Obstacle detection method and device based on millimeter wave radar and electronic equipment

Abstract

The application discloses a kind of millimeter wave radar-based obstacle detection method, device and electronic equipment.Therein, the method comprises: in the process of vehicle driving, the target frame point cloud data collected by millimeter wave radar is acquired, the target frame point cloud data is used to characterize the obstacle around vehicle, and the millimeter wave radar is installed in front of vehicle;Curve fitting is carried out based on target frame point cloud data, and target fitting curve is obtained, target fitting curve includes first fitting curve and second fitting curve, and first fitting curve and second fitting curve are used to indicate the obstacle at different positions in front of vehicle;The confidence of target fitting curve is determined, and the confidence is used to determine whether there is obstacle around vehicle;Based on confidence, the target detection result of obstacle is determined.The application solves the technical problem that the detection efficiency of obstacle in road based on millimeter wave radar in the related art is low.

Description

Technical Field

[0001] This invention relates to the field of radar, and more specifically, to an obstacle detection method, apparatus, and electronic device based on millimeter-wave radar. Background Technology

[0002] Fences on highways or urban roads are generally divided into permanent and temporary fences. Permanent fences serve as median strips between two lanes, while temporary fences are used for municipal construction site enclosures. Permanent fences typically have double yellow lines on the ground, but temporary fences have no markings on the road surface. In actual driving, road fences are generally identified visually or by using lidar on the vehicle. However, relying on visual identification is difficult due to conditions such as snow accumulation and unclear lane markings. Furthermore, millimeter-wave radar detection suffers from low data processing precision, leading to low accuracy in fence identification. Lidar detection is limited by its short detection range and high cost, resulting in low efficiency within a limited budget.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides an obstacle detection method, apparatus, and electronic device based on millimeter-wave radar, to at least solve the technical problem of low detection efficiency of obstacles in roads based on millimeter-wave radar in related technologies.

[0005] According to one aspect of the present invention, an obstacle detection method based on millimeter-wave radar is provided, comprising: acquiring target frame point cloud data collected by millimeter-wave radar during vehicle operation, wherein the target frame point cloud data is used to characterize obstacles around the vehicle, and the millimeter-wave radar is installed in front of the vehicle; performing curve fitting based on the target frame point cloud data to obtain a target fitting curve, wherein the target fitting curve includes a first fitting curve and a second fitting curve, wherein the first fitting curve and the second fitting curve are used to represent obstacles at different positions in front of the vehicle; determining the confidence level of the target fitting curve, wherein the confidence level is used to determine whether there are obstacles around the vehicle; and determining the target detection result of the obstacle based on the confidence level.

[0006] Optionally, curve fitting is performed based on the target frame point cloud data to obtain a target fitting curve, including: dividing the points into gradients based on the positions of the points in the target frame point cloud data to obtain multiple first point set sets and multiple second point set sets; determining the number of first points contained in the first point set among the multiple first point set sets, and the number of second points contained in the second point set among the multiple second point set sets; determining the first point set corresponding to the maximum number of first points in the multiple first point set sets as the first target point set, and determining the second point set corresponding to the maximum number of second points in the multiple second point set sets as the second target point set; performing curve fitting on the first points contained in the first target point set and the second points contained in the second target point set respectively to obtain a first fitting curve and a second fitting curve.

[0007] Optionally, determining the confidence level of the target fitted curve includes: determining the first target coefficient corresponding to the first fitted curve and the second target coefficient corresponding to the second fitted curve; determining the first initial confidence level based on the first target coefficient and the second initial confidence level based on the second target coefficient; determining the first target confidence level based on the first distance between the first point and the first fitted curve, and determining the second target confidence level based on the second distance between the second point and the second fitted curve; obtaining the sum of the first initial confidence level and the first target confidence level, and the sum of the second initial confidence level and the second target confidence level, respectively, to obtain the confidence level of the first fitted curve and the confidence level of the second fitted curve.

[0008] Optionally, curve fitting is performed on the first traces contained in the first target trace set to obtain a first fitted curve, including: obtaining a first historical fitted curve and a first historical confidence level of the first historical fitted curve, wherein the first historical fitted curve is used to characterize the fitted curve obtained by curve fitting of historical frame point cloud data, and the first historical fitted curve and the first fitted curve are used to represent obstacles at the same position in front of the vehicle; in response to the first historical confidence level being greater than a first preset confidence level, the first historical traces contained in the first historical fitted curve are merged into the first target trace set to obtain a first merged trace set; in response to the number of traces of the first merged traces contained in the first merged trace set being greater than a first threshold, curve fitting is performed on the first merged traces to obtain the first fitted curve.

[0009] Optionally, curve fitting is performed on the second traces contained in the second target trace set to obtain a second fitted curve, including: obtaining a second historical fitted curve and a second historical confidence level of the second historical fitted curve, wherein the second historical fitted curve is used to characterize the fitted curve obtained by curve fitting of historical frame point cloud data, and the second historical fitted curve and the second fitted curve are used to represent obstacles at the same position in front of the vehicle; in response to the second historical confidence level being greater than a second preset confidence level, the second historical traces contained in the second historical fitted curve are merged into the second target trace set to obtain a second merged trace set; in response to the number of traces of the second merged traces contained in the second merged trace set being greater than a second threshold, curve fitting is performed on the second merged traces to obtain a second fitted curve.

