Deceleration strip determination method and device, storage medium and vehicle

Abstract

The invention relates to a deceleration strip determination method and device, a storage medium and a vehicle, and relates to the technical field of vehicles. The method comprises the following steps: acquiring a deceleration strip initial detection result corresponding to a to-be-detected image, clustering each sub-initial detection result included in the deceleration strip initial detection result based on a preset clustering rule to obtain a clustering result, and aiming at any cluster included in the clustering result and based on the sub-initial detection results included in the cluster, determining whether the to-be-detected image corresponds to the initial detection result of the to-be-detected image. And determining a sub-target detection result in the to-be-detected image, and obtaining a deceleration strip target detection result corresponding to the to-be-detected image based on all sub-target detection results in the to-be-detected image. The invention provides a novel method for processing each sub-initial detection result through clustering, and by adopting the method, the accuracy of determining the deceleration strip can be improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of vehicles, and in particular, to a speed bump determination method, apparatus, storage medium, and vehicle. Background Art

[0002] In the related art, a road surface is detected by using a vehicle perception system (such as a camera, a lidar, etc.). When a speed bump is detected, a suspension actuator makes an adaptive adjustment in advance, so that the vehicle can drive more smoothly over an uneven road surface, thereby enabling passengers to obtain an optimal comfort experience.

[0003] However, the speed bump detection algorithm in the related art has limited detection ability, and may output more detection results than the actual number of speed bumps on the road, reducing the accuracy of speed bump detection. Summary of the Invention

[0004] The purpose of the present disclosure is to provide a speed bump determination method, apparatus, storage medium, and vehicle.

[0005] According to a first aspect of an embodiment of the present disclosure, there is provided a speed bump determination method, including:

[0006] Obtaining an initial speed bump detection result corresponding to an image to be detected, where the initial speed bump detection result includes a plurality of sub-initial detection results;

[0007] Clustering each sub-initial detection result based on a preset clustering rule to obtain a clustering result;

[0008] For any cluster included in the clustering result, determining a sub-target detection result in the image to be detected based on the sub-initial detection results included in the cluster;

[0009] Based on all the sub-target detection results in the image to be detected, obtaining a target speed bump detection result corresponding to the image to be detected.

[0010] Optionally, the clustering each sub-initial detection result based on a preset clustering rule to obtain a clustering result includes:

[0011] Determining a clustering evaluation parameter between a first sub-initial detection result and a second sub-initial detection result based on the first sub-initial detection result and the second sub-initial detection result, where the first sub-initial detection result is any sub-initial detection result in the initial speed bump detection result, and the second sub-initial detection result is a sub-initial detection result different from the first sub-initial detection result in the initial speed bump detection result;

[0012] When the clustering evaluation parameter satisfies a preset parameter threshold, clustering the first sub-initial detection result and the second sub-initial detection result into the same cluster.

[0013] Optionally, the clustering evaluation parameter includes at least one of the following:

[0014] The directional angle difference between the line segments indicated by the two sub-initial detection results respectively;

[0015] The length of the merged line segment after merging the line segments indicated by the two sub-initial detection results respectively according to a preset merging strategy;

[0016] Among the line segments indicated by the two sub-initial detection results respectively, the maximum distance from the two endpoints of one line segment to the other line segment.

[0017] Optionally, the clustering evaluation parameter includes the length of the merged line segment after merging the line segments indicated by the two sub-initial detection results respectively according to a preset merging strategy. Determining the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result based on the first sub-initial detection result and the second sub-initial detection result includes:

[0018] Obtain the coordinates of the perpendicular foot points from the two endpoints of the first line segment to the second line segment respectively. The first line segment is the line segment indicated by the first sub-initial detection result, and the second line segment is the line segment indicated by the second sub-initial detection result;

[0019] Based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points, determine the length of the merged line segment.

[0020] Optionally, the determining the length of the merged line segment based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points includes:

[0021] Determine the abscissa of the left endpoint of the merged line segment as the minimum value among the abscissa of the left endpoint of the first line segment, the abscissa of the left endpoint of the second line segment, and the abscissa of each of the two perpendicular foot points;

[0022] Determine the ordinate of the left endpoint of the merged line segment as the minimum value among the ordinate of the left endpoint of the first line segment, the ordinate of the left endpoint of the second line segment, and the ordinate of each of the two perpendicular foot points;

[0023] Determine the abscissa of the right endpoint of the merged line segment as the maximum value among the abscissa of the right endpoint of the first line segment, the abscissa of the right endpoint of the second line segment, and the abscissa of each of the two perpendicular foot points;

[0024] Determine the ordinate of the right endpoint of the merged line segment as the maximum value among the ordinates of the right endpoint of the first line segment, the ordinate of the right endpoint of the second line segment, and the ordinates of the two foot points respectively;

[0025] Based on the abscissa of the left endpoint of the merged line segment, the ordinate of the left endpoint of the merged line segment, the ordinate of the right endpoint of the merged line segment, and the ordinate of the right endpoint of the merged line segment, obtain the length of the merged line segment.

[0026] Optionally, the method further includes:

[0027] Obtain an initial parameter threshold corresponding to the clustering evaluation parameter;

[0028] Determine a scale normalization factor according to the size parameter of the image to be detected;

[0029] Adjust the initial parameter threshold based on the scale normalization factor to obtain the preset parameter threshold.

[0030] Optionally, the determining a scale normalization factor according to the size parameter of the image to be detected includes:

[0031] Determine the width parameter included in the scale normalization factor based on the ratio of a preset value to the maximum value among the size parameters.

[0032] Optionally, for any cluster included in the clustering result, based on the sub-initial detection results included in the cluster, determining a sub-target detection result in the image to be detected includes:

[0033] For any cluster included in the clustering result, when the cluster includes one sub-initial detection result, determine the sub-initial detection result in the cluster as a sub-target detection result in the image to be detected;

[0034] For any cluster included in the clustering result, when the cluster includes multiple sub-initial detection results, obtain the minimum bounding rectangle of all the sub-initial detection results in the cluster, and determine a sub-target detection result in the image to be detected based on the midpoints of the two short sides of the minimum bounding rectangle.

