacement of the feature points. The speed curves obtained from each part are products of the local intervals in that part. The speed curves are regressed to generate a speed curve that connects the various speed curves and spans the entire interval. The speed curve closer to the middle part has higher reliability. If multiple speed curves are obtained in a specific interval, the speed curve with higher reliability is given priority.

[0011] Optionally, in step S1, the vehicle equipped with the distance sensor travels horizontally to obtain a spatiotemporal image, and delineates the edges formed by consecutive corresponding points in the right-hand image, which contain depth information. The points constituting the epipolar distance image EPDI are processed using the depth information for edge detection.

[0012]

[0013] In the formula, m is the slope of the side; Δx represents the horizontal movement distance of the image from the sensor; Δy represents the distance the feature moves in the image plane, i.e., the change in the image in each scan frame; k is the interval between the scanning lines, with the standard unit being meters. -1 Δt represents the scan time; F0 is the frame rate of the line scan sensor, with standard units of Hz or s. -1 V is the moving speed of the sensor;

[0014] The depth D and disparity u are correlated using the following method based on the distance traveled:

[0015]

[0016]

[0017] In the formula, ΔU represents the parallax change at different positions; u1 and u2 represent the parallax of the image distance sensor at positions 1 and 2, respectively; ΔX represents the horizontal movement distance of the image distance sensor; X represents the horizontal distance between the reference point P and the image sensor at position 1; and h represents the vertical distance between the camera plane and the image plane.

[0018] Combining equations (1) and (3), the velocity is obtained as:

[0019]

[0020] Optionally, in step S5, since the acquired spatiotemporal distance image data contains multiple cluster planes, these cluster planes are designated as S1, S2, ..., S... n This means that if S n Part from the kth n Frame start to k' n If it exists within the range up to frame k, the estimated velocity value R for frame k is...n The credibility of (k) is defined as follows:

[0021]

[0022]

[0023] 6σ=k' n -k n (7)

[0024] In the formula, k n k' represents the frame preceding the k-th frame. n σ represents the frame following the k-th frame; μ represents the average value of the frames before and after the k-th frame; σ represents the scaling factor of the difference between the frame following the k-th frame and the frame before the k-th frame.

[0025] From S n The obtained velocity curve is regarded as V n If (k), the estimated velocity curve V(k) spanning the entire interval is represented as follows:

[0026]

[0027]

[0028] In the formula, n is the Sth... n The set of frames in a cluster plane; R is the set of frames in the Sth cluster plane; n The sum of the confidence levels of the velocity estimates for each frame in each cluster plane;

[0029] Connect the speed curves corresponding to each part of their respective trust levels.

[0030] The advantages of this invention are as follows: Based on the principles of visual kinematics, this invention utilizes images acquired by an onboard camera and distance sensor to calculate the vehicle's speed by observing the movement of feature points in the images. Compared to existing speed detection methods based on physical sensors, this method is lower in cost, simpler in principle, and more accurate. This method can be used alone or combined with data from other sensors to improve detection accuracy and stability. It can be applied to various vehicle speed detection and driver assistance systems, and is particularly suitable for low-speed applications, enabling intelligent vehicle speed detection.

[0031] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0032] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0033] Figure 1 This is a schematic diagram of the formation of the outer polar surface image;

[0034] Figure 2 This is a depth-parallax relationship diagram;

[0035] Figure 3 This is a flowchart for speed estimation processing. Detailed Implementation

[0036] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0037] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0038] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0039] This method, based on visual kinematics, utilizes images acquired by an onboard camera and distance sensor to calculate vehicle speed by observing the movement of feature points within the images. Compared to existing speed detection methods based on physical sensors, this method is lower in cost, simpler in principle, and more accurate. It can be used alone or combined with data from other sensors to improve detection accuracy and stability. It can be applied to various vehicle speed detection and driver assistance systems, and is particularly suitable for low-speed applications, enabling intelligent vehicle speed detection.

[0040] The principle of the vehicle speed detection method based on epipolar images is as follows: Figure 1 As shown.

[0041] exist Figure 1 In the image on the left, a vehicle equipped with a distance sensor is traveling horizontally. The resulting spatiotemporal image depicts the edges formed by the continuous corresponding points that make up the image on the right. It is an image containing depth information. Although the points that make up the epipolar distance image (EPDI) are in the same coordinate system, they form multiple planes. The depth information can be directly used for edge detection.

[0042] in:

[0043]

[0044] In the formula, m is the slope of the side; Δx represents the horizontal distance the image moves from the sensor, which is... Figure 1 The change in the horizontal axis; Δy represents the distance the feature moves in the image plane, i.e., the change in the image in each scan frame; k is the interval between the scan lines, with the standard unit being meters (m). -1 Δt represents the scan time; F0 is the frame rate of the line scan sensor, with standard units of Hz or s. -1 V is the moving speed of the sensor.

[0045] like Figure 2 As shown, the depth D and disparity u are correlated using the following method based on the distance traveled:

[0046]

[0047]

[0048] In the formula, ΔU represents the change in disparity at different positions; u1 and u2 represent the disparity of the image distance sensor at positions 1 and 2, respectively; ΔX represents the horizontal movement distance of the image distance sensor; X represents the reference point P (e.g., ...). Figure 2 (as shown) is the horizontal distance between the camera plane and the image sensor at position 1; h represents the vertical distance between the camera plane and the image plane.