[0010] Optionally, curve fitting is performed on the first traces contained in the first target trace set to obtain a first fitted curve, including: filtering the first target trace set in response to a first historical confidence level being less than or equal to a first preset confidence level; and filtering the first merged trace set in response to a first historical confidence level being greater than the first preset confidence level.

[0011] Optionally, curve fitting is performed on the second traces contained in the second target trace set to obtain a second fitted curve, including: filtering the second target trace set in response to the second historical confidence level being less than or equal to the second preset confidence level; and filtering the second merged trace set in response to the second historical confidence level being greater than the second preset confidence level.

[0012] Optionally, the method further includes: comparing the confidence level of the first fitted curve with the confidence level of the second fitted curve to obtain a first difference; if the first difference is greater than a first preset difference, comparing the confidence level of the first fitted curve with a third preset confidence level to obtain a second difference, and comparing the confidence level of the second fitted curve with the third preset confidence level to obtain a third difference; removing the first fitted curve when the second difference is greater than a second preset difference, and removing the second fitted curve when the third difference is greater than a second preset difference.

[0013] According to another aspect of the present invention, an obstacle detection device based on millimeter-wave radar is also provided, comprising: an acquisition module, configured to acquire target frame point cloud data collected by millimeter-wave radar during vehicle operation, the target frame point cloud data being used to characterize obstacles around the vehicle, the millimeter-wave radar being installed in front of the vehicle; a fitting module, configured to perform curve fitting based on the target frame point cloud data to obtain a target fitting curve, the target fitting curve including a first fitting curve and a second fitting curve, the first fitting curve and the second fitting curve being used to represent obstacles at different positions in front of the vehicle; a first determination module, configured to determine the confidence level of the target fitting curve, the confidence level being used to determine whether there are obstacles around the vehicle; and a second determination module, configured to determine the target detection result of the obstacle based on the confidence level.

[0014] According to another aspect of the present invention, a computer-readable storage medium is also provided, characterized in that the computer-readable storage medium includes a stored program, wherein the program executes the obstacle detection method based on millimeter-wave radar as described above when it is run.

[0015] According to another aspect of the present invention, an electronic device is also provided, comprising: one or more processors; a storage device for storing one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors perform any of the above-described obstacle detection methods based on millimeter-wave radar.

[0016] In the obstacle detection method based on millimeter-wave radar provided in this embodiment of the invention, target frame point cloud data collected by millimeter-wave radar is acquired during vehicle operation. This target frame point cloud data is used to characterize obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle. Curve fitting is performed based on the target frame point cloud data to obtain a target fitting curve. The target fitting curve includes a first fitting curve and a second fitting curve, which represent obstacles at different positions in front of the vehicle. The confidence level of the target fitting curve is determined, and this confidence level is used to determine whether obstacles exist around the vehicle. Based on the confidence level, the target detection result of the obstacle is determined. It is noteworthy that millimeter-wave radar can be used to acquire point cloud data at different positions in front of the vehicle, and a first fitting curve and a second fitting curve can be fitted. By calculating the confidence levels of the first and second fitting curves, the presence of obstacles around the vehicle can be determined, further improving the detection efficiency of obstacles in the road based on millimeter-wave radar and further solving the technical problem of low detection efficiency of obstacles in the road based on millimeter-wave radar in related technologies. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0018] Figure 1 This is a flowchart of an obstacle detection method based on millimeter-wave radar according to an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of a point cloud distribution according to an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of an obstacle detection device based on millimeter-wave radar according to an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] Example 1

[0024] Figure 1 This is a flowchart of an obstacle detection method based on millimeter-wave radar according to an embodiment of the present invention, as follows: Figure 1 As shown, the method includes the following steps:

[0025] Step S102: During the vehicle's operation, acquire target frame point cloud data collected by millimeter-wave radar. The target frame point cloud data is used to characterize obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle.

[0026] The vehicles mentioned above can be either autonomous or manually driven; there are no restrictions on the type of vehicle.

[0027] The aforementioned target frame point cloud data can be data collected by the vehicle's front radar to represent information about the road the vehicle is traveling on. The vehicle's front radar can be a millimeter-wave radar. Optionally, the information about the road the vehicle is traveling on can include the vehicle's position, obstacles encountered during the vehicle's journey, etc.

[0028] The obstacles around the vehicle mentioned above can be fences on both sides of the road on which the vehicle travels. Optionally, the fences can be permanent or temporary.

[0029] In one optional embodiment, a millimeter-wave radar can be installed in front of the vehicle. During the vehicle's movement, the millimeter-wave radar collects data in real time to obtain target frame point cloud data. Optionally, by analyzing the target frame point cloud data, it can be determined whether there are obstacles on the road where the vehicle is traveling.

[0030] Step S104: Perform curve fitting based on the target frame point cloud data to obtain the target fitting curve. The target fitting curve includes a first fitting curve and a second fitting curve. The first fitting curve and the second fitting curve are used to represent obstacles at different positions in front of the vehicle.