[0035] According to a second aspect of the embodiments of the present disclosure, there is provided a speed bump determination device, including:

[0036] A first acquisition module, configured to acquire an initial speed bump detection result corresponding to an image to be detected, where the initial speed bump detection result includes multiple sub-initial detection results;

[0037] A clustering module, configured to cluster each sub-initial detection result based on a preset clustering rule to obtain a clustering result;

[0038] A first determination module, configured to, for any cluster included in the clustering result, determine a sub-target detection result in the to-be-detected image based on the sub-initial detection results included in the cluster;

[0039] A second determination module, configured to obtain a speed bump target detection result corresponding to the to-be-detected image based on all the sub-target detection results in the to-be-detected image.

[0040] According to a third aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium, on which computer program instructions are stored, and when the program instructions are executed by a processor, the steps of the speed bump determination method provided in the first aspect of the present disclosure are implemented.

[0041] According to a fourth aspect of the embodiments of the present disclosure, there is provided a vehicle, including:

[0042] A processor;

[0043] A memory for storing executable instructions executable by the processor;

[0044] Wherein, the processor is configured to execute the executable instructions stored in the memory to implement the steps of the speed bump determination method provided in the first aspect of the present disclosure.

[0045] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: By obtaining the speed bump initial detection result corresponding to the to-be-detected image, clustering a plurality of sub-initial detection results included in the speed bump initial detection result based on a preset clustering rule to obtain a clustering result, then for any cluster included in the clustering result, determining a sub-target detection result in the to-be-detected image based on the sub-initial detection results included in the cluster, and finally, a speed bump target detection result corresponding to the to-be-detected image can be obtained based on all the sub-target detection results in the to-be-detected image. The embodiments of the present disclosure provide a new method for processing each sub-initial detection result through clustering. By clustering and merging each sub-initial detection result into a cluster, and further obtaining the sub-target detection result corresponding to a final speed bump based on the sub-initial detection results within the same cluster, it enables each finally determined speed bump to refer to the sub-initial detection results within the corresponding cluster, improving the accuracy of speed bump determination.

[0046] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. Description of the Drawings

[0047] The accompanying drawings here are incorporated into and constitute a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure.

[0048] Figure 1 is a flowchart of a speed bump determination method shown according to an exemplary embodiment.

[0049] Figure 2 is a flowchart of another speed bump determination method shown according to an exemplary embodiment.

[0050] Figure 3 is a block diagram of a speed bump determination device shown according to an exemplary embodiment.

[0051] Figure 4 is a block diagram of a vehicle shown according to an exemplary embodiment. Detailed Description of the Embodiments

[0052] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0053] It should be noted that all actions of obtaining signals, information, or data in this application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where the location is located and obtaining the authorization given by the owner of the corresponding device.

[0054] In the related art, speed bump detection usually relies on a speed bump detection model to locate the edge contour position of the speed bump. The detection model obtains the speed bump contour line by predicting the positions of the center point and the two end points at the bottom of the speed bump. However, due to the limited detection ability of the detection model itself, there are usually the following two types of problems in actual applications: 1. The detection model may detect a speed bump as multiple small disconnected speed bumps, which is also called the "broken detection" problem; 2. The detection model may output two or more detection results for the same speed bump on the road, which is also called the "duplicate detection" problem.

[0055] It can be seen that both of the above two situations will cause the detection model to output more detection results than the actual number of speed bumps on the road, reducing the accuracy of speed bump detection.

[0056] In view of the above problems, the embodiments of the present disclosure provide a speed bump determination method, apparatus, storage medium, and vehicle, so as to provide a new method for processing multiple sub-initial detection results included in the initial detection result of the detected speed bump, thereby improving the accuracy of the final detection result of the speed bump.

[0057] Figure 1 It is a flowchart of a speed bump determination method shown according to an exemplary embodiment. As Figure 1 shown, the embodiments of the present disclosure provide a speed bump determination method, which can be executed by a vehicle, specifically, can be executed by a speed bump determination apparatus, and the apparatus can be implemented in a software and / or hardware manner and configured in the vehicle. As Figure 1 shown, the method may include the following steps.

[0058] In step S110, obtain the initial detection result of the speed bump corresponding to the image to be detected, and the initial detection result of the speed bump includes multiple sub-initial detection results.

[0059] Among them, the detection result corresponding to each detected speed bump in the initial detection result of the speed bump is called a sub-initial detection result.

[0060] In some embodiments, the initial detection result of the speed bump may be the detection result output by the speed bump detection model or the detection result output by other speed bump detection algorithms. Taking the example of obtaining the initial detection result of the speed bump through the speed bump detection model, a frame of road image (i.e., the image to be detected) collected by the vehicle camera can be used as the input, and after passing through the speed bump detection model, the initial detection result of the speed bump output by the speed bump detection model can be obtained, and the initial detection result of the speed bump may be a set composed of multiple sub-initial detection results.

[0061] In step S120, based on a preset clustering rule, cluster each sub-initial detection result to obtain a clustering result.

[0062] In some embodiments, the clustering result may include one or more clusters, and each cluster may include one or more sub-initial detection results.

[0063] In the embodiments of the present disclosure, considering that there may be sub-initial detection results corresponding to the same real speed bump among multiple sub-initial detection results, therefore, each sub-initial detection result can be clustered based on a certain clustering rule to cluster the sub-initial detection results corresponding to the same real speed bump into the same cluster.

[0064] In step S130, for any cluster included in the clustering result, based on the sub-initial detection results included in the cluster, determine a sub-target detection result in the image to be detected.

[0065] As can be seen from the foregoing, after clustering, one cluster can correspond to one real speed bump. Therefore, in the embodiments of the present disclosure, for any cluster included in the clustering result, one sub-target detection result in the image to be detected can be determined based on the sub-initial detection results included in the cluster.

