Vehicle speed estimation method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a vehicle speed estimation method and device, electronic equipment and a storage medium, and relates to the technical field of data processing, and the method comprises the steps: obtaining a range Doppler (RD) image generated by a radar carried by a vehicle, the RD image is an image obtained after the radar performs fast Fourier transform on detected echo data in a distance dimension and a speed dimension, the transverse direction of the RD image corresponds to a distance index, and the longitudinal direction of the RD image corresponds to a Doppler index; extracting amplitude edge features of the RD image; based on the amplitude edge features, a target Doppler index corresponding to a target position in the RD image is determined, and the target position is a position representing that the intensity of echoes received by the radar in the RD image is suddenly changed; and estimating the driving speed of the vehicle according to the target Doppler index and the Doppler resolution of the radar. By applying the technical scheme provided by the embodiment of the invention, the accuracy of vehicle speed estimation can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a vehicle speed estimation method, device, electronic device and storage medium. Background Art

[0002] As autonomous driving technology matures, many vehicles have L2 or even L3 autonomous driving functions. The vehicle speed is a crucial parameter in vehicle obstacle avoidance, alarm, braking and other functions.

[0003] At present, inertial sensors and millimeter-wave radars can be used to estimate the speed of the vehicle. The vehicle uses an inertial sensor to measure the acceleration to obtain the vehicle speed. However, due to the inaccuracy of the acceleration measured during the vehicle startup phase and the rapid acceleration and deceleration phase, the final measured vehicle speed is not accurate. In the traditional millimeter-wave radar vehicle speed estimation method, an equation is constructed based on the relationship or variant of the Doppler value measured by the on-board millimeter-wave radar, the azimuth of the target relative to the radar, and the vehicle's own driving speed, and then the equation is solved to obtain the vehicle speed. The above method relies on two measurement information, the azimuth and the Doppler value. The measurement errors of the two will affect the final solved vehicle speed, resulting in a large error in the estimated vehicle speed and low accuracy. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a vehicle speed estimation method, device, electronic device and storage medium to improve the accuracy of vehicle speed estimation. The specific technical solution is as follows:

[0005] In a first aspect of an embodiment of the present application, a vehicle speed estimation method is provided, comprising:

[0006] Acquire a RD (Range Doppler) image generated by a radar carried by the vehicle, wherein the RD image is an image obtained by performing fast Fourier transform on the detected echo data in the range dimension and the velocity dimension respectively, the lateral direction of the RD image corresponds to the range index, and the longitudinal direction of the RD image corresponds to the Doppler index;

[0007] Extracting amplitude edge features of the RD image;

[0008] Based on the amplitude edge feature, determining a target Doppler index corresponding to a target position in the RD image, wherein the target position is: a position in the RD image where a sudden change in the intensity of the echo received by the radar occurs;

[0009] The driving speed of the vehicle is estimated according to the target Doppler index and the Doppler resolution of the radar.

[0010] In a possible implementation manner, extracting the amplitude edge feature of the RD image includes:

[0011] Extracting the amplitude edge feature of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel;

[0012] For each pixel, if the absolute value of the first amplitude edge characterization value of the pixel is greater than a first preset threshold, the first amplitude edge characterization value of the pixel is updated to a first value; if the absolute value of the first amplitude edge characterization value of the pixel is less than or equal to the first preset threshold, the first amplitude edge characterization value of the pixel is updated to a second value.

[0013] In a possible implementation manner, extracting the amplitude edge feature of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel includes:

[0014] Filling the border of the RD image, wherein the number of filled pixel rows and / or pixel columns is determined according to the size of the first convolution kernel;

[0015] The first convolution kernel is used to perform convolution transformation on the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel.

[0016] In a possible implementation manner, determining the target Doppler index corresponding to the target position in the RD image based on the amplitude edge feature includes:

[0017] Accumulating first amplitude edge representation values ​​of pixels corresponding to each Doppler index in the RD image respectively to obtain first accumulated values ​​corresponding to each Doppler index;

[0018] Determine the maximum maximum value and the second maximum maximum value in the first accumulated value corresponding to each Doppler index arranged according to the size of the index value, and use the Doppler indexes corresponding to the maximum maximum value and the second maximum maximum value as the target Doppler index.

[0019] In a possible implementation manner, estimating the vehicle's speed according to the target Doppler index and the Doppler resolution of the radar includes:

[0020] Determine a Doppler index corresponding to the smallest first accumulated value among candidate Doppler indexes, wherein the candidate Doppler indexes are all Doppler indexes between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second largest maximum value;

[0021] The product of the determined Doppler index and the Doppler resolution of the radar is calculated, and the obtained product is used as the driving speed of the vehicle.

[0022] In a possible implementation manner, extracting the amplitude edge feature of the RD image includes:

[0023] Accumulating pixel values ​​of pixels corresponding to each Doppler index in the RD image respectively to obtain second accumulated values ​​corresponding to each Doppler index;

[0024] The amplitude edge features of the second accumulated values ​​corresponding to the Doppler indexes arranged according to the size of the index values ​​are extracted to obtain the second amplitude edge representation values ​​corresponding to the Doppler indexes.

[0025] In a possible implementation manner, extracting the amplitude edge feature of the second accumulated value corresponding to each Doppler index arranged according to the size of the index value to obtain the second amplitude edge representation value corresponding to each Doppler index includes:

[0026] Using the second convolution kernel, convolution transformation is performed one by one on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, so as to obtain the second amplitude edge representation value corresponding to each Doppler index.

[0027] In a possible implementation manner, determining the target Doppler index corresponding to the target position in the RD image based on the amplitude edge feature includes:

[0028] Determine a maximum value among the second amplitude edge representation values, and use a Doppler index corresponding to the determined maximum value as a target Doppler index;

[0029] The estimating the vehicle's speed according to the target Doppler index and the Doppler resolution of the radar comprises:

[0030] The product of the target Doppler index and the Doppler resolution of the radar is calculated, and the obtained product is used as the driving speed of the vehicle.

[0031] In a second aspect of an embodiment of the present application, a vehicle speed estimation device is provided, comprising:

[0032] an acquisition module, configured to acquire a range Doppler RD image generated by a radar carried by the vehicle, wherein the RD image is an image obtained by performing fast Fourier transform on the detected echo data in the distance dimension and the speed dimension respectively, the lateral direction of the RD image corresponds to the distance index, and the longitudinal direction of the RD image corresponds to the Doppler index;

[0033] An extraction module, used for extracting amplitude edge features of the RD image;

[0034] A determination module, configured to determine a target Doppler index corresponding to a target position in the RD image based on the amplitude edge feature, wherein the target position is a position in the RD image where a sudden change in the intensity of the echo received by the radar occurs;

[0035] The estimation module is used to estimate the vehicle's travel speed according to the target Doppler index and the Doppler resolution of the radar.

[0036] In a possible implementation manner, the extraction module is specifically used to: extract the amplitude edge features of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel; for each pixel, if the absolute value of the first amplitude edge representation value of the pixel is greater than a first preset threshold, update the first amplitude edge representation value of the pixel to a first value; if the absolute value of the first amplitude edge representation value of the pixel is less than or equal to the first preset threshold, update the first amplitude edge representation value of the pixel to a second value.

[0037] In a possible implementation, the extraction module is specifically used to: fill the boundaries of the RD image, wherein the number of filled pixel rows and / or pixel columns is determined according to the size of the first convolution kernel; and perform a convolution transformation on the RD image pixel by pixel along the distance dimension of the RD image using the first convolution kernel to obtain a first amplitude edge representation value of each pixel.