[0031] The first fitting curve mentioned above can be formed by fitting the point cloud containing the points collected by millimeter-wave radar to represent the information of the left front of the vehicle.

[0032] The second fitting curve mentioned above can be formed by fitting the point cloud containing the points collected by the millimeter-wave radar to represent the information of the right front of the vehicle.

[0033] Figure 2 This is a schematic diagram of a point cloud distribution according to an embodiment of the present invention, such as... Figure 2 As shown, after acquiring the target frame point cloud data, the target frame point cloud data can be classified to obtain the point cloud data of the vehicle's left front collected by the millimeter-wave radar and the point cloud data of the vehicle's right front collected by the millimeter-wave radar.

[0034] In one optional embodiment, the point cloud at the left front of the vehicle can be curve-fitted to obtain a first fitted curve, and the point cloud at the right front of the vehicle can be fitted to obtain a second fitted curve. Optionally, the point cloud at the left front of the vehicle and the point cloud at the right front of the vehicle can be curve-fitted using the Random Sample Consensus (RANSAC) method and the Gauss-Newton method. RANSAC calculates a reasonable mathematical model by randomly selecting subsamples from all noisy sample datasets. The Gauss-Newton method is an iterative method that uses least squares to calculate regression parameters in a nonlinear regression model.

[0035] RANSAC is used for curve fitting, which involves randomly sampling the target frame point cloud. Each sampling sample represents a portion of the total number of points in the current point cloud, such as 2 / 3. Iterative calculations are then performed, with the number of iterations set by those skilled in the art; in this application, 10 iterations are used for illustration. This allows for curve fitting of the selected point cloud points. Optionally, Gauss-Newton's method can be used for curve fitting, assuming that all point clouds can be fitted with a quadratic curve y = a0 + a1x + a2x. 2 +noise, where a0, a1, and a2 are the constant term, the coefficient of the first term, and the coefficient of the second term, respectively; x represents the coordinates of the point cloud; and noise is the residual value. The residual value can be determined by the distances between all points used for fitting and the fitted curve. The best fitted curve has the smallest residual value. The residual value corresponding to the best fitted curve can be calculated using the following formula. Where a0, a1, and a2 are the constant term, the coefficient of the linear term, and the coefficient of the quadratic term, respectively, and x i Coordinates used to represent point clouds.

[0036] Step S106: Determine the confidence level of the target fitted curve. The confidence level is used to determine whether there are obstacles around the vehicle.

[0037] The confidence level of the target fitted curve mentioned above may include the confidence level of the first fitted curve and the confidence level of the second fitted curve. Optionally, the confidence level may include the residual confidence level and the distance length confidence level.

[0038] In an optional embodiment, the confidence level of the residual value can be obtained by normalizing the total residual value to all points used for curve fitting. Optionally, assuming the total residual value has a minimum of 0 and a maximum of 8, the maximum and minimum values ​​of the residual value confidence level can be preset. For example, assuming the maximum residual value confidence level is 70 and the minimum is 0, then when the total residual value is 0, the residual value confidence level is the maximum value of 70, and when the total residual value is 8, the residual value confidence level is the minimum value of 0. That is, the larger the residual value, the smaller the residual value confidence level. Further, when the number of points used for curve fitting is greater than a preset value, the number of points falling at a distance greater than a threshold from the fitted curve can be determined first. The number of points falling outside the fitted curve is then compared with the number of points used for curve fitting, and the product of this ratio and the maximum residual value is used as the residual value confidence level. The preset value can be set by those skilled in the art; in this embodiment, a preset value of 30 is used for illustration.

[0039] In another optional embodiment, the distance confidence level can be determined in the following way: for example, it can be assumed that the distance confidence level is calculated starting when the distance between the fitted curve and the center point of the rear wheel of the vehicle is greater than or equal to 50 meters. Optionally, if the range of the distance confidence level is assumed to be 0-30, then when the distance between the fitted curve and the center point of the rear wheel of the vehicle is greater than or equal to 50 meters and less than or equal to 80 meters, the distance confidence level can be considered to be 0. When the distance between the fitted curve and the center point of the rear wheel of the vehicle is greater than 80 meters and less than or equal to 110 meters, the distance confidence level can be considered to increase linearly from 0. It can be specified that the distance confidence level is 30 when the fitted curve is 110 meters away from the center point of the rear wheel of the vehicle. Optionally, the above data are for illustrative purposes only, and specific settings can be made by those skilled in the art.

[0040] Furthermore, after determining the confidence levels of the first and second fitted curves, it is possible to determine whether there are obstacles around the vehicle based on the confidence levels.

[0041] Step S108: Determine the target detection result of the obstacle based on the confidence level.

[0042] In one optional embodiment, after obtaining the confidence levels of the first and second fitted curves, the presence of obstacles around the vehicle can be determined based on the magnitude of the confidence levels. Optionally, this can be determined by setting a threshold; when the confidence level is greater than the threshold, it can be considered that an obstacle exists near the vehicle.