[0066] Among them, one sub-target detection result can be understood as the detection result corresponding to one real speed bump in the image to be detected.

[0067] In step S140, based on all the sub-target detection results in the image to be detected, the speed bump target detection result corresponding to the image to be detected is obtained.

[0068] Among them, the speed bump target detection result can be understood as the detection result of the speed bump in the image to be detected finally determined.

[0069] In the embodiments of the present disclosure, after obtaining the sub-target detection results corresponding to all the real speed bumps in the image to be detected, the sub-target detection results corresponding to all the real speed bumps can be summarized as the final speed bump detection result of the image to be detected, that is, the speed bump target detection result is obtained.

[0070] By using the above method, by obtaining the initial speed bump detection result corresponding to the image to be detected, clustering the multiple sub-initial detection results included in the initial speed bump detection result based on a preset clustering rule to obtain a clustering result, and then for any cluster included in the clustering result, determining one sub-target detection result in the image to be detected based on the sub-initial detection results included in the cluster, and finally, based on all the sub-target detection results in the image to be detected, the speed bump target detection result corresponding to the image to be detected can be obtained. The embodiments of the present disclosure provide a new method for processing each sub-initial detection result through clustering. By clustering and merging each sub-initial detection result into a cluster, and further obtaining the sub-target detection result corresponding to one final speed bump based on the sub-initial detection results within the same cluster, the finally determined speed bump can refer to the sub-initial detection results within the corresponding cluster, improving the accuracy of speed bump determination.

[0071] In some embodiments, each sub-initial detection result can be clustered by sequentially determining whether two parameters can be clustered. Among them, for whether two parameters are clustered into one class, some clustering evaluation parameters can be used to judge. In this case, in step S120, clustering each sub-initial detection result based on a preset clustering rule to obtain a clustering result may include the following steps:

[0072] Based on the first sub-initial detection result and the second sub-initial detection result, determine the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result. The first sub-initial detection result is any one of the sub-initial detection results in the speed bump initial detection result, and the second sub-initial detection result is the sub-initial detection result in the speed bump initial detection result that is different from the first sub-initial detection result;

[0073] When the clustering evaluation parameter meets the preset parameter threshold, cluster the first sub-initial detection result and the second sub-initial detection result into the same cluster.

[0074] In the embodiments of the present disclosure, for any two sub-initial detection results in the speed bump initial detection result, for example, the first sub-initial detection result and the second initial detection result, the clustering evaluation parameter between these two sub-initial detection results can be calculated. If the clustering evaluation parameter between these two sub-initial detection results meets the preset parameter threshold, then these two sub-initial detection results can be clustered into the same cluster, that is, it is determined that these two sub-initial detection results belong to the same class.

[0075] In some embodiments, for the convenience of calculation, each sub-initial detection result in the speed bump initial detection result can be traversed in sequence, and then the clustering evaluation parameter is calculated between the currently traversed sub-initial detection result and other sub-initial detection results respectively, and based on the sequentially calculated clustering evaluation parameters, other sub-initial detection results that belong to the same cluster as the currently traversed sub-initial detection result are determined.

[0076] It can be understood that during the sequential traversal process, combinations for which the clustering evaluation parameter has already been calculated can be skipped. Exemplarily, when the currently traversed sub-initial detection result is the first sub-initial detection result, the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result has been calculated. Then, when the currently traversed sub-initial detection result is the second sub-initial detection result, the calculation of the clustering evaluation parameter between the second sub-initial detection result and the first sub-initial detection result can be skipped.

[0077] In addition, in some extreme scenarios, assuming that the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result meets the preset parameter threshold, at this time, the first sub-initial detection result and the second sub-initial detection result are clustered into the same cluster. At the same time, assuming that the clustering evaluation parameter between the first sub-initial detection result and the third sub-initial detection result meets the preset parameter threshold, at this time, the first sub-initial detection result and the third sub-initial detection result are also clustered into the same cluster. At the same time, assuming that when the second sub-initial detection result is traversed currently, the clustering evaluation parameter between the second sub-initial detection result and the third sub-initial detection result does not meet the preset parameter threshold. At this time, according to the clustering rule, the second sub-initial detection result and the third sub-initial detection result are not clustered into the same cluster. However, considering that the single sub-initial detection threshold has little impact on the finally determined speed bump, therefore, at this time, it can be considered to cluster the first sub-initial detection result, the second sub-initial detection result, and the third sub-initial detection result into the same cluster, or it can be considered to cluster the third sub-initial detection result into a cluster different from the first sub-initial detection result and the second sub-initial detection result.

[0078] In some embodiments, the clustering evaluation parameter includes at least one of the following:

[0079] The direction angle difference between the line segments indicated by the two sub-initial detection results respectively;

[0080] The length of the merged line segment after merging the line segments indicated by the two sub-initial detection results respectively according to the preset merging strategy;

[0081] Among the line segments indicated by the two sub-initial detection results respectively, the maximum distance from the two endpoints of one line segment to the other line segment.

[0082] In some embodiments, the above three parameters can be selected as the clustering evaluation parameters simultaneously. In this case, when the direction angle difference between the line segments indicated by the two sub-initial detection results respectively meets the corresponding preset angle threshold, and the length of the merged line segment after merging the line segments indicated by the two sub-initial detection results respectively according to the preset merging strategy meets the corresponding preset length threshold, and among the line segments indicated by the two sub-initial detection results respectively, the maximum distance from the two endpoints of one line segment to the other line segment meets the corresponding preset distance threshold, it is determined that the clustering evaluation parameter between the two sub-initial detection results meets the preset parameter threshold, and thus these two sub-initial detection results are clustered into the same cluster.

[0083] In some other embodiments, any two of the above parameters can also be selected simultaneously as the clustering evaluation parameters. For example, the direction angle difference between the line segments indicated by the two sub-initial detection results is selected simultaneously, and the length of the merged line segment after merging the line segments indicated by the two sub-initial detection results according to the preset merging strategy, or the length of the merged line segment after merging the line segments indicated by the two sub-initial detection results according to the preset merging strategy is selected simultaneously, and the maximum distance from the two endpoints of one line segment to the other line segment among the line segments indicated by the two sub-initial detection results.