[0038] In a possible implementation manner, the determination module is specifically used to: respectively accumulate the first amplitude edge representation values ​​of the pixels corresponding to each Doppler index in the RD image to obtain the first accumulated values ​​corresponding to each Doppler index; determine the maximum maximum value and the second maximum value among the first accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, and use the Doppler indexes corresponding to the maximum maximum value and the second maximum value as the target Doppler index.

[0039] In a possible implementation manner, the estimation module is specifically used to: determine the Doppler index corresponding to the smallest first accumulated value among the candidate Doppler indexes, wherein the candidate Doppler indexes are all Doppler indexes between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second largest maximum value; calculate the product of the determined Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

[0040] In a possible implementation, the extraction module is specifically used to: accumulate pixel values ​​of pixels corresponding to each Doppler index in the RD image respectively to obtain second accumulated values ​​corresponding to each Doppler index; extract amplitude edge features of the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value to obtain second amplitude edge representation values ​​corresponding to each Doppler index.

[0041] In a possible implementation, the extraction module is specifically used to: use a second convolution kernel to perform convolution transformation on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value one by one, so as to obtain a second amplitude edge representation value corresponding to each Doppler index.

[0042] In a possible implementation, the determination module is specifically used to: determine the maximum value among the second amplitude edge characterization values, and use the Doppler index corresponding to the determined maximum value as the target Doppler index; the estimation module is specifically used to: calculate the product of the target Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

[0043] In a third aspect of the embodiments of the present application, an electronic device is provided, including:

[0044] Memory, used to store computer programs;

[0045] The processor is used to implement the steps of any method described in the first aspect when executing the program stored in the memory.

[0046] In a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the steps of any method described in the first aspect are implemented.

[0047] In a fifth aspect of the embodiments of the present application, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the steps of any one of the methods described in the first aspect above.

[0048] Beneficial effects of the embodiments of the present application:

[0049] In the technical solution provided in the embodiment of the present application, the target Doppler index corresponding to the target position where the echo intensity received by the radar changes suddenly is determined by extracting the amplitude edge features in the RD image generated by the radar, and then the vehicle's driving speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0050] The relationship between the Doppler d, azimuth angle θ and vehicle speed V of a stationary target is: d = V × cosθ. The azimuth angle of a distant stationary object relative to the radar is relatively small and close to 0°, so d≈V can be obtained, that is, the Doppler value of a distant stationary object measured by the radar is approximately equal to the vehicle speed. In addition, the echo intensity reflected by a stationary object is much greater than the ground clutter or noise, so it is manifested on the RD image as a brighter strip along the distance dimension. In the above scheme, by extracting the amplitude edge features in the RD image, the brighter strip formed by the echo reflected by the distant stationary object along the distance dimension can be found, and then the target Doppler index corresponding to the target position where the echo intensity reflected by the distant stationary object changes suddenly can be determined, and then the vehicle's speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0051] In the above scheme, the vehicle's speed can be estimated by relying on only one measurement information, namely the Doppler value corresponding to the target Doppler index in the RD image, which reduces the amount of measurement information used for vehicle speed estimation, and also reduces the impact of errors of multiple measurement information on vehicle speed estimation, thereby improving the accuracy of vehicle speed estimation.

[0052] Of course, implementing any product or method of the present application does not necessarily require achieving all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other embodiments can also be obtained based on these drawings.

[0054] Figure 1 A first schematic diagram of an RD image provided in an embodiment of the present application;

[0055] Figure 2 A schematic diagram of the geometric relationship between the radar and the stationary object provided in the embodiment of the present application;

[0056] Figure 3 A schematic diagram of a first flow chart of a vehicle speed estimation method provided in an embodiment of the present application;

[0057] Figure 4 A first detailed flow chart of step S302 provided in an embodiment of the present application;

[0058] Figure 5 A second schematic diagram of an RD image provided in an embodiment of the present application;

[0059] Figure 6 A detailed flow chart of step S401 provided in an embodiment of the present application;

[0060] Figure 7 A schematic diagram of a filled RD image provided in an embodiment of the present application;

[0061] Figure 8 A detailed flow chart of step S303 provided in an embodiment of the present application;

[0062] Fig. 9 A detailed flow chart of step S304 provided in an embodiment of the present application;

[0063] Fig.10 A second flow chart of the vehicle speed estimation method provided in an embodiment of the present application;

[0064] Fig.11 A third schematic diagram of an RD image provided in an embodiment of the present application;

[0065] Fig.12 A schematic diagram of a first amplitude edge representation value of each pixel in an example provided in an embodiment of the present application;

[0066] Fig.13 A binary image obtained by binarizing the first amplitude edge representation value of each pixel in an example provided in an embodiment of the present application;

[0067] Fig.14 A schematic diagram of a first accumulated value corresponding to each Doppler index in an example provided in an embodiment of the present application;

[0068] Fig.15 A second detailed flow diagram of step S302 provided in an embodiment of the present application;

[0069] Fig.16 A schematic diagram of the structure of a vehicle speed estimation device provided in an embodiment of the present application;

[0070] Fig.17 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field based on the present application belong to the scope of protection of the present application.

[0072] The following first describes some of the nouns mentioned in the embodiments of the present application.

[0073] RD image: RD image is an image obtained by performing FFT (Fast Fourier Transform) on the echo data detected by the radar in the distance dimension and the velocity dimension respectively. Figure 1 In the RD image shown, the horizontal axis is the distance index, that is, the horizontal direction of the RD image corresponds to the distance index, which represents the distance between the target detected by the radar and the radar; the vertical axis is the Doppler index, that is, the vertical direction of the RD image corresponds to the Doppler index, which represents the projection size of the speed of the target detected by the radar relative to the radar on the line connecting the target and the center of the radar. The pixel value of each pixel in the RD image represents the echo amplitude received by the radar, and the unit of the echo amplitude is dB (decibel). The larger the echo amplitude received by the radar, the larger the pixel value of the corresponding pixel in the RD image, which is reflected in the higher brightness of the pixel in the RD image.

[0074] Range resolution: When two targets are at the same azimuth relative to the radar but at different distances from the radar, the minimum distance at which the radar can distinguish the two is called the range resolution. The range resolution can be written as ΔR. The product of the range index and the range resolution is the distance between the radar and the measured target.

[0075] Velocity resolution: Velocity resolution is the minimum velocity interval that the radar can distinguish. Velocity resolution can be recorded as ΔV. Velocity resolution can also be called Doppler resolution or velocity measurement resolution. The product of Doppler index and velocity resolution is Doppler value. Doppler is a radial velocity. Doppler value is the projection of the velocity of the target relative to the radar on the line connecting the center of the target and the radar.

[0076] As autonomous driving technology matures, many vehicles have L2 or even L3 autonomous driving functions. The vehicle speed is a crucial parameter in vehicle obstacle avoidance, alarm, braking and other functions.

[0077] At present, inertial sensors and millimeter-wave radars can be used to estimate the speed of the vehicle. The vehicle uses an inertial sensor to measure the acceleration to obtain the vehicle speed. However, due to the inaccuracy of the acceleration measured during the vehicle startup phase and the rapid acceleration and deceleration phase, the final measured vehicle speed is not accurate. In the traditional millimeter-wave radar vehicle speed estimation method, an equation is constructed based on the relationship or variant of the Doppler value measured by the on-board millimeter-wave radar, the azimuth of the target relative to the radar, and the vehicle's own driving speed, and then the equation is solved to obtain the vehicle speed. The above method relies on two measurement information, the azimuth and the Doppler value. The measurement errors of the two will affect the final solved vehicle speed, resulting in a large error in the estimated vehicle speed and low accuracy.