[0043] In the obstacle detection method based on millimeter-wave radar provided in this embodiment of the invention, target frame point cloud data collected by millimeter-wave radar is acquired during vehicle operation. This target frame point cloud data is used to characterize obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle. Curve fitting is performed based on the target frame point cloud data to obtain a target fitting curve. The target fitting curve includes a first fitting curve and a second fitting curve, which represent obstacles at different positions in front of the vehicle. The confidence level of the target fitting curve is determined, and this confidence level is used to determine whether obstacles exist around the vehicle. Based on the confidence level, the target detection result of the obstacle is determined. It is noteworthy that millimeter-wave radar can be used to acquire point cloud data at different positions in front of the vehicle, and a first fitting curve and a second fitting curve can be fitted. By calculating the confidence levels of the first and second fitting curves, the presence of obstacles around the vehicle can be determined, further improving the detection efficiency of obstacles in the road based on millimeter-wave radar and further solving the technical problem of low detection efficiency of obstacles in the road based on millimeter-wave radar in related technologies.

[0044] Optionally, curve fitting is performed based on the target frame point cloud data to obtain a target fitting curve, including: dividing the points into gradients based on the positions of the points in the target frame point cloud data to obtain multiple first point set sets and multiple second point set sets; determining the number of first points contained in the first point set among the multiple first point set sets, and the number of second points contained in the second point set among the multiple second point set sets; determining the first point set corresponding to the maximum number of first points in the multiple first point set sets as the first target point set, and determining the second point set corresponding to the maximum number of second points in the multiple second point set sets as the second target point set; performing curve fitting on the first points contained in the first target point set and the second points contained in the second target point set respectively to obtain a first fitting curve and a second fitting curve.

[0045] The first set of points mentioned above can be the points contained in the point cloud collected by millimeter-wave radar to represent the left front of the vehicle.

[0046] The second set of points mentioned above can be the points contained in the point cloud collected by millimeter-wave radar to represent the right front of the vehicle.

[0047] In one optional embodiment, the target frame point cloud data is analyzed using relevant models or functions to determine the obstacle information at different locations on the road where the vehicle is traveling, represented by different point clouds in the target frame point cloud data. Furthermore, multiple first point set sets and multiple second point set sets can be tracked to form a point list, and the points can be stored.

[0048] In another optional embodiment, after obtaining the target frame point cloud data, the distance between adjacent points among multiple points contained in the target point cloud data can be determined. This distance can be a lateral distance. A distance threshold is set; when the distance between adjacent points is greater than the distance threshold, gradient partitioning can be performed on the multiple points to obtain multiple first point set sets and multiple second point set sets. Optionally, the multiple first point set sets can be compared, and the first point set containing the most points can be determined as the first target point set. Similarly, the multiple second point set sets can be compared, and the second point set containing the most points can be determined as the second target point set. Optionally, curve fitting can be performed using the points contained in the first target point set to obtain a first fitted curve, and curve fitting can be performed using the points contained in the second target point set to obtain a second fitted curve.

[0049] To illustrate, consider the following example: Assume the target frame point cloud data contains 100 points, with a distance threshold of 1. During gradient partitioning, distance detection reveals that the distance between any two adjacent points from the first to the tenth point is less than the distance threshold of 1. However, the distance between the tenth and eleventh points is greater than the distance threshold. Therefore, all points from the first to the tenth point can be defined as the first set of points. Optionally, gradient partitioning can be performed on the remaining eleventh to one hundredth points using the same method. For instance, points from the eleventh to the eightieth point can be defined as the second set of first points, and points from the eighty-first to the one hundredth point as the third set of first points. Furthermore, by comparing the number of points in each first set of points, it can be seen that the second set of first points determined by the eleventh to eightieth points has the largest number of points. Therefore, the second set of first points determined by the eleventh to eightieth points can be determined as the first target set of points, and the points in the first target set of points can be used for curve fitting.

[0050] Optionally, determining the confidence level of the target fitted curve includes: determining the first target coefficient corresponding to the first fitted curve and the second target coefficient corresponding to the second fitted curve; determining the first initial confidence level based on the first target coefficient and the second initial confidence level based on the second target coefficient; determining the first target confidence level based on the first distance between the first point and the first fitted curve, and determining the second target confidence level based on the second distance between the second point and the second fitted curve; obtaining the sum of the first initial confidence level and the first target confidence level, and the sum of the second initial confidence level and the second target confidence level, respectively, to obtain the confidence level of the first fitted curve and the confidence level of the second fitted curve.

[0051] The first target coefficient and the second target coefficient mentioned above can be the same or different. In this application, it is explained that the first target coefficient and the second target coefficient can be the same.

[0052] The first initial confidence level mentioned above can be the confidence level of the residual values ​​of the first fitted curve, and the second initial confidence level can be the confidence level of the residual values ​​of the second fitted curve.

[0053] The first target confidence level can be the distance length confidence level of the first fitted curve, and the second target confidence level can be the distance length confidence level of the second fitted curve.

[0054] In an alternative embodiment, the optimal residual value can be determined using the following formula. Where a0, a1, and a2 are the constant term, the coefficient of the linear term, and the coefficient of the quadratic term, respectively, and x i The coordinates used to represent the point cloud. Optionally, after obtaining the optimal residual values, the confidence level of the residual values ​​can be obtained by normalizing the total residual values ​​to all the points used to fit the curve.