[0084] In some embodiments, when the clustering evaluation parameter includes the length of the merged line segment after merging the line segments indicated by the two sub-initial detection results according to the preset merging strategy, based on the first sub-initial detection result and the second sub-initial detection result, determining the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result may include the following steps of calculating the length of the merged line segment:

[0085] Obtain the coordinates of the perpendicular foot points from the two endpoints of the first line segment to the second line segment, where the first line segment is the line segment indicated by the first sub-initial detection result and the second line segment is the line segment indicated by the second sub-initial detection result;

[0086] Based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points, determine the length of the merged line segment.

[0087] Among them, the coordinates of the perpendicular foot points from the two endpoints of the first line segment to the second line segment can be regarded as the coordinates of the points where the two endpoints of the first line segment are respectively mapped to the second line segment.

[0088] In the embodiments of the present disclosure, after obtaining the coordinates of the perpendicular foot points from the two endpoints of the first line segment to the second line segment, the length of the merged line segment can be determined based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points.

[0089] In some embodiments, to simplify the calculation process, the first line segment can be the shorter one, and the second line segment can be the longer one. That is, the line segment indicated by the first sub-initial detection result is shorter than the line segment indicated by the second sub-initial detection result.

[0090] In some embodiments, based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points, determining the length of the merged line segment may include the following steps:

[0091] Determine the abscissa of the left endpoint of the merged line segment as the minimum value among the abscissa of the left endpoint of the first line segment, the abscissa of the left endpoint of the second line segment, and the abscissas of the two foot points respectively;

[0092] Determine the ordinate of the left endpoint of the merged line segment as the minimum value among the ordinate of the left endpoint of the first line segment, the ordinate of the left endpoint of the second line segment, and the ordinates of the two foot points respectively;

[0093] Determine the abscissa of the right endpoint of the merged line segment as the maximum value among the abscissa of the right endpoint of the first line segment, the abscissa of the right endpoint of the second line segment, and the abscissas of the two foot points respectively;

[0094] Determine the ordinate of the right endpoint of the merged line segment as the maximum value among the ordinate of the right endpoint of the first line segment, the ordinate of the right endpoint of the second line segment, and the ordinates of the two foot points respectively;

[0095] Based on the abscissa of the left endpoint of the merged line segment, the ordinate of the left endpoint of the merged line segment, the ordinate of the right endpoint of the merged line segment, and the ordinate of the right endpoint of the merged line segment, obtain the length of the merged line segment.

[0096] Next, combined with an example, an exemplary description will be given of the calculation process when the above clustering evaluation parameters include different specific parameters:

[0097] Suppose the initial detection result of the speed bump corresponding to the image I to be detected can be recorded as: L res ={l 1 ,l 2 ,..l n}, where l i represents the i-th sub-initial detection result, and L res represents the set of all sub-initial detection results, that is, the initial detection result of the speed bump. Each sub-initial detection result contains the coordinates of two endpoints, which can be denoted as the left endpoint P 1 (x 1 ,y 1 ), and the right endpoint P 2 (x 2 ,y 2 ).

[0098] In some embodiments, when the clustering evaluation parameter includes the direction angle difference between the line segments indicated by two sub-initial detection results, the direction angle difference between the line segments indicated by any two sub-initial detection results (for example, l i and l j ) can be calculated by the following calculation formula:

[0099] α i =atan((y i1 -y i2) / (x i1 -x i2 ))*180

[0100] α j =atan((y j1 -y j2 ) / (x j1 -x j2 ))*180

[0101] Δα ij =min(|(α i -α j )|,|180-(α i -α j )|)

[0102] where α i and α j respectively represent the direction angles of the line segments indicated by the i-th and j-th sub-initial detection results. The point P i1 (x i1 , y i1 ), and P i2 (x i2 , y i2 ) respectively represent the left and right endpoint coordinates of the i-th sub-initial detection result. The point P j1 (x j1 , y j1 ), and P j2 (x j2 , y j2 ) respectively represent the left and right endpoint coordinates of the j-th sub-initial detection result. Δα ij represents the difference in the direction angles of the line segments indicated by the i-th and j-th sub-initial detection results.

[0103] In some embodiments, when the clustering evaluation parameter includes the length of the merged line segment after the line segments indicated by two sub-initial detection results are merged according to a preset merging strategy, the length of the merged line segment after the line segments indicated by any two sub-initial detection results (e.g., l i and l j ) are merged according to the preset merging strategy can be calculated by the following calculation formula:

[0104] First, calculate the foot of the perpendicular point (in the embodiments of the present disclosure, taking the first line segment as the shorter one and the second line segment as the longer one as an example):

[0105]

[0106]

[0107] where m i , mj are the initial detection results l i ,l j The length of the indicated line segment. Compare m i , m j Size, assuming m i <m j , then solve l i The left and right endpoints are l j The foot point of the indicated line segment is denoted by (x v ,y v ). The formula for finding the foot point is as follows:

[0108] k j =(y i1 -y i2 ) / (x i1 -x i2 )

[0109] a=k j

[0110] b=-1.0

[0111] c=y i1 -k j *y i2

[0112] x v =(b*b*x i1 -a*b*y i1 -a*c) / (a*a+b*b)

[0113] y v =(b*b*y i1 -a*b*x i1 -a*c) / (a*a+b*b)

[0114] According to the above formula, we can calculate l i The left and right endpoints are l j The foot points of the perpendicular are respectively denoted as points V 1 (x 1v ,y 1v ), click V 2 (x 2v ,y 2v ).