[0078] In order to solve the above problems, an embodiment of the present application provides a vehicle speed estimation method to improve the accuracy of vehicle speed estimation.

[0079] In order to more clearly illustrate the vehicle speed estimation method provided in the embodiment of the present application, the embodiment of the present application provides the following Figure 2 The geometric relationship diagram shown first describes the implementation idea of ​​the vehicle speed estimation method provided in the embodiment of the present application.

[0080] like Figure 2 As shown, the vehicle is moving forward at a speed of V. There is a stationary object at an azimuth angle of θ in the left front. The radar carried by the vehicle can measure the Doppler value d based on the echo reflected by the stationary object, that is, the projection of the speed of the stationary object relative to the radar carried by the vehicle on the line connecting the stationary object and the center of the radar is d. Based on the mathematical relationship between the Doppler value d measured by the radar, the azimuth angle θ between the radar and the stationary object, and the vehicle speed V, the formula can be obtained: d=V×cos(θ). The farther the stationary object in front of the radar is from the radar, the closer the azimuth angle is to 0, that is, θ≈0°. According to the above formula, d≈V×cos(0°) can be obtained, that is, d≈V, that is, the Doppler value d measured based on the echo reflected by the distant stationary object in front of the radar can be approximately equal to the vehicle speed V. Therefore, the vehicle speed V can be estimated based on the Doppler value d measured by the echo reflected by the distant stationary object in front of the radar.

[0081] In addition, since the echo intensity reflected by a stationary object is much greater than the ground clutter and noise, that is, the signal-to-noise ratio of a stationary object is much greater than the noise, after the radar receives the echo reflected by the stationary object, the amplitude of the corresponding pixel of the stationary object on the generated RD image will be larger, that is, the pixel value will be larger, which is reflected in the RD image as the color of the corresponding pixel will be brighter. Among them, the echo intensity represents the energy of the radar receiving the target reflected echo. The echo intensity is proportional to the echo amplitude. The greater the echo intensity, the greater the echo amplitude. And because the Doppler value measured based on the reflected echo from the distant stationary object directly in front of the radar is close to the vehicle speed, in the RD image generated after the radar receives the echo reflected from the distant stationary object directly in front, the amplitude of the pixel corresponding to the distant stationary object is much larger than the ground clutter and noise, that is, the pixel value of the pixel corresponding to the distant stationary object is much larger than the pixel value of the pixel corresponding to the ground clutter and noise. This is reflected in the RD image as the pixel corresponding to the distant stationary object is brighter than the pixel corresponding to the ground clutter and noise. Thus, a bright stripe is formed along the distance dimension on the generated RD image, and the Doppler value corresponding to the distant stationary object directly in front of the radar falls within the range of the bright stripe.

[0082] According to the above implementation ideas, the Doppler value within the highlighted strip range in the RD image can be used to estimate the vehicle's speed.

[0083] The vehicle speed estimation method provided in the embodiment of the present application can be applied to electronic devices such as desktop computers, laptop computers, mobile phones or servers. For ease of description, the execution subject of the vehicle speed estimation method provided in the embodiment of the present application is collectively referred to as an electronic device.

[0084] like Figure 3 As shown, the vehicle speed estimation method provided in the embodiment of the present application includes the following steps S301-S304.

[0085] S301, acquiring an RD image generated by a radar mounted on the vehicle.

[0086] The above radar may be a millimeter wave radar. The radar may be installed at the bumper directly in front of the vehicle. After the radar transmits a detection wave and receives an echo reflected by an object, an RD image may be generated. The electronic device may actively obtain the RD image generated by the radar according to a preset period; the electronic device may also receive the RD image reported by the radar itself.

[0087] S302, extracting amplitude edge features of the RD image.

[0088] The amplitude edge feature is the edge along the distance dimension on the RD image, representing the location where the pixel value of the pixel in the RD image changes greatly.

[0089] The electronic device can perform edge detection on the RD image using an edge detection algorithm, and extract the edge along the distance dimension on the RD image as the amplitude edge feature of the RD image. The edge detection algorithm can be a Prewitt (edge ​​detection of a first-order differential operator) algorithm. Figure 1 The RD image shown is horizontal along the distance dimension. The electronic device performs edge detection on the RD image to extract the horizontal edges, and uses the extracted horizontal edges as amplitude edge features. The obtained amplitude edge features include the upper and lower boundaries of the rectangular frame 101, and other horizontal edges within the rectangular frame 101 (not shown in the figure).

[0090] S303, determining a target Doppler index corresponding to the target position in the RD image based on the amplitude edge feature.

[0091] The target position is the position where the intensity of the echo received by the radar in the RD image changes suddenly.

[0092] The pixel values ​​of the pixels on both sides of the amplitude edge feature are quite different, that is, the position of the amplitude edge feature in the RD image is the position where the intensity of the echo received by the radar changes. The electronic device can determine the Doppler index corresponding to the position of one or more amplitude edge features with the largest amplitude change from all the amplitude edge features as the target Doppler index. For example, Figure 1As shown, the amplitude edge features extracted by the electronic device include the upper and lower boundaries of the rectangular frame 101, and other lateral edges in the rectangular frame 101 (not shown in the figure). The electronic device can determine the Doppler indexes corresponding to the upper and lower boundaries of the rectangular frame 101 with the largest amplitude change as the target Doppler index.

[0093] S304, estimating the vehicle's travel speed according to the target Doppler index and the radar's Doppler resolution.

[0094] The specific implementation of step S304 can be found in the following text and will not be described in detail here.

[0095] In the technical solution provided in the embodiment of the present application, the target Doppler index corresponding to the target position where the echo intensity received by the radar changes suddenly is determined by extracting the amplitude edge features in the RD image generated by the radar, and then the vehicle's driving speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0096] The relationship between the Doppler d, azimuth angle θ and vehicle speed V of a stationary target is: d = V × cosθ. The azimuth angle of a distant stationary object relative to the radar is relatively small and close to 0°, so d≈V can be obtained, that is, the Doppler value of a distant stationary object measured by the radar is approximately equal to the vehicle speed. In addition, the echo intensity reflected by a stationary object is much greater than the ground clutter or noise, so it is manifested on the RD image as a brighter strip along the distance dimension. In the above scheme, by extracting the amplitude edge features in the RD image, the brighter strip formed by the echo reflected by the distant stationary object along the distance dimension can be found, and then the target Doppler index corresponding to the target position where the echo intensity reflected by the distant stationary object changes suddenly can be determined, and then the vehicle's speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0097] In the above scheme, the vehicle's speed can be estimated by relying on only one measurement information, namely the Doppler value corresponding to the target Doppler index in the RD image, which reduces the amount of measurement information used for vehicle speed estimation, and also reduces the impact of errors of multiple measurement information on vehicle speed estimation, thereby improving the accuracy of vehicle speed estimation.

[0098] The present application also provides a method for extracting amplitude edge features, such as Figure 4 A detailed flow chart of step S302 is shown, and the extraction method may include the following steps S401-S404.

[0099] S401 , extracting the amplitude edge feature of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel.

[0100] The first amplitude edge characterization value of a pixel characterizes the amplitude variation of pixels adjacent to the pixel along the Doppler dimension.

[0101] The electronic device may apply an edge detection operator to each pixel one by one along the distance dimension of the RD image, thereby obtaining a first amplitude edge representation value of each pixel. The edge detection operator may be a Prewitt operator.