[0055] In another optional embodiment, the distance length confidence level can be determined by determining the distances between the first fitted curve and the vehicle, and the distances between the second fitted curve and the vehicle. For example, it can be assumed that the distance length confidence level is calculated starting when the distance between the first fitted curve and the center point of the vehicle's rear wheel is 50 meters, and the range of the distance length confidence level is assumed to be 0-30. Optionally, when the distance between the first fitted curve and the center point of the vehicle's rear wheel is 50-80 meters, the distance length confidence level can be considered to be 0. When the distance between the first fitted curve and the center point of the vehicle's rear wheel is 81-110 meters, the distance length confidence level can be considered to increase linearly from 0. It can be specified that the distance length confidence level is 30 when the distance between the first fitted curve and the center point of the vehicle's rear wheel is 110 meters. Optionally, the above data are for illustrative purposes only, and specific settings can be made by those skilled in the art. Optionally, the distance length confidence level of the second fitted curve can also be calculated using this method.

[0056] Furthermore, after obtaining the first initial confidence level and the first target confidence level, the first initial confidence level and the first target confidence level can be added together to obtain the confidence level of the first fitted curve. After obtaining the second initial confidence level and the second target confidence level, the second initial confidence level and the second target confidence level can be added together to obtain the confidence level of the second fitted curve.

[0057] Furthermore, after obtaining the confidence level of the target fitting curve, a threshold can be set. Target fitting curves with a confidence level greater than the threshold are arranged and stored according to the distance between the target fitting curve and the vehicle. This allows the position of the point list at the next moment to be predicted based on vehicle speed, yaw rate, and time difference between adjacent frames. Thus, when the number of points contained in the future frame point cloud data is less than a certain value, prediction can be made using historical frame point cloud data with higher confidence, thereby improving the accuracy of obstacle detection.

[0058] Specifically, it can be done using the following formula:

[0059] theta = yawrate × deltaT, s = velocity × deltaT, where theta is the yaw angle, yawrate is the yaw rate, deltaT is the time difference, s is the distance the vehicle travels, and velocity is the vehicle speed.

[0060]

[0061] Optionally, curve fitting is performed on the first traces contained in the first target trace set to obtain a first fitted curve, including: obtaining a first historical fitted curve and a first historical confidence level of the first historical fitted curve, wherein the first historical fitted curve is used to characterize the fitted curve obtained by curve fitting of historical frame point cloud data, and the first historical fitted curve and the first fitted curve are used to represent obstacles at the same position in front of the vehicle; in response to the first historical confidence level being greater than a first preset confidence level, the first historical traces contained in the first historical fitted curve are merged into the first target trace set to obtain a first merged trace set; in response to the number of traces of the first merged traces contained in the first merged trace set being greater than a first threshold, curve fitting is performed on the first merged traces to obtain the first fitted curve.

[0062] The aforementioned first pre-set reliability can be set by those skilled in the art.

[0063] The aforementioned first threshold can be set by those skilled in the art.

[0064] In an optional embodiment, when performing curve fitting using the first target point set to form a first fitted curve, a first historical fitted curve can be obtained first, and the confidence level of the first historical fitted curve can be determined. When the confidence level of the first historical fitted curve is greater than a first preset confidence level, the first historical points contained in the first historical fitted curve can be merged into the first target point set to obtain a first merged point set. Furthermore, if the number of points in the first merged point set is greater than a first threshold, the first merged point set can be used for curve fitting to obtain the first fitted curve.

[0065] Optionally, curve fitting is performed on the second traces contained in the second target trace set to obtain a second fitted curve, including: obtaining a second historical fitted curve and a second historical confidence level of the second historical fitted curve, wherein the second historical fitted curve is used to characterize the fitted curve obtained by curve fitting of historical frame point cloud data, and the second historical fitted curve and the second fitted curve are used to represent obstacles at the same position in front of the vehicle; in response to the second historical confidence level being greater than a second preset confidence level, the second historical traces contained in the second historical fitted curve are merged into the second target trace set to obtain a second merged trace set; in response to the number of traces of the second merged traces contained in the second merged trace set being greater than a second threshold, curve fitting is performed on the second merged traces to obtain a second fitted curve.

[0066] The aforementioned second pre-set reliability can be set by those skilled in the art.

[0067] The aforementioned second threshold can be set by those skilled in the art.

[0068] In an optional embodiment, when using the second target point set to perform curve fitting to form the first fitted curve, a second historical fitted curve can be obtained first, and the confidence level of the second historical fitted curve can be determined. When the confidence level of the second historical fitted curve is greater than a second preset confidence level, the second historical points contained in the second historical fitted curve can be merged into the second target point set to obtain a second merged point set. Furthermore, if the number of points in the second merged point set is greater than a second threshold, the second merged point set can be used to perform curve fitting to obtain the second fitted curve.

[0069] Optionally, curve fitting is performed on the first traces contained in the first target trace set to obtain a first fitted curve, including: filtering the first target trace set in response to a first historical confidence level being less than or equal to a first preset confidence level; and filtering the first merged trace set in response to a first historical confidence level being greater than the first preset confidence level.