[0115] Second, calculate the length of the merged line segment:

[0116] Using the initial detection result l i , l j Left and right endpoints P i1 , P i2 , P j1 , P j2Coordinates and the foot of the perpendicular point V 1 , V 2 coordinates, calculate the coordinates of the two endpoints of the combined line segment after combining the two line segments. The left endpoint of the combined line segment is denoted as S 1 (x 1s , y 1s ), and the right endpoint of the combined line segment is denoted as S 2 (x 2s , y 2s ). The calculation formulas are as follows:

[0117] x 1s = min(x i1 , x j1 , x 1v , x 2v )

[0118] x 2s = max(x i1 , x j1 , x 1v , x 2v )

[0119] y 1s = min(y i1 , y j1 , y 1v , y 2v )

[0120] y 2s = max(y i1 , y j1 , y 1v , y 2v )

[0121] According to the coordinates of the two endpoints S 1 and S 2 of the obtained combined line segment, calculate the length of the combined line segment. The calculation formula is as follows:

[0122]

[0123] Among them, m s is the length of the combined line segment after combining the line segments indicated by the sub-initial detection result l i according to the preset combination strategy. j

[0124] In some embodiments, when the clustering evaluation parameter includes the maximum distance from the two endpoints of one of the line segments indicated by the two sub-initial detection results to the other line segment, the maximum distance from the two endpoints of one of the line segments to the other line segment can be calculated by the following calculation formula:

[0125] In the embodiments of the present disclosure, using the sub-initial detection result l​i Left and right endpoints P i1 , P i2 coordinates, and l i The perpendicular foot points V j from the left and right endpoints to l 1 , V 2 coordinates (still taking the first line segment as the shorter one and the second line segment as the longer one as an example), calculate the maximum distance from the two endpoints of one line segment to the other line segment. The calculation formula is as follows:

[0126]

[0127]

[0128] h max = max(h 1 , h 2 )

[0129] where h 1 , h 2 are respectively the distances from the left and right endpoints of the sub-initial detection result l i to the line segment indicated by the sub-initial detection result l j , and h max is the maximum distance.

[0130] In some embodiments, that the clustering evaluation parameter meets the preset parameter threshold may mean that the clustering evaluation parameter is less than the preset parameter threshold.

[0131] In this case, when the clustering evaluation parameter includes the direction angle difference between the line segments indicated by the two sub-initial detection results respectively, that the direction angle difference meets the corresponding preset angle threshold may mean that the direction angle difference is less than the preset angle threshold.

[0132] Assume that the direction angle difference between the line segments indicated by the two sub-initial detection results is represented as Δα ij , and the preset angle threshold is th Δα , then the above process can be expressed by the calculation formula as Δα ij < th Δα .

[0133] Similarly, when the clustering evaluation parameter includes the length of the merged line segment after the line segments indicated by the two sub-initial detection results are merged according to the preset merging strategy, that the length of the merged line segment meets the corresponding preset length threshold may mean that the length of the merged line segment is less than the preset length threshold.

[0134] Assume that the length of the merged line segment is represented as m s , and the preset length threshold is th m , then the above process can be expressed by the calculation formula as ms <th m 。

[0135] Similarly, among the line segments indicated by the two sub-initial detection results included in the clustering evaluation parameters, when the maximum distance from the two endpoints of one line segment to the other line segment satisfies the corresponding preset distance threshold, the maximum distance may satisfy being less than the preset distance threshold.

[0136] Assume the maximum distance is represented as h max , and the preset distance threshold is th h , then the above process can be expressed by the calculation formula as h max <th h 。

[0137] In some embodiments, the preset parameter thresholds can be set in advance. For example, an angle threshold, a length threshold, and a distance threshold are set in advance.

[0138] In addition, considering that the image sizes of the images to be detected collected by the camera may be inconsistent, setting the same parameter thresholds may result in a decrease in the accuracy of the subsequent determined speed bumps. Therefore, in order to further improve the accuracy of speed bump determination, in some embodiments, the present disclosure embodiments can also adaptively adjust the pre-set angle threshold, length threshold, distance threshold, etc. based on the size parameters of the image. Therefore, the method of the present disclosure embodiments may further include the following steps:

[0139] Obtain the initial parameter thresholds corresponding to the clustering evaluation parameters;

[0140] Determine a scale normalization factor according to the size parameters of the image to be detected;

[0141] Adjust the initial parameter thresholds based on the scale normalization factor to obtain the preset parameter thresholds.

[0142] In the embodiments of the present disclosure, basic initial parameter thresholds can be set according to actual engineering experience. Subsequently, the corresponding initial parameter thresholds can be adaptively adjusted based on the size parameters of the image to be detected, so that the preset parameter thresholds can adapt to their respective image sizes.

[0143] Exemplarily, the angle threshold, length threshold, and distance threshold can be set to 5, 4, and 20 pixels respectively in advance.

[0144] In some embodiments, the product of the scale normalization factor and the initial parameter thresholds can be determined as the preset parameter thresholds.

[0145] Exemplarily, assume the initial parameter thresholds include an initial angle threshold, an initial length threshold, and an initial distance threshold, which are respectively represented as TH Δα 、THm , TH h , the scale normalization factor is denoted as scale. Then, the process of adjusting the initial parameter threshold based on the scale normalization factor to obtain the preset parameter threshold can be respectively expressed as follows:

[0146] th Δα = TH Δα * scale

[0147] th m = TH m * scale

[0148] th h = TH h * scale

[0149] In some embodiments, according to the size parameters of the image to be detected, determining the scale normalization factor may include the following steps:

[0150] Based on the ratio of the preset value to the parameter value with the largest value among the size parameters, determine the width parameter included in the scale normalization factor.

[0151] In some embodiments, the size parameters of the image to be detected may include width and height. In this case, the maximum value of the width and height can be selected, and then the ratio of the preset value to the maximum value of the width and height is used to determine the width parameter included in the scale normalization factor.

[0152] In some embodiments, the preset value can be taken as 640. In this case, the process of determining the scale normalization factor scale can be expressed as the following calculation formula:

[0153] scale = 640 / max(w, h)

[0154] where w and h are the width parameter and height parameter of the image to be detected respectively.