[0102] For example, Figure 5 In the RD image shown, the Doppler index can be recorded as n, n = 1, 2, ..., N. Wherein, N is the maximum value of the Doppler index, and its value is 64. The row index of a row of pixels with a Doppler index of n can be recorded as dn. The distance index can be recorded as m, m = 1, 2, ..., M. Wherein, M is the maximum value of the distance index, and its value is 200. The column index of a column of pixels with a distance index of m can be recorded as rm. Figure 5 The pixel value of each pixel in the RD image shown can be expressed as:

[0103] RD(n,m);

[0104] Among them, RD(n,m) represents Figure 5 The pixel value of the pixel in row dn and column rm.

[0105] The electronic device extracts the first amplitude edge representation values ​​of the pixels in the columns r1, r2, ..., r200 from left to right pixel by pixel along the rows d1, d2, ..., d64, respectively. The first amplitude edge representation value of each pixel can be expressed as:

[0106] A(n,m);

[0107] Where n = 1, 2, ..., N; m = 1, 2, ..., M; A(n, m) means Figure 5 The first amplitude edge representation value of the pixel in row dn and column rm.

[0108] S402, for each pixel, determine whether the absolute value of the first amplitude edge representation value of the pixel is greater than a first preset threshold; if so, execute step S403, if not, execute step S404.

[0109] For each pixel, the greater the difference in pixel values ​​between adjacent pixels along the Doppler dimension, the greater the absolute value of the first amplitude edge representation value obtained. In order to determine edge points where the amplitude changes of adjacent pixels are large, the electronic device can determine whether the absolute value of the first amplitude edge representation value of each pixel is greater than a first preset threshold, thereby determining that the pixel whose absolute value of the first amplitude edge representation value is greater than the first preset threshold is an edge point.

[0110] When the absolute value of the first amplitude edge characterization value is greater than the first preset threshold, it means that the pixel values ​​of the pixels adjacent to the pixel along the Doppler dimension are relatively different, and the amplitudes of the pixels adjacent to the pixel change significantly. The electronic device can use the pixel as an edge point and execute step S403 to update the first amplitude edge characterization value of the pixel to a first value. The first value indicates that the pixel is an edge point constituting an amplitude edge feature. The first value can be set to 1.

[0111] When the absolute value of the first amplitude edge characterization value is less than or equal to the first preset threshold, it means that the difference between the pixel values ​​of the adjacent pixels of the pixel along the Doppler dimension is small, and the amplitudes of the adjacent pixels of the pixel are similar. The electronic device can determine that the pixel is not an edge point, and execute step S404 to update the first amplitude edge characterization value of the pixel to a second value. The second value indicates that the pixel is not an edge point. The second value can be set to 0.

[0112] Continuing with the above example, the electronic device extracts the first amplitude edge representation value of each pixel as follows:

[0113] A(n,m);

[0114] Where n = 1, 2, ..., N; m = 1, 2, ..., M; A(n, m) means Figure 5 The first amplitude edge representation value of the pixel in row dn and column rm.

[0115] The first preset threshold is thr, thr=25dB, the first value is 1, the second value is 0, and the electronic device can update the first amplitude edge characterization value according to the following expression:

[0116]

[0117] Where n = 1, 2, ..., N; m = 1, 2, ..., M; B (n, m) means Figure 5 The first amplitude edge representation value after the pixel in row dn and column rm is updated. |A(n,m)| represents the absolute value of A(n,m).

[0118] S403: Update the first amplitude edge representation value of the pixel to a first value.

[0119] S404: Update the first amplitude edge representation value of the pixel to a second value.

[0120] In the technical solution provided by the embodiment of the present application, after obtaining the first amplitude edge characterization value of each pixel, the electronic device determines the edge point with a large amplitude change of adjacent pixels along the Doppler dimension according to whether the absolute value of the first amplitude edge characterization value is greater than the first preset threshold value, and updates the first amplitude edge characterization value of the edge point to the first value, and updates the first amplitude edge characterization value of the non-edge point to the second value. In this way, the edge points with large amplitude changes of adjacent pixels can be more clearly distinguished, and subsequently based on the updated first amplitude edge characterization value, the target position where the echo intensity suddenly changes and the corresponding target Doppler index can be determined more quickly, thereby improving the efficiency of vehicle speed estimation based on the target Doppler index.

[0121] The embodiment of the present application also provides a method for obtaining a first amplitude edge representation value of each pixel by using convolution transformation, such as Figure 6 A detailed flow chart of step S401 is shown, and the method includes the following steps S601-S602.

[0122] S601, performing border filling on the RD image.

[0123] The number of filled pixel rows and / or pixel columns is determined according to the size of the first convolution kernel.

[0124] The first convolution kernel can be used to detect the change in pixel value of pixels adjacent to each pixel in the RD image along the Doppler dimension. The size of the first convolution kernel can be 3×3, 5×5, 7×7, etc., and there is no limitation on this, as long as the edge in the image can be detected. When the electronic device fills the border of the RD image, it is necessary to determine the number of filled pixel rows and / or pixel columns according to the size of the first convolution kernel to ensure that each pixel in the RD image can be convolved using the first convolution kernel. For example, when the size of the first convolution kernel is 3×3, the electronic device needs to fill 1 row of pixel rows on the upper and lower sides of the RD image, and 1 row of pixel columns on the left and right sides of the RD image, to ensure that the pixels at the boundary of the RD image can also be convolved using the first convolution kernel.

[0125] In addition, when filling the border of the RD image, for the Doppler dimension, the pixel values ​​of the last row in the original RD image can be filled before the first row, and the pixel values ​​of the first row in the original RD image can be filled after the last row. If multiple rows of pixels need to be filled, the same method can be used to fill the pixel values ​​of the last rows before the first row, and the pixel values ​​of the first rows after the last row. For the distance dimension, both sides are filled with 0. Figure 5For the RD image shown in FIG. 1 , when the electronic device performs border filling on the RD image, the pixel value of row d64 may be filled before row d1 of the RD image, the pixel value of row d1 may be filled after row d64 of the RD image, and a column of 0 may be filled on the left side of column r1 and on the right side of column r200 of the RD image. The filled RD image is as shown in FIG. Figure 7 shown.

[0126] S602 : along the distance dimension of the RD image, use a first convolution kernel to perform convolution transformation on the RD image pixel by pixel to obtain a first amplitude edge representation value of each pixel.

[0127] The electronic device can use the first convolution kernel to place each pixel at the center of the first convolution kernel along the distance dimension of the RD image after boundary filling, perform convolution calculations on the first convolution kernel with each pixel and the values ​​adjacent to the pixel pixel by pixel, and use the calculation result as the first amplitude edge representation value of the pixel at the center position of the convolution calculation.

[0128] For example, the first convolution kernel 3×3 for:

[0129]

[0130] For ease of description, Figure 7 The first row is labeled d0, the last row is labeled d65, the first column is labeled r0, and the last column is labeled r201. Figure 7 The padded RD image shown can be expressed as:

[0131] RD1(n,m);

[0132] Where n = 0, 1, ..., N, N + 1; m = 0, 1, ..., M, M + 1; RD1 (n, m) means Figure 7 The pixel value of the pixel in row dn and column rm.

[0133] Use the first convolution kernel 3×3 The first amplitude edge representation value A(n,m) of each pixel obtained by performing convolution transformation on the RD image pixel by pixel is:

[0134]

[0135] Where n = 1, 2, ..., N; m = 1, 2, ..., M; i and j are kernel 3×3The row and column labels of each component in , i = -1, 0 or 1, j = -1, 0 or 1; Σ represents summation, and A(n,m) represents the first amplitude edge representation value of the pixel of dn rows and rm columns. When the electronic device performs convolution transformation on the pixel of dn rows and rm columns, the pixel value of the pixel is multiplied by the element at the center position of the first convolution kernel, the pixel value of the pixel adjacent to the pixel is multiplied by the element at the edge position of the first convolution kernel, and the obtained products are summed to obtain the first amplitude edge representation value of the pixel.