[0070] In one optional embodiment, when the first historical confidence level is less than or equal to the first preset confidence level, it is not necessary to merge the first historical points contained in the first historical fitted curve into the first target point set. In this case, the first target point set can be filtered to remove noise points. When the first historical confidence level is greater than the first preset confidence level, it is necessary to merge the first historical points contained in the first historical fitted curve into the first target point set to obtain the first merged point set. In this case, the first target point set can be filtered to remove noise points.

[0071] Optionally, curve fitting is performed on the second traces contained in the second target trace set to obtain a second fitted curve, including: filtering the second target trace set in response to the second historical confidence level being less than or equal to the second preset confidence level; and filtering the second merged trace set in response to the second historical confidence level being greater than the second preset confidence level.

[0072] In one optional embodiment, when the second historical confidence level is less than or equal to the second preset confidence level, it is not necessary to merge the second historical points contained in the second historical fitting curve into the second target point set. In this case, the second target point set can be filtered to remove noise points. When the second historical confidence level is greater than the second preset confidence level, it is necessary to merge the second historical points contained in the second historical fitting curve into the second target point set to obtain the second merged point set. In this case, the second target point set can be filtered to remove noise points.

[0073] Optionally, the method further includes: comparing the confidence level of the first fitted curve with the confidence level of the second fitted curve to obtain a first difference; if the first difference is greater than a first preset difference, comparing the confidence level of the first fitted curve with a third preset confidence level to obtain a second difference, and comparing the confidence level of the second fitted curve with the third preset confidence level to obtain a third difference; removing the first fitted curve when the second difference is greater than a second preset difference, and removing the second fitted curve when the third difference is greater than a second preset difference.

[0074] The first preset difference, the third preset confidence level, and the second preset difference mentioned above can all be set by those skilled in the art.

[0075] In an optional embodiment, after obtaining the confidence levels of the first and second fitted curves, the two-sided curves can be validated. Optionally, the difference between the confidence levels of the first and second fitted curves, i.e., the first difference, can be determined first. If the first difference is greater than the first preset difference, the confidence level of the first fitted curve is compared with the third preset confidence level to obtain the second difference. At the same time, the confidence level of the second fitted curve is compared with the third preset confidence level to obtain the third difference. Optionally, the first fitted curve can be removed when the second difference is greater than the second preset difference, and the second fitted curve can be removed when the third difference is greater than the second preset difference, thereby improving the detection accuracy of obstacles.

[0076] The solution described in this invention has the following beneficial effects:

[0077] 1) Compared to visual sensors, it requires less computation, does not require neural network inference, and can be applied to low-computing platforms such as microcontrollers.

[0078] 2) Compared with lidar, millimeter-wave radar can significantly reduce detection costs.

[0079] 3) Compared with traditional millimeter-wave radar, the accuracy of detection is significantly improved due to the addition of point tracking and gradient division processes.

[0080] Example 2

[0081] According to another aspect of the present invention, an obstacle detection device based on millimeter-wave radar is also provided. This device can execute the obstacle detection method based on millimeter-wave radar in the above embodiments. The specific implementation and preferred application scenarios are the same as those in the above embodiments, and will not be repeated here.

[0082] Figure 3 This is a schematic diagram of an obstacle detection device based on millimeter-wave radar according to an embodiment of the present invention, as shown below. Figure 3 As shown, the device includes the following components:

[0083] The acquisition module 302 is used to acquire target frame point cloud data collected by millimeter-wave radar during vehicle operation. The target frame point cloud data is used to characterize obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle.

[0084] The fitting module 304 is used to perform curve fitting based on the target frame point cloud data to obtain the target fitting curve. The target fitting curve includes a first fitting curve and a second fitting curve. The first fitting curve and the second fitting curve are used to represent obstacles at different positions in front of the vehicle.

[0085] The first determining module 306 is used to determine the confidence level of the target fitted curve, which is used to determine whether there are obstacles around the vehicle.

[0086] The second determination module 308 is used to determine the target detection result of the obstacle based on the confidence level.

[0087] Optionally, the fitting module 304 includes: a partitioning unit, used to perform gradient partitioning on the points based on the positions of the points in the target frame point cloud data, to obtain multiple first point set sets and multiple second point set sets; a first determining unit, used to determine the number of first points contained in the first point set among the multiple first point set sets, and the number of second points contained in the second point set among the multiple second point set sets; a second determining unit, used to determine the first point set corresponding to the maximum number of first points in the multiple first point set sets as the first target point set, and to determine the second point set corresponding to the maximum number of second points in the multiple second point set sets as the second target point set; and a fitting unit, used to perform curve fitting on the first points contained in the first target point set and the second points contained in the second target point set, respectively, to obtain a first fitting curve and a second fitting curve.