[0155] In some embodiments, after obtaining the clustering result, for any cluster included in the clustering result, based on the sub-initial detection results included in the cluster, determining a sub-target detection result in the image to be detected may include the following steps:

[0156] For any cluster included in the clustering result, in the case where the cluster includes one sub-initial detection result, determine the sub-initial detection result in the cluster as a sub-target detection result in the image to be detected;

[0157] For any cluster included in the clustering result, when the cluster includes multiple sub-initial detection results, obtain the minimum bounding rectangle of all the sub-initial detection results in the cluster, and determine a sub-target detection result in the image to be detected based on the midpoints of the two short sides of the minimum bounding rectangle.

[0158] As can be seen from the foregoing, the clustering result may include multiple clusters, and each cluster may include one or more sub-initial detection results.

[0159] Then, if a certain cluster in the clustering result includes only one sub-initial detection result, the sub-initial detection result can be directly determined as a sub-target detection result in the image to be detected. If a certain cluster in the clustering result includes multiple (i.e., two or more) sub-initial detection results, the minimum bounding rectangle of all the sub-initial detection results in the cluster can be obtained, and a sub-target detection result in the image to be detected can be determined based on the midpoints of the two short sides of the minimum bounding rectangle.

[0160] Exemplarily, assume that a certain cluster contains multiple sub-initial detection results, denoted as {l 1 ,l 2 ...l K}, a total of K sub-initial detection results. Take the left and right endpoint coordinates of all the sub-initial detection results, a total of 2K endpoints, use the minimum bounding rectangle solving function (such as the opencv minAreaRect function) to solve the minimum bounding rectangle of all the endpoints, return the four vertex coordinates of the minimum bounding rectangle, use the four vertex coordinates to judge the long and short sides, and respectively take the midpoints of the two short sides as the combined result of the K sub-initial detection results in the cluster, that is, a sub-target detection result in the image to be detected.

[0161] In the embodiments of the present disclosure, various detection results may include two endpoints, namely left and right.

[0162] By adopting the above method, clustering is performed by introducing multi-dimensional conditions such as direction angle difference, combined line segment length, and maximum distance, and all sub-initial detection results in the cluster are combined based on the minimum bounding rectangle method, which can simultaneously solve the problems of missed detection and repeated detection of speed bumps, improve the accuracy of speed bump determination, and save a large amount of manpower and time costs without data mining and retraining the model; in addition, by adaptively calculating the preset parameter threshold, it is suitable for different images to be detected, improving the applicable range of the speed bump determination method; in addition, since the direction angle difference, combined line segment length, maximum distance, and adaptive threshold are not affected by distortion, angle, and direction, good results can also be obtained for complex scenarios such as distorted and curved speed bumps, speed bumps at various angles and directions.

[0163] Figure 2 is a flowchart of another speed bump determination method shown according to an exemplary embodiment. AsFigure 2 As shown, an embodiment of the present disclosure provides a speed bump determination method, which can be executed by a vehicle, specifically by a speed bump determination device, which can be implemented in a software and / or hardware manner and configured in the vehicle. As Figure 2 shown, the method may include the following steps.

[0164] In step S201, the image to be detected is input into the speed bump detection model.

[0165] In step S202, a plurality of sub-initial detection results output by the speed bump detection model are obtained.

[0166] In step S203, based on the first sub-initial detection result and the second sub-initial detection result, the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result is determined.

[0167] Among them, the first sub-initial detection result is any one of the sub-initial detection results in the initial speed bump detection result, and the second sub-initial detection result is the sub-initial detection result in the initial speed bump detection result that is different from the first sub-initial detection result.

[0168] Among them, the clustering evaluation parameter includes the direction angle difference between the line segments indicated by the two sub-initial detection results respectively, the length of the merged line segment after the line segments indicated by the two sub-initial detection results are merged according to the preset merging strategy, and the maximum distance from the two endpoints of one of the line segments indicated by the two sub-initial detection results to the other line segment.

[0169] In step S204, the initial parameter threshold corresponding to the clustering evaluation parameter is obtained.

[0170] In step S205, based on the ratio of the preset value to the maximum value among the size parameters of the image to be detected, the scale normalization factor is determined.

[0171] In step S206, the initial parameter threshold is adjusted based on the scale normalization factor to obtain the preset parameter threshold.

[0172] In step S207, when the clustering evaluation parameter is less than the corresponding preset parameter threshold, the first sub-initial detection result and the second sub-initial detection result are clustered into the same cluster.

[0173] Among them, after clustering each sub-initial detection result in the initial speed bump detection result, a clustering result can be obtained.

[0174] In step S208, for any cluster included in the clustering result, when the cluster includes one sub-initial detection result, the sub-initial detection result in the cluster is determined as a sub-target detection result in the image to be detected.

[0175] In step S209, for any cluster included in the clustering result, when the cluster includes multiple sub-initial detection results, obtain the minimum bounding rectangle of all the sub-initial detection results in the cluster, and determine a sub-target detection result in the image to be detected based on the midpoints of the two shorter sides of the minimum bounding rectangle.

[0176] In step S210, based on all the sub-target detection results in the image to be detected, obtain the speed bump target detection result corresponding to the image to be detected.

[0177] Among them, the detailed descriptions of steps S201 - S210 can refer to the foregoing embodiments, and will not be elaborated here.

[0178] Among them, Figure 2 the order of some steps can be adjusted. For example, steps S204 - S206 can be executed before step S201. Or steps S204 - S206 are optional. Thus, in some scenarios, steps S204 - S206 can be not executed.

[0179] In addition, it should be noted that the speed bump determination method of the embodiments of the present disclosure is general and not necessarily limited to the speed bump scenario, and is also applicable to various detection algorithms for strip-shaped, linear or rod-shaped objects.