[0136] In the technical solution provided in the embodiment of the present application, the RD image is firstly border-filled, so as to ensure that the pixels at the border in the RD image can also be convolution-transformed, and then the first amplitude edge representation value of each pixel in the RD image can be obtained. Then, the amplitude edge feature is extracted by convolution transformation, which can effectively identify the edge in the image, and the calculation complexity is low, saving calculation resources.

[0137] The present application also provides a method for obtaining a target Doppler index. Figure 4 The provided method extracts the first amplitude edge characterization value of the RD image and uses it to determine the target Doppler index. Figure 8 A detailed flow chart of step S303 is shown, and the acquisition method includes the following steps S801-S802.

[0138] S801 , respectively accumulating first amplitude edge representation values ​​of pixels corresponding to respective Doppler indexes in the RD image to obtain first accumulated values ​​corresponding to respective Doppler indexes.

[0139] The electronic device may accumulate the first amplitude edge representation value of each pixel in sequence according to the distance index for each Doppler index to obtain a first accumulated value corresponding to each Doppler index.

[0140] For example, Figure 5 The first amplitude edge characterization value corresponding to the RD image shown is expressed as:

[0141] B(n,m);

[0142] Where n = 1, 2, ..., N; m = 1, 2, ..., M; B (n, m) means Figure 5 The first amplitude edge representation value after the pixel in row dn and column rm is updated.

[0143] For the Doppler index n, the electronic device can accumulate the first amplitude edge representation value of each pixel in sequence according to the distance index, and obtain the first accumulated value B sum (n) is:

[0144]

[0145] Where n = 1, 2, ..., N; m = 1, 2, ..., M; Σ represents summation, B sum (n) is the first accumulated value corresponding to the Doppler index n.

[0146] S802, determining the maximum maximum value and the second maximum value in the first accumulated values ​​corresponding to the Doppler indices arranged according to the index value size, and taking the Doppler indexes corresponding to the maximum maximum value and the second maximum value as the target Doppler index.

[0147] The electronic device may arrange the first accumulated values ​​in order according to the size of the index values. For example, the first accumulated value is B sum (n), n = 1, 2, ..., N; the electronic device adds the first accumulated value B sum (n) Arranged in order from small to large according to the index value: B sum (1) B sum (2), ..., B sum (N). Afterwards, the electronic device can perform function fitting for each Doppler index and the corresponding first accumulated value to obtain a function of the first accumulated value relative to the Doppler index, and then obtain the maximum maximum value and the second maximum value of the first accumulated value corresponding to each Doppler index by solving all maximum values ​​of the function.

[0148] The electronic device can also traverse the first accumulated values ​​of each Doppler index in turn. When the first accumulated value corresponding to a Doppler index is greater than the first accumulated values ​​of two adjacent Doppler indexes, the first accumulated value of the Doppler index is recorded as a maximum value, and then the maximum maximum value and the second largest maximum value are determined from all the recorded maximum values.

[0149] Since there is usually a certain offset in the measurement of RD images, the electronic device can first determine the Doppler indexes corresponding to the positions with the largest echo intensity mutations, namely the maximum maximum value and the second maximum value, and then determine the Doppler index used to estimate the vehicle speed based on the Doppler indexes corresponding to the maximum maximum value and the second maximum value, thereby reducing the impact of the offset, improving the accuracy of the Doppler index obtained for estimating the vehicle speed, and thus improving the accuracy of the vehicle speed estimation.

[0150] Based on the above Figure 8 After the target Doppler index is determined by the provided method, the embodiment of the present application can estimate the vehicle's driving speed by using the determined target Doppler index through the following method 1 or method 2.

[0151] Method 1: Select any Doppler index from the candidate Doppler indexes, and calculate the product of the selected Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

[0152] Among them, the candidate Doppler indices are all the Doppler indices between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second maximum maximum value. The electronic device can randomly select any Doppler index among the candidate Doppler indices, calculate the product of the selected Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

[0153] In this way, without additional calculations, the electronic device can reduce the offset of the Doppler value generated by the measurement of the RD image, improve the accuracy of the obtained Doppler index for estimating the vehicle speed, and further improve the accuracy of the vehicle speed estimation.

[0154] Method 2, a refined flowchart of step S304 as Fig. 9 shown, Method 2 may include the following steps S901 - S902.

[0155] S901, determine the Doppler index corresponding to the smallest first cumulative value among the candidate Doppler indices.

[0156] The electronic device can sequentially compare the first cumulative values corresponding to adjacent indices among the candidate Doppler indices, determine the smallest first cumulative value among the first cumulative values corresponding to the candidate Doppler indices, and further determine the Doppler index corresponding to this first cumulative value.

[0157] For example, the Doppler index corresponding to the second maximum maximum value is i, the Doppler index corresponding to the maximum maximum value is j, and i < j. The electronic device can first sequentially compare the first cumulative values corresponding to each index among the candidate Doppler indices to determine the smallest first cumulative value:

[0158]

[0159] where k = i, i + 1,..., j; min{} represents finding the minimum value of the multiple first cumulative values within the brackets, and B min is the smallest first cumulative value. After the electronic device determines the smallest first cumulative value, it can determine the Doppler index k1 corresponding to this first cumulative value.

[0160] S902, calculate the product of the determined Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

[0161] Continuing the above example, the electronic device can calculate the product of the determined Doppler index k1 and the resolution ΔV of the radar to obtain k1×ΔV = V, and the electronic device can use the obtained product V as the driving speed of the vehicle.

[0162] Since there is usually a certain offset in the measurement of RD images, the electronic device determines the Doppler index corresponding to the smallest first accumulated value among all the candidate Doppler indices between the Doppler indices corresponding to the maximum maximum value and the second largest maximum value for estimating the vehicle speed, thereby reducing the impact of the offset and improving the accuracy of the Doppler index obtained for estimating the vehicle speed, thereby improving the accuracy of the vehicle speed estimation.

[0163] For a clearer explanation based on the above Figure 3 , Figure 4 , Figure 6 , Figure 8 and Fig. 9 The vehicle speed estimation method implemented in this application also provides the following Fig.10 A complete embodiment is shown, which includes the following steps S1001-S1010.

[0164] S1001, obtaining an RD image generated by a radar mounted on a vehicle, corresponding to the above step S301.

[0165] S1002, filling the border of the RD image, corresponding to the above step S601.

[0166] S1003, along the distance dimension of the RD image, use the first convolution kernel to perform convolution transformation on the RD image pixel by pixel to obtain a first amplitude edge representation value of each pixel, which corresponds to the above step S602.

[0167] S1004, for each pixel, determine whether the absolute value of the first amplitude edge representation value of the pixel is greater than a first preset threshold; if so, execute step S1005, if not, execute step S1006, corresponding to the above step S402.

[0168] S1005: Update the first amplitude edge representation value of the pixel to a first value, corresponding to the above step S403.

[0169] S1006: Update the first amplitude edge representation value of the pixel to a second value, corresponding to the above step S404.

[0170] S1007, respectively accumulating the first amplitude edge representation values ​​of the pixels corresponding to each Doppler index in the RD image to obtain the first accumulated value corresponding to each Doppler index, which corresponds to the above step S801.

[0171] S1008, determining the maximum maximum value and the second maximum maximum value in the first accumulated value corresponding to each Doppler index arranged according to the index value size, and taking the Doppler index corresponding to the maximum maximum value and the second maximum maximum value as the target Doppler index, which corresponds to the above step S802.