[0088] Optionally, the first determining module 306 includes: a third determining unit, used to determine a first target coefficient corresponding to a first fitted curve and a second target coefficient corresponding to a second fitted curve; a fourth determining unit, used to determine a first initial confidence level based on the first target coefficient and a second initial confidence level based on the second target coefficient; a fifth determining unit, used to determine a first target confidence level based on a first distance between a first point and the first fitted curve, and a second target confidence level based on a second distance between a second point and the second fitted curve; and an acquiring unit, used to acquire the sum of the first initial confidence level and the first target confidence level, and the sum of the second initial confidence level and the second target confidence level, respectively, to obtain the confidence level of the first fitted curve and the confidence level of the second fitted curve.

[0089] Optionally, the fitting unit includes: a first acquisition subunit, used to acquire a first historical fitting curve and a first historical confidence level of the first historical fitting curve, wherein the first historical fitting curve is used to characterize the fitting curve obtained by curve fitting of historical frame point cloud data, and the first historical fitting curve and the first fitting curve are used to represent obstacles at the same position in front of the vehicle; a first merging subunit, used to merge the first historical points contained in the first historical fitting curve into a first target point set in response to the first historical confidence level being greater than a first preset confidence level, to obtain a first merged point set; and a first fitting subunit, used to perform curve fitting on the first merged point set in response to the number of points of the first merged point set being greater than a first threshold, to obtain a first fitting curve.

[0090] Optionally, the fitting unit includes: a second acquisition subunit, used to acquire a second historical fitting curve and a second historical confidence level of the second historical fitting curve, wherein the second historical fitting curve is used to characterize the fitting curve obtained by curve fitting of historical frame point cloud data, and the second historical fitting curve and the second fitting curve are used to represent obstacles at the same position in front of the vehicle; a second merging subunit, used to merge the second historical points contained in the second historical fitting curve into a second target point set in response to the second historical confidence level being greater than a second preset confidence level, to obtain a second merged point set; and a second fitting subunit, used to perform curve fitting on the second merged point set in response to the number of points of the second merged point set being greater than a second threshold, to obtain a second fitting curve.

[0091] Optionally, the fitting unit further includes: a first filtering unit, configured to filter the first target point set in response to a first historical confidence level being less than or equal to a first preset confidence level; and a second filtering unit, configured to filter the first merged point set in response to a first historical confidence level being greater than the first preset confidence level.

[0092] Optionally, the fitting unit further includes: a third filtering unit for filtering the second target point set in response to a second historical confidence level being less than or equal to a second preset confidence level; and a fourth filtering unit for filtering the second merged point set in response to a second historical confidence level being greater than a second preset confidence level.

[0093] Optionally, the method apparatus includes: a first comparison module, configured to compare the confidence level of a first fitted curve with the confidence level of a second fitted curve to obtain a first difference; a second comparison module, configured to compare the confidence level of the first fitted curve with a third preset confidence level to obtain a second difference when the first difference is greater than a first preset difference, and to compare the confidence level of the second fitted curve with the third preset confidence level to obtain a third difference; and a removal module, configured to remove the first fitted curve when the second difference is greater than a second preset difference, and to remove the second fitted curve when the third difference is greater than a second preset difference.

[0094] Example 3

[0095] According to another aspect of the present invention, a computer-readable storage medium is also provided, characterized in that the computer-readable storage medium includes a stored program, wherein the program executes the obstacle detection method based on millimeter-wave radar as described above when it is run.

[0096] Example 4

[0097] According to another aspect of the present invention, an electronic device is also provided, comprising: one or more processors; a storage device for storing one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors perform any of the above-described obstacle detection methods based on millimeter-wave radar.

[0098] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0099] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0101] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0102] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0103] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0104] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An obstacle detection method based on millimeter-wave radar, characterized in that, include: During vehicle operation, target frame point cloud data collected by millimeter-wave radar is acquired. The target frame point cloud data is used to characterize the fence obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle. Determine the distances between multiple points contained in the target frame point cloud data; If the distance between adjacent points is greater than a distance threshold, the points are divided by gradient to obtain multiple sets of first points and multiple sets of second points. Determine the number of first traces in the plurality of first trace sets that contain first traces, and the number of second traces in the plurality of second trace sets that contain second traces; The first set of first points corresponding to the maximum number of first points among the plurality of first point sets is determined as the first target point set, and the second set of second points corresponding to the maximum number of second points among the plurality of second point sets is determined as the second target point set; Curve fitting is performed on the first trace contained in the first target trace set and the second trace contained in the second target trace set to obtain a first fitting curve and a second fitting curve. The first fitting curve and the second fitting curve are used to represent fence obstacles at different positions in front of the vehicle. Determine the first target coefficient and the first point cloud coordinates corresponding to the first fitted curve, and the second target coefficient and the second point cloud coordinates corresponding to the second fitted curve; The first residual confidence level is determined based on the first target coefficient and the first point cloud coordinates, and the second residual confidence level is determined based on the second target coefficient and the second point cloud coordinates; Based on the first distance between the first point and the first fitted curve, a first distance length confidence level is determined, and based on the second distance between the second point and the second fitted curve, a second distance length confidence level is determined. The confidence scores of the first residual and the first distance length are obtained respectively, and the confidence scores of the second residual and the second distance length are obtained respectively, to obtain the confidence scores of the first fitted curve and the second fitted curve. The confidence scores are used to determine whether there are fence obstacles around the vehicle. If the confidence level is greater than the confidence level threshold, it is determined that there is an obstacle in front of the vehicle.