[0180] Figure 3 is a block diagram of a speed bump determination device shown according to an exemplary embodiment. Referring to Figure 3 , the speed bump determination device 300 includes:

[0181] A first acquisition module 310, configured to acquire a speed bump initial detection result corresponding to the image to be detected, where the speed bump initial detection result includes multiple sub-initial detection results;

[0182] A clustering module 320, configured to cluster each sub-initial detection result based on a preset clustering rule to obtain a clustering result;

[0183] A first determination module 330, configured to, for any cluster included in the clustering result, determine a sub-target detection result in the image to be detected based on the sub-initial detection results included in the cluster;

[0184] A second determination module 340, configured to obtain a speed bump target detection result corresponding to the image to be detected based on all the sub-target detection results in the image to be detected.

[0185] Optionally, the clustering module 320 includes:

[0186] A first determination sub-module, configured to determine a clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result based on the first sub-initial detection result and the second sub-initial detection result, where the first sub-initial detection result is any one of the sub-initial detection results in the speed bump initial detection result, and the second sub-initial detection result is a sub-initial detection result different from the first sub-initial detection result in the speed bump initial detection result.

[0187] A clustering sub-module, configured to cluster the first sub-initial detection result and the second sub-initial detection result into the same cluster when the clustering evaluation parameter meets a preset parameter threshold.

[0188] Optionally, the clustering evaluation parameter includes at least one of the following:

[0189] The direction angle difference between the line segments indicated by the two sub-initial detection results respectively;

[0190] The length of the merged line segment after the line segments indicated by the two sub-initial detection results are merged according to a preset merging strategy;

[0191] Among the line segments indicated by the two sub-initial detection results respectively, the maximum distance from the two endpoints of one line segment to the other line segment.

[0192] Optionally, the clustering evaluation parameter includes the length of the merged line segment after the line segments indicated by the two sub-initial detection results are merged according to a preset merging strategy. The first determination sub-module includes:

[0193] An acquisition unit, configured to acquire the coordinates of the perpendicular foot points of the two endpoints of the first line segment to the second line segment respectively, where the first line segment is the line segment indicated by the first sub-initial detection result, and the second line segment is the line segment indicated by the second sub-initial detection result;

[0194] A determination unit, configured to determine the length of the merged line segment based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points.

[0195] Optionally, the determination unit is further configured to determine the minimum value among the abscissa of the left endpoint of the first line segment, the abscissa of the left endpoint of the second line segment, and the abscissas of the two foot points as the abscissa of the left endpoint of the merged line segment; determine the minimum value among the ordinate of the left endpoint of the first line segment, the ordinate of the left endpoint of the second line segment, and the ordinates of the two foot points as the ordinate of the left endpoint of the merged line segment; determine the maximum value among the abscissa of the right endpoint of the first line segment, the abscissa of the right endpoint of the second line segment, and the abscissas of the two foot points as the abscissa of the right endpoint of the merged line segment; determine the maximum value among the ordinate of the right endpoint of the first line segment, the ordinate of the right endpoint of the second line segment, and the ordinates of the two foot points as the ordinate of the right endpoint of the merged line segment; and obtain the length of the merged line segment based on the abscissa of the left endpoint of the merged line segment, the ordinate of the left endpoint of the merged line segment, the ordinate of the right endpoint of the merged line segment, and the abscissa of the right endpoint of the merged line segment.

[0196] Optionally, the speed bump determination device 300 further includes:

[0197] A second acquisition module, configured to acquire an initial parameter threshold corresponding to the clustering evaluation parameter;

[0198] A third determination module, configured to determine a scale normalization factor according to the size parameter of the image to be detected;

[0199] An adjustment module, configured to adjust the initial parameter threshold based on the scale normalization factor to obtain the preset parameter threshold.

[0200] Optionally, the third determination module is further configured to determine the width parameter included in the scale normalization factor based on the ratio of a preset value to the largest parameter value among the size parameters.

[0201] Optionally, the first determination module 330 includes:

[0202] A second determination sub-module, configured to, for any one cluster included in the clustering result, when the cluster includes one sub-initial detection result, determine the sub-initial detection result in the cluster as a sub-target detection result in the image to be detected;

[0203] A third determination sub-module, configured to, for any one cluster included in the clustering result, when the cluster includes multiple sub-initial detection results, obtain the minimum bounding rectangle of all sub-initial detection results in the cluster, and determine a sub-target detection result in the image to be detected based on the midpoints of the two short sides of the minimum bounding rectangle.

[0204] Regarding the speed bump determination device 300 in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0205] The present disclosure also provides a computer-readable storage medium, on which computer program instructions are stored, and when the program instructions are executed by a processor, the steps of the speed bump determination method provided by the present disclosure are implemented.

[0206] Figure 4 is a block diagram of a vehicle shown according to an exemplary embodiment. For example, the vehicle 400 may be a hybrid vehicle, or a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. The vehicle 400 may be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0207] Referring to Figure 4 , the vehicle 400 may include various subsystems. For example, the infotainment system 410, the perception system 420, the decision control system 430, the drive system 440, and the computing platform 450. Among them, the vehicle 400 may also include more or fewer subsystems, and each subsystem may include multiple components. In addition, each subsystem and each component of the vehicle 400 may be interconnected by wired or wireless means.

[0208] In some embodiments, the infotainment system 410 may include a communication system, an entertainment system, and a navigation system, etc.

[0209] The perception system 420 may include several types of sensors for sensing information about the environment around the vehicle 400. For example, the perception system 420 may include a global positioning system (the global positioning system may be a GPS system, or a Beidou system, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter wave radar, ultrasonic radar, and a camera device.

[0210] The decision control system 430 may include a computing system, a vehicle controller, a steering system, an accelerator, and a braking system.

[0211] The drive system 440 may include components that provide power motion for the vehicle 400. In one embodiment, the drive system 440 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of an internal combustion engine, an electric motor, and an air compression engine. The engine can convert the energy provided by the energy source into mechanical energy.

[0212] Some or all of the functions of vehicle 400 are controlled by computing platform 450. Computing platform 450 may include at least one processor 451 and a memory 452, and the processor 451 may execute instructions 453 stored in the memory 452.

[0213] Processor 451 can be any conventional processor, such as a commercially available CPU. The processor may also include, for example, a Graphic Process Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), or a combination thereof.