[0172] S1009, determining the Doppler index corresponding to the smallest first accumulated value among the candidate Doppler indexes, corresponding to the above step S901.

[0173] S1010, calculating the product of the determined Doppler index and the Doppler resolution of the radar, and taking the obtained product as the driving speed of the vehicle, which corresponds to the above step S902.

[0174] Figure 11-Figure 14 For applications such as Fig.10 The illustrated embodiment performs one example of vehicle speed estimation.

[0175] Fig.11 is an original RD image. After the electronic device executes step S1001 to obtain the RD image generated by the radar carried by the vehicle, it executes steps S1002-S1003 to fill the RD image and perform convolution transformation on the filled RD image to obtain the first amplitude edge representation value of each pixel as shown in Fig.12 shown.

[0176] The electronic device then executes steps S1004-S1006 to update the first amplitude edge representation value of each pixel to the first value or the second value, that is, to perform binarization processing, and obtain a binary image of the first amplitude edge representation value of each pixel after processing. Fig.13 shown.

[0177] The electronic device then executes step S1007 to obtain the first accumulated value corresponding to each Doppler index. Fig.14 shown.

[0178] The electronic device then executes step S1008 to determine the maximum maximum value and the second maximum maximum value, such as Fig.14 The first accumulated values ​​corresponding to points B and A in the figure are respectively, wherein the first accumulated value corresponding to point A is the second largest maximum value, the Doppler index corresponding to point A is i, the first accumulated value corresponding to point B is the largest maximum value, and the Doppler index corresponding to point B is j.

[0179] The electronic device then executes step S1009 to determine the Doppler index with the smallest first accumulated value corresponding to all Doppler indices between i and j, such as Fig.14 The Doppler index corresponding to point C in , the first accumulated value corresponding to point C is the smallest among the first accumulated values ​​corresponding to all Doppler indices between i and j, and the Doppler index corresponding to point C is k1.

[0180] Afterwards, the electronic device executes step S1010 to calculate the product V of the determined Doppler index k1 and the Doppler resolution ΔV of the radar, and uses V as the driving speed of the vehicle.

[0181] Using the above Figure 4-Figure 14The vehicle speed estimation method implemented by the described content is: first extract the first amplitude edge characterization value as the amplitude edge feature from the RD image pixel by pixel, then accumulate and sum the updated first amplitude edge characterization value to obtain a first cumulative value, and then determine the target Doppler index according to the first cumulative value, and then estimate the vehicle speed according to the target Doppler index, mainly using a method of first extracting the amplitude edge feature and then summing it. The following also provides a method of first accumulating and summing the RD image and then extracting the amplitude edge feature to implement the vehicle speed estimation method provided in the embodiment of the present application.

[0182] like Fig.15 In a detailed flow chart of step S302 shown, the RD image is first cumulatively summed, and then the second amplitude edge representation value is extracted. Step S302 may include the following steps S1501-S1502.

[0183] S1501 , respectively accumulating pixel values ​​of pixels corresponding to each Doppler index in the RD image to obtain second accumulated values ​​corresponding to each Doppler index.

[0184] The electronic device may accumulate the pixel values ​​of each pixel in sequence according to the distance index for each Doppler index to obtain a second accumulated value corresponding to each Doppler index.

[0185] For example, Figure 5 In the RD image shown, the pixel value of each pixel can be expressed as:

[0186] RD(n,m);

[0187] Where n = 1, 2, ..., N; m = 1, 2, ..., M; RD (n, m) means Figure 5 The pixel value of the pixel in row dn and column rm.

[0188] For the Doppler index n, the electronic device can accumulate the pixel values ​​of each pixel in sequence according to the distance index to obtain a second accumulated value A sum (n) is:

[0189]

[0190] Where n = 1, 2, ..., N; Σ represents the sum, A sum (n) is the second accumulated value corresponding to the Doppler index n.

[0191] S1502, extracting the amplitude edge features of the second accumulated values ​​corresponding to the Doppler indexes arranged according to the size of the index values, and obtaining the second amplitude edge representation values ​​corresponding to the Doppler indexes.

[0192] The second amplitude edge characterization value corresponding to the Doppler index characterizes the amplitude difference between all pixels corresponding to the Doppler index and all pixels corresponding to the adjacent Doppler index. The electronic device can apply the edge detection operator to the second accumulated value corresponding to each Doppler index one by one according to the index value, so as to obtain the second amplitude edge characterization value corresponding to each Doppler index.

[0193] For example, Figure 5 The RD image shown in FIG. 1 is summed horizontally according to step S1501 to obtain a second accumulated value A. sum (n), the edge detection operator f is applied to the second accumulated value corresponding to the Doppler index n, and the second amplitude edge representation value B(n) corresponding to the Doppler index n can be obtained: sum (n).

[0194] In one possible implementation, the electronic device may use a second convolution kernel to perform convolution transformation on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, and obtain a second amplitude edge characterization value corresponding to each Doppler index. The size of the second convolution kernel may be 3×1, 5×1, etc., and is composed of multiple 1×1 one-dimensional convolution blocks stacked together. The use of one-dimensional convolution transformation to extract amplitude edge features can effectively identify edges in images, and has low computational complexity, saving computing resources.

[0195] For example, still Figure 5 Taking the RD image shown in FIG. 1 as an example, after performing horizontal summation on the RD image according to step S1501, a second accumulated value A is obtained. sum (n), n = 1, 2, ..., N; the second convolution kernel is kernal3, kernal3 = (k -1 , k0, k1) = (0.2, 0.5, 0.3), the electronic device uses the second convolution kernel kernal3 to convolve the second accumulated value A sum (n) Perform convolution transformation one by one to obtain the second amplitude edge representation value corresponding to each Doppler index:

[0196]

[0197] Wherein, n=1, 2, ..., N; i is the subscript of each component in kernal3, i=-1, 0, 1; B(n) is the second amplitude edge representation value corresponding to the Doppler index n.

[0198] Above Fig.15In the provided technical solution, the pixel values ​​of the pixels corresponding to each Doppler index are first accumulated, and then the second amplitude edge representation value is extracted. There is no need to extract the first amplitude edge representation value of each pixel one by one, which greatly reduces the amount of calculation for extracting the amplitude edge feature, and thus greatly saves the computing resources required for extracting the amplitude edge feature, thereby greatly saving the computing resources required for vehicle speed estimation.

[0199] Based on the above Fig.15 After extracting the second amplitude edge characterization value by the method shown, the electronic device can perform step S303 by the following method provided in an embodiment of the present application to determine the target Doppler index: determine the maximum value among the second amplitude edge characterization values, and use the Doppler index corresponding to the determined maximum value as the target Doppler index.

[0200] The electronic device can perform function fitting on each Doppler index and the corresponding second amplitude edge characterization value to obtain a function of the second amplitude edge characterization value relative to the Doppler index, and then obtain the maximum value of the second amplitude edge characterization value corresponding to each Doppler index by solving the maximum value of the function.

[0201] The electronic device can also traverse the second amplitude edge characterization values ​​of each Doppler index in turn, compare two second amplitude edge characterization values ​​each time and record the larger second amplitude edge characterization value and the corresponding Doppler index. After traversing all Doppler indexes, the electronic device can determine the maximum value among the second amplitude edge characterization values ​​and determine the Doppler index corresponding to the maximum value as the target Doppler index.

[0202] In an example, the second amplitude edge characterization values ​​corresponding to the respective Doppler indexes are: B(1), B(2), ..., B(N), and the electronic device needs to determine the maximum value of the second amplitude edge characterization values:

[0203] B max =max i {B(1),B(2),…B(i),…,B(N)};

[0204] Among them, B max is the second largest amplitude edge representation value.