2. The method according to claim 1, characterized in that, Curve fitting is performed on the first trace contained in the first target trace set and the second trace contained in the second target trace set, respectively, to obtain a first fitted curve and a second fitted curve, including: Obtain a first historical fitting curve and a first historical confidence level of the first historical fitting curve. The first historical fitting curve is used to characterize the fitting curve obtained by curve fitting of historical frame point cloud data, and the first historical fitting curve and the first fitting curve are used to represent fence obstacles at the same position in front of the vehicle. In response to the first historical confidence level being greater than the first preset confidence level, the first historical points contained in the first historical fitted curve are merged into the first target point set to obtain the first merged point set; In response to the fact that the number of first merged traces contained in the first merged trace set is greater than a first threshold, curve fitting is performed on the first merged traces to obtain the first fitted curve.

3. The method according to claim 1, characterized in that, Curve fitting is performed on the first trace contained in the first target trace set and the second trace contained in the second target trace set, respectively, to obtain a first fitted curve and a second fitted curve, including: A second historical fitting curve and a second historical confidence level of the second historical fitting curve are obtained. The second historical fitting curve is used to characterize the fitting curve obtained by curve fitting of historical frame point cloud data, and the second historical fitting curve and the second fitting curve are used to represent fence obstacles at the same position in front of the vehicle. In response to the second historical confidence level being greater than the second preset confidence level, the second historical points contained in the second historical fitted curve are merged into the second target point set to obtain the second merged point set; In response to the fact that the number of second merged traces contained in the second merged trace set is greater than the second threshold, curve fitting is performed on the second merged traces to obtain the second fitted curve.

4. The method according to claim 2, characterized in that, Curve fitting is performed on the first trace contained in the first target trace set and the second trace contained in the second target trace set, respectively, to obtain a first fitted curve and a second fitted curve, including: In response to the first historical confidence level being less than or equal to the first preset confidence level, the first target point set is filtered; In response to the first historical confidence level being greater than the first preset confidence level, the first merged trace set is filtered.

5. The method according to claim 3, characterized in that, Curve fitting is performed on the first trace contained in the first target trace set and the second trace contained in the second target trace set, respectively, to obtain a first fitted curve and a second fitted curve, including: In response to the second historical confidence level being less than or equal to the second preset confidence level, the second target point set is filtered; In response to the second historical confidence level being greater than the second preset confidence level, the second merged trace set is filtered.

6. The method according to claim 1, characterized in that, The method further includes: The confidence scores of the first fitted curve and the second fitted curve are compared to obtain the first difference. If the first difference is greater than the first preset difference, the confidence level of the first fitted curve is compared with the third preset confidence level to obtain the second difference, and the confidence level of the second fitted curve is compared with the third preset confidence level to obtain the third difference. When the second difference is greater than the second preset difference, the first fitted curve is removed, and when the third difference is greater than the second preset difference, the second fitted curve is removed.

7. An obstacle detection device based on millimeter-wave radar, characterized in that, include: The acquisition module is used to acquire target frame point cloud data collected by millimeter-wave radar during vehicle operation. The target frame point cloud data is used to characterize the fence obstacles around the vehicle. The millimeter-wave radar is installed in front of the vehicle. The fitting module is used to determine the distance between multiple points contained in the target frame point cloud data; when the distance between adjacent points is greater than a distance threshold, the points are divided by gradient to obtain multiple first point sets and multiple second point sets. The number of first traces contained in the first trace set among the plurality of first trace sets is determined, and the number of second traces contained in the second trace set among the plurality of second trace sets is determined; the first trace set corresponding to the maximum number of first traces among the plurality of first trace sets is determined as the first target trace set, and the second trace set corresponding to the maximum number of second traces among the plurality of second trace sets is determined as the second target trace set; curve fitting is performed on the first traces contained in the first target trace set and the second traces contained in the second target trace set, respectively, to obtain a first fitting curve and a second fitting curve, wherein the first fitting curve and the second fitting curve are used to represent fence obstacles at different positions in front of the vehicle; The first determining module is used to determine the first target coefficient and the first point cloud coordinates corresponding to the first fitted curve, and the second target coefficient and the second point cloud coordinates corresponding to the second fitted curve; determine the first residual confidence level based on the first target coefficient and the first point cloud coordinates, and determine the second residual confidence level based on the second target coefficient and the second point cloud coordinates; Based on the first distance between the first point and the first fitted curve, a first distance length confidence level is determined, and based on the second distance between the second point and the second fitted curve, a second distance length confidence level is determined; the sum of the first residual confidence level and the first distance length confidence level, and the sum of the second residual confidence level and the second distance length confidence level are obtained respectively, to obtain the confidence level of the first fitted curve and the confidence level of the second fitted curve, wherein the confidence level is used to determine whether there are fence obstacles around the vehicle; The second determining module is used to determine that there is an obstacle in front of the vehicle when the confidence level is greater than the confidence level threshold.

8. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the obstacle detection method based on millimeter-wave radar according to any one of claims 1 to 6.

Images