[0214] Memory 452 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk.

[0215] In addition to instructions 453, memory 452 may also store data, such as road maps, route information, data on the position, direction, speed, etc. of the vehicle. The data stored in memory 452 can be used by computing platform 450.

[0216] In an embodiment of the present disclosure, processor 451 may execute instructions 453 to complete all or part of the steps of the above-mentioned speed bump determination method.

[0217] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program that can be executed by a programmable device, and the computer program has a code portion for executing the above-mentioned speed bump determination method when executed by the programmable device.

[0218] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0219] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A speed bump determination method, characterized in that, comprising: Obtaining an initial detection result of a speed bump corresponding to an image to be detected, where the initial detection result of the speed bump includes a plurality of sub-initial detection results; Based on a preset clustering rule, clustering each sub-initial detection result to obtain a clustering result; For any cluster included in the clustering result, based on the sub-initial detection results included in the cluster, determining a sub-target detection result in the image to be detected; Based on all the sub-target detection results in the image to be detected, obtaining a target detection result of the speed bump corresponding to the image to be detected.

2. The method according to claim 1, characterized in that, The step of clustering each sub-initial detection result based on a preset clustering rule to obtain a clustering result includes: Based on a first sub-initial detection result and a second sub-initial detection result, determining a clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result, where the first sub-initial detection result is any one of the sub-initial detection results in the initial detection result of the speed bump, and the second sub-initial detection result is a sub-initial detection result in the initial detection result of the speed bump that is different from the first sub-initial detection result; When the clustering evaluation parameter meets a preset parameter threshold, clustering the first sub-initial detection result and the second sub-initial detection result into the same cluster.

3. The method according to claim 2, characterized in that, The clustering evaluation parameter includes at least one of the following: The direction angle difference between the line segments indicated by two sub-initial detection results respectively; The length of the merged line segment after merging the line segments indicated by two sub-initial detection results according to a preset merging strategy; Among the line segments indicated by two sub-initial detection results, the maximum distance from the two endpoints of one line segment to the other line segment.

4. The method according to claim 2, characterized in that, The clustering evaluation parameter includes the length of the merged line segment after merging the line segments indicated by two sub-initial detection results according to a preset merging strategy. The step of determining the clustering evaluation parameter between the first sub-initial detection result and the second sub-initial detection result based on the first sub-initial detection result and the second sub-initial detection result includes: Obtaining the coordinates of the perpendicular foot points of the two endpoints of the first line segment to the second line segment respectively, where the first line segment is the line segment indicated by the first sub-initial detection result, and the second line segment is the line segment indicated by the second sub-initial detection result; Based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points, determining the length of the merged line segment.

5. The method according to claim 4, characterized in that, The step of determining the length of the merged line segment based on the coordinates of the two endpoints of the first line segment, the coordinates of the two endpoints of the second line segment, and the coordinates of the two perpendicular foot points includes: Determining the abscissa of the left endpoint of the merged line segment as the minimum value among the abscissa of the left endpoint of the first line segment, the abscissa of the left endpoint of the second line segment, and the abscissas of the two perpendicular foot points respectively; Determine the ordinate of the left endpoint of the combined line segment as the minimum value among the ordinates of the left endpoint of the first line segment, the ordinate of the left endpoint of the second line segment, and the ordinates of the two foot points respectively; Determine the abscissa of the right endpoint of the combined line segment as the maximum value among the abscissas of the right endpoint of the first line segment, the abscissa of the right endpoint of the second line segment, and the abscissas of the two foot points respectively; Determine the ordinate of the right endpoint of the combined line segment as the maximum value among the ordinates of the right endpoint of the first line segment, the ordinate of the right endpoint of the second line segment, and the ordinates of the two foot points respectively; Based on the abscissa of the left endpoint of the combined line segment, the ordinate of the left endpoint of the combined line segment, the ordinate of the right endpoint of the combined line segment, and the abscissa of the right endpoint of the combined line segment, obtain the length of the combined line segment.

6. The method according to claim 2, wherein, the method further includes: Obtain an initial parameter threshold corresponding to the clustering evaluation parameter; Determine a scale normalization factor according to the size parameter of the image to be detected; Adjust the initial parameter threshold based on the scale normalization factor to obtain the preset parameter threshold.

7. The method according to claim 6, wherein, the determining the scale normalization factor according to the size parameter of the image to be detected includes: Based on the ratio of a preset value to the maximum value among the size parameters, determine the width parameter included in the scale normalization factor.

8. The method according to claim 1, wherein, for any one of the clusters included in the clustering result, based on the sub-initial detection results included in the cluster, determining a sub-target detection result in the image to be detected includes: For any one of the clusters included in the clustering result, when the cluster includes one sub-initial detection result, determine the sub-initial detection result in the cluster as a sub-target detection result in the image to be detected; For any one of the clusters included in the clustering result, when the cluster includes multiple sub-initial detection results, obtain the minimum bounding rectangle of all the sub-initial detection results in the cluster, and based on the midpoints of the two shorter sides of the minimum bounding rectangle, determine a sub-target detection result in the image to be detected.

9. A speed bump determination device, wherein, it includes: A first acquisition module, configured to acquire an initial speed bump detection result corresponding to the image to be detected, where the initial speed bump detection result includes multiple sub-initial detection results; A clustering module, configured to cluster each sub-initial detection result based on a preset clustering rule to obtain a clustering result; A first determination module, configured to, for any one of the clusters included in the clustering result, based on the sub-initial detection results included in the cluster, determine a sub-target detection result in the image to be detected; A second determination module, configured to obtain a target speed bump detection result corresponding to the image to be detected based on all the sub-target detection results in the image to be detected.

10. A computer-readable storage medium, on which computer program instructions are stored, wherein, When the program instruction is executed by a processor, it implements the steps of the method according to any one of claims 1-8.

11. A vehicle, characterized in that it comprises: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the executable instructions stored in the memory to implement the steps of the method according to any one of claims 1-8.

Images