[0205] The electronic device can first compare the size of B(1) and B(2), where B(1) < B(2), then the electronic device records B(2) and the corresponding Doppler index 2; the electronic device then compares B(2) with B(3), where B(2) < B(3), then the electronic device records B(3) and the corresponding Doppler index 3; and so on. The largest second amplitude edge representation value B is determined at the end of the traversal. max After that, the electronic device can determine B maxThe corresponding target Doppler index k2.

[0206] After determining the target Doppler index based on the above method, the electronic device can execute step S304 through the following method provided in an embodiment of the present application to estimate the vehicle's driving speed: calculate the product of the target Doppler index and the Doppler resolution of the radar, and use the obtained product as the vehicle's driving speed.

[0207] Continuing with the above example, the electronic device may calculate the product of the determined target Doppler index k2 and the resolution ΔV of the radar to obtain k2×ΔV=V, and the electronic device may use the obtained product V as the driving speed of the vehicle.

[0208] In the above embodiment, the Doppler index corresponding to the maximum value of the second amplitude edge characterization value is determined as the target Doppler index. The determination process is relatively simple and can more accurately find the location where the echo intensity has a sudden change. When estimating the vehicle speed, the product of the target Doppler index and the Doppler resolution of the radar is used as the vehicle's driving speed. This simplifies the calculation process while ensuring the accuracy of the vehicle speed estimation.

[0209] The vehicle speed estimation method provided in the embodiment of the present application has a relatively simple calculation process, which transforms the speed solution problem into the edge detection problem of the RD image. It no longer depends on the azimuth measurement accuracy of the radar, but is only related to the speed measurement accuracy of the radar. The speed measurement accuracy is measured by the speed measurement resolution of the radar, which is also called speed resolution and Doppler resolution. The higher the speed measurement resolution of the radar, the more sensitive it is to speed micro-movements, the more accurate the measured Doppler value, and the more accurate the vehicle speed estimated using the Doppler value.

[0210] In addition, the vehicle speed estimation method provided in the embodiment of the present application is more sensitive to sudden changes in speed. When the vehicle is just started or is in the stage of rapid acceleration or deceleration, the vehicle acceleration further obtained by the vehicle speed estimated by the vehicle speed estimation method provided in the embodiment of the present application is more accurate than the acceleration measured by the inertial sensor.

[0211] In addition, accurate speed estimation can not only help the radar perceive the vehicle's own speed more accurately, and assist in more accurate measurement of azimuth through speed compensation, but can also be used as an odometer to assist the vehicle in accurate navigation and positioning, SLAM (Simultaneous Localization and Mapping), etc.

[0212] Corresponding to the above-mentioned vehicle speed estimation method, the embodiment of the present application also provides a vehicle speed estimation device, such as Fig.16 As shown, the device comprises:

[0213] The acquisition module 1601 is used to acquire a range Doppler RD image generated by a radar carried by the vehicle, wherein the RD image is an image obtained by performing fast Fourier transform on the detected echo data in the distance dimension and the speed dimension respectively, the horizontal direction of the RD image corresponds to the distance index, and the vertical direction of the RD image corresponds to the Doppler index;

[0214] An extraction module 1602 is used to extract amplitude edge features of the RD image;

[0215] The determination module 1603 is used to determine the target Doppler index corresponding to the target position in the RD image based on the amplitude edge feature, wherein the target position is: a position where the intensity of the echo received by the radar in the RD image changes suddenly;

[0216] The estimation module 1604 is used to estimate the vehicle's speed according to the target Doppler index and the Doppler resolution of the radar.

[0217] In the technical solution provided in the embodiment of the present application, the target Doppler index corresponding to the target position where the echo intensity received by the radar changes suddenly is determined by extracting the amplitude edge features in the RD image generated by the radar, and then the vehicle's driving speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0218] The relationship between the Doppler d, azimuth angle θ and vehicle speed V of a stationary target is: d = V × cosθ. The azimuth angle of a distant stationary object relative to the radar is relatively small and close to 0°, so d≈V can be obtained, that is, the Doppler value of a distant stationary object measured by the radar is approximately equal to the vehicle speed. In addition, the echo intensity reflected by a stationary object is much greater than the ground clutter or noise, so it is manifested on the RD image as a brighter strip along the distance dimension. In the above scheme, by extracting the amplitude edge features in the RD image, the brighter strip formed by the echo reflected by the distant stationary object along the distance dimension can be found, and then the target Doppler index corresponding to the target position where the echo intensity reflected by the distant stationary object changes suddenly can be determined, and then the vehicle's speed is estimated based on the target Doppler index and the Doppler resolution of the radar.

[0219] In the above scheme, the vehicle's speed can be estimated by relying on only one measurement information, namely the Doppler value corresponding to the target Doppler index in the RD image, which reduces the amount of measurement information used for vehicle speed estimation, and also reduces the impact of errors of multiple measurement information on vehicle speed estimation, thereby improving the accuracy of vehicle speed estimation.

[0220] In a possible implementation, the extraction module 1602 may be specifically configured to: extract the amplitude edge features of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel; for each pixel, if the absolute value of the first amplitude edge representation value of the pixel is greater than a first preset threshold, update the first amplitude edge representation value of the pixel to a first value; if the absolute value of the first amplitude edge representation value of the pixel is less than or equal to the first preset threshold, update the first amplitude edge representation value of the pixel to a second value.

[0221] In a possible implementation, the extraction module 1602 can be specifically used to: fill the boundaries of the RD image, wherein the number of filled pixel rows and / or pixel columns is determined based on the size of the first convolution kernel; and perform a convolution transformation on the RD image pixel by pixel along the distance dimension of the RD image using the first convolution kernel to obtain a first amplitude edge representation value for each pixel.

[0222] In a possible implementation, the determination module 1603 can be specifically used to: respectively accumulate the first amplitude edge representation values ​​of the pixels corresponding to each Doppler index in the RD image to obtain the first accumulated values ​​corresponding to each Doppler index; determine the maximum maximum value and the second maximum value in the first accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, and use the Doppler indexes corresponding to the maximum maximum value and the second maximum value as the target Doppler index.

[0223] In one possible implementation, the estimation module 1604 can be specifically used to: determine the Doppler index corresponding to the smallest first accumulated value among the candidate Doppler indexes, wherein the candidate Doppler indexes are all Doppler indexes between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second largest maximum value; calculate the product of the determined Doppler index and the Doppler resolution of the radar, and use the obtained product as the vehicle's driving speed.

[0224] In one possible implementation, the extraction module 1602 can be specifically used to: accumulate the pixel values ​​of the pixels corresponding to each Doppler index in the RD image respectively to obtain the second accumulated value corresponding to each Doppler index; extract the amplitude edge features of the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, to obtain the second amplitude edge representation value corresponding to each Doppler index.

[0225] In a possible implementation, the extraction module 1602 may be specifically configured to: use a second convolution kernel to perform convolution transformation on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value one by one, so as to obtain a second amplitude edge representation value corresponding to each Doppler index.

[0226] In a possible implementation, the determination module 1603 can be specifically used to: determine the maximum value among the second amplitude edge characterization values, and use the Doppler index corresponding to the determined maximum value as the target Doppler index; the estimation module is specifically used to: calculate the product of the target Doppler index and the Doppler resolution of the radar, and use the obtained product as the vehicle's driving speed.

[0227] The present application also provides an electronic device, such as Fig.17 As shown, including:

[0228] Memory 1701, used for storing computer programs;

[0229] The processor 1702 is used to implement the steps of any of the above-mentioned vehicle speed estimation methods when executing the program stored in the memory 1701.