[0049] Combining equations (1) and (3), the velocity is obtained as:

[0050]

[0051] The speed detection and calculation program proposed in this invention has the following steps:

[0052] (1) Data acquisition: In order to obtain a spatiotemporal distance image, a distance sensor that performs repeated route scanning needs to be mounted on the vehicle.

[0053] (2) Segmentation of the spatiotemporal distance image: Since the spatiotemporal distance image is composed of multiple curve clusters, it is necessary to divide these curve clusters into several parts for the convenience of subsequent calculations.

[0054] (3) Corresponding curves of each part: The obtained parts represent the appearance of each object within the sensor's field of view over time, indicating the vehicle's position and speed changes as detected by the sensor.

[0055] (4) Calculation of the velocity curve for each part: The curves corresponding to each part in the previous step are expressed by analytical formulas, which can show the positional changes of each local short interval. The velocity curve of each part can be obtained by analytical differentiation.

[0056] (5) Generation of the full-range velocity curve: The velocity curves obtained from each part are products of the localized intervals within that part. Here, a velocity curve spanning the entire interval is generated by connecting the various velocity curves. This is obtained by regression processing of these velocity curves; therefore, the velocity curves near the center have high reliability, while those near the ends have low reliability. When multiple velocity curves can be obtained for a specific interval, the velocity curve with higher reliability should be given priority.

[0057] Specifically, if S n Part from K n Frame start to K' n If the velocity estimate for the Kth frame exists within the range up to the previous frame, the reliability of the estimate is defined as follows:

[0058]

[0059]

[0060] 6σ=k' n -k n (7)

[0061] In the formula, k n k' represents the frame preceding the k-th frame. nσ represents the frame following the k-th frame; μ represents the average value of the frames before and after the k-th frame; σ represents the scaling factor of the difference between the frame following the k-th frame and the frame before the k-th frame.

[0062] From S n The obtained velocity curve is regarded as V n If (k), the estimated velocity curve V(k) spanning the entire interval is represented as follows:

[0063]

[0064]

[0065] N: The set containing the portion of the Kth frame.

[0066] Ultimately, the speed curves corresponding to each part of their respective trust levels can be smoothly connected.

[0067] Figure 3 This is a flowchart for speed estimation processing.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A vehicle speed detection method based on epipolar images, characterized in that: The method includes: S1: A distance sensor that performs repeated route scanning is mounted on the vehicle to periodically capture images of the road ahead of the vehicle and obtain spatiotemporal distance images; S2: Divide the obtained spatiotemporal distance image into several parts; S3: The curves corresponding to each part are obtained, representing how each object changes over time within the sensor's field of view, and representing the vehicle's position and speed changes as detected by the sensor. S4: Express the curves corresponding to each part in S3 using analytical expressions to represent the positional changes of each local short interval; obtain the velocity curves of each part by analytical differentiation; S5: Generation of the full-range speed curve. The vehicle speed is calculated based on the displacement of the feature points. The speed curves obtained from each part are products of the local intervals in that part. The speed curves are regressed to generate a speed curve that connects the various speed curves and spans the entire interval. The speed curve closer to the middle part has higher reliability. If multiple speed curves are obtained in a specific interval, the speed curve with higher reliability is given priority. In step S1, the vehicle equipped with the distance sensor travels horizontally to obtain a spatiotemporal image. The edges formed by consecutive corresponding points in the spatiotemporal image are then depicted, containing depth information. The depth information is used to detect the edges of the points constituting the epipolar distance image EPDI. (1) In the formula, It is the slope of the edge; This indicates the horizontal distance the image has moved from the sensor; It represents the distance a feature moves in the image plane, that is, the amount of change in the image in each scan frame; It refers to the spacing between the scan lines, a standard unit. ; Indicates the scan time; It is the frame rate of the line scan sensor, and the standard unit is... or ; It is the moving speed of the sensor; Depth D and parallax The following method is used to associate data based on the distance traveled: (2) (3) In the formula, This represents the amount of parallax variation at different locations; and These represent the parallax of the image distance sensor at positions 1 and 2, respectively; This indicates the horizontal distance the image has moved from the sensor; This represents the horizontal distance between reference point P and the image sensor at position 1; This represents the perpendicular distance between the camera plane and the image plane; Combining equations (1) and (3), the velocity is obtained as: (4) In step S5, since the acquired spatiotemporal distance image data contains multiple cluster planes, these cluster planes are used... It means that if Part from the first Frame start to If it exists within the range up to the frame, then the first Estimated frame rate Credibility is defined in the following way: (5) (6) (7) In the formula, Indicates the first The previous frame; Indicates the first The frame following the previous frame; Indicates the first The average value of the frame before and after it; Indicates the first The scaling factor of the difference between the next frame and the previous frame; From The obtained velocity curve is regarded as If so, the estimated velocity curve spans the entire interval Represented in the following way: (8) (9) In the formula, For the first The set of frames in a cluster plane; For the first The sum of the confidence levels of the velocity estimates for each frame in each cluster plane; Connect the speed curves corresponding to each part of their respective trust levels.

Images