[0230] In addition, the electronic device may further include a communication bus and / or a communication interface, and the processor 1702, the communication interface, and the memory 1701 communicate with each other via the communication bus.

[0231] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0232] The communication interface is used for communication between the above electronic device and other devices.

[0233] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0234] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0235] In another embodiment provided in the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned vehicle speed estimation methods are implemented.

[0236] In another embodiment provided in the present application, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute the steps of any vehicle speed estimation method in the above embodiments.

[0237] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a solid-state hard disk (SSD), etc.

[0238] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0239] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, electronic device, and storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0240] The above description is only a preferred embodiment of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application are included in the protection scope of the present application.

Claims

1. A vehicle speed estimation method, characterized in that: include: Acquire a range Doppler RD image generated by a radar carried by the vehicle, wherein the RD image is an image obtained by performing fast Fourier transform on the detected echo data in the distance dimension and the speed dimension respectively, the lateral direction of the RD image corresponds to the distance index, and the longitudinal direction of the RD image corresponds to the Doppler index; Extracting amplitude edge features of the RD image; Based on the amplitude edge feature, determining a target Doppler index corresponding to a target position in the RD image, wherein the target position is: a position in the RD image where a sudden change in the intensity of the echo received by the radar occurs; The driving speed of the vehicle is estimated according to the target Doppler index and the Doppler resolution of the radar.

2. The method according to claim 1, characterized in that The step of extracting the amplitude edge feature of the RD image comprises: Extracting the amplitude edge feature of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel; For each pixel, if the absolute value of the first amplitude edge characterization value of the pixel is greater than a first preset threshold, the first amplitude edge characterization value of the pixel is updated to a first value; if the absolute value of the first amplitude edge characterization value of the pixel is less than or equal to the first preset threshold, the first amplitude edge characterization value of the pixel is updated to a second value.

3. The method according to claim 2, characterized in that The step of extracting the amplitude edge feature of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel includes: Filling the border of the RD image, wherein the number of filled pixel rows and / or pixel columns is determined according to the size of the first convolution kernel; The first convolution kernel is used to perform convolution transformation on the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel.

4. The method according to claim 2 or 3, characterized in that: The step of determining a target Doppler index corresponding to a target position in the RD image based on the amplitude edge feature includes: Accumulating first amplitude edge representation values ​​of pixels corresponding to each Doppler index in the RD image respectively to obtain first accumulated values ​​corresponding to each Doppler index; Determine the maximum maximum value and the second maximum maximum value in the first accumulated value corresponding to each Doppler index arranged according to the size of the index value, and use the Doppler indexes corresponding to the maximum maximum value and the second maximum maximum value as the target Doppler index.

5. The method according to claim 4, characterized in that The estimating the vehicle's speed according to the target Doppler index and the Doppler resolution of the radar comprises: Determine a Doppler index corresponding to the smallest first accumulated value among candidate Doppler indexes, wherein the candidate Doppler indexes are all Doppler indexes between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second largest maximum value; The product of the determined Doppler index and the Doppler resolution of the radar is calculated, and the obtained product is used as the driving speed of the vehicle.

6. The method according to claim 1, characterized in that The step of extracting the amplitude edge feature of the RD image comprises: Accumulating pixel values ​​of pixels corresponding to each Doppler index in the RD image respectively to obtain second accumulated values ​​corresponding to each Doppler index; The amplitude edge features of the second accumulated values ​​corresponding to the Doppler indexes arranged according to the size of the index values ​​are extracted to obtain the second amplitude edge representation values ​​corresponding to the Doppler indexes.

7. The method according to claim 6, characterized in that The step of extracting the amplitude edge features of the second accumulated values ​​corresponding to the Doppler indexes arranged according to the size of the index values ​​to obtain the second amplitude edge representation values ​​corresponding to the Doppler indexes includes: Using the second convolution kernel, convolution transformation is performed one by one on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, so as to obtain the second amplitude edge representation value corresponding to each Doppler index.

8. The method according to claim 6 or 7, characterized in that: The step of determining a target Doppler index corresponding to a target position in the RD image based on the amplitude edge feature includes: Determine a maximum value among the second amplitude edge representation values, and use a Doppler index corresponding to the determined maximum value as a target Doppler index; The estimating the vehicle's speed according to the target Doppler index and the Doppler resolution of the radar comprises: The product of the target Doppler index and the Doppler resolution of the radar is calculated, and the obtained product is used as the driving speed of the vehicle.

9. A vehicle speed estimation device, characterized in that: include: an acquisition module, configured to acquire a range Doppler RD image generated by a radar carried by the vehicle, wherein the RD image is an image obtained by performing fast Fourier transform on the detected echo data in the distance dimension and the speed dimension respectively, the lateral direction of the RD image corresponds to the distance index, and the longitudinal direction of the RD image corresponds to the Doppler index; An extraction module, used for extracting amplitude edge features of the RD image; A determination module, configured to determine a target Doppler index corresponding to a target position in the RD image based on the amplitude edge feature, wherein the target position is a position in the RD image where a sudden change in the intensity of the echo received by the radar occurs; The estimation module is used to estimate the vehicle's travel speed according to the target Doppler index and the Doppler resolution of the radar.

10. The device according to claim 9, characterized in that The extraction module is specifically used to: extract the amplitude edge features of each pixel in the RD image pixel by pixel along the distance dimension of the RD image to obtain a first amplitude edge representation value of each pixel; for each pixel, if the absolute value of the first amplitude edge representation value of the pixel is greater than a first preset threshold, update the first amplitude edge representation value of the pixel to a first value; if the absolute value of the first amplitude edge representation value of the pixel is less than or equal to the first preset threshold, update the first amplitude edge representation value of the pixel to a second value; and / or The extraction module is specifically used to: perform border filling on the RD image, wherein the number of filled pixel rows and / or pixel columns is determined according to the size of the first convolution kernel; perform convolution transformation on the RD image pixel by pixel along the distance dimension of the RD image using the first convolution kernel to obtain a first amplitude edge representation value of each pixel; and / or The determination module is specifically used to: respectively accumulate the first amplitude edge representation values ​​of the pixels corresponding to each Doppler index in the RD image to obtain the first accumulated values ​​corresponding to each Doppler index; determine the maximum maximum value and the second maximum value in the first accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value, and use the Doppler indexes corresponding to the maximum maximum value and the second maximum value as the target Doppler index; and / or The estimation module is specifically used to: determine the Doppler index corresponding to the smallest first accumulated value among the candidate Doppler indexes, wherein the candidate Doppler indexes are all Doppler indexes between the Doppler index corresponding to the maximum maximum value and the Doppler index corresponding to the second largest maximum value; calculate the product of the determined Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle; and / or The extraction module is specifically used to: accumulate the pixel values ​​of the pixels corresponding to each Doppler index in the RD image respectively to obtain the second accumulated value corresponding to each Doppler index; extract the amplitude edge features of the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value to obtain the second amplitude edge representation value corresponding to each Doppler index; and / or The extraction module is specifically used to: use a second convolution kernel to perform convolution transformation on the second accumulated values ​​corresponding to each Doppler index arranged according to the size of the index value one by one, so as to obtain a second amplitude edge representation value corresponding to each Doppler index; and / or The determination module is specifically used to determine the maximum value among the second amplitude edge characterization values, and use the Doppler index corresponding to the determined maximum value as the target Doppler index; the estimation module is specifically used to calculate the product of the target Doppler index and the Doppler resolution of the radar, and use the obtained product as the driving speed of the vehicle.

11. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, for implementing the steps of the method described in any one of claims 1 to 8 when executing a program stored in a memory.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images