Obstacle moving speed determination method, device, equipment and medium

Abstract

The present disclosure provides a method and device for determining the moving speed of an obstacle, and a medium. Specifically, the present disclosure relates to the technical fields of autonomous driving, assisted driving, point cloud speed measurement, intelligent transportation, traffic violation detection, cloud computing, etc. The specific implementation scheme is as follows: determining at least two target point clouds from candidate point clouds; determining a reference point cloud and at least one to-be-matched point cloud from the at least two target point clouds according to the point cloud information quantity of the target point clouds, taking the two-dimensional projection image corresponding to the reference point cloud and the two-dimensional projection image corresponding to the to-be-matched point cloud as the reference projection image and the to-be-matched projection image respectively; performing matching filtering according to the reference projection image and the to-be-matched projection image, determining a matching filtering result, and determining the pixel point translation amount between the reference projection image and the to-be-matched projection image according to the matching filtering result; and determining the moving speed of the target obstacle according to the pixel point translation amount. The present disclosure achieves the effect of improving the accuracy of point cloud speed measurement.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular to the technical field of automatic driving, assisted driving, point cloud speed measurement, intelligent transportation, traffic violation detection and cloud computing, and particularly relates to a method and device for determining moving speed of an obstacle, equipment and medium. BACKGROUND

[0002] With the development of data acquisition technology, the form of data acquisition for obstacles is no longer limited to two-dimensional data acquisition, such as two-dimensional image acquisition, but more and more tends to three-dimensional data acquisition, and point cloud data acquisition is a typical three-dimensional data acquisition form.

[0003] Point cloud data represents data obtained by detecting obstacles by a radar sensor. Point cloud data is applied to various application scenarios, and using point cloud data to measure the speed of obstacles is a typical application scenario. SUMMARY

[0004] The present disclosure provides a method and device for determining moving speed of an obstacle to improve the accuracy of point cloud speed measurement.

[0005] According to an aspect of the present disclosure, a method for determining moving speed of an obstacle is provided, comprising:

[0006] determining at least two frames of target point clouds from candidate point clouds; wherein the candidate point clouds represent point clouds collected for a target obstacle;

[0007] determining a reference point cloud and at least one frame of to-be-matched point clouds from the at least two frames of target point clouds according to the amount of point cloud information of the target point clouds;

[0008] taking a two-dimensional projection image corresponding to the reference point cloud and a two-dimensional projection image corresponding to the to-be-matched point cloud as a reference projection image and a to-be-matched projection image, respectively;

[0009] performing matching filtering according to the reference projection image and the to-be-matched projection image, determining a matching filtering result, and determining a pixel point translation amount between the reference projection image and the to-be-matched projection image according to the matching filtering result;

[0010] determining the moving speed of the target obstacle according to the pixel point translation amount.

[0011] According to another aspect of the present disclosure, a device for determining moving speed of an obstacle is provided, comprising:

[0012] a first point cloud determination module configured to determine at least two frames of target point clouds from candidate point clouds; wherein the candidate point clouds represent point clouds collected for a target obstacle;

[0013] The second point cloud determining module is configured to determine a reference point cloud and at least one frame of to-be-matched point cloud from the at least two frames of target point clouds according to a point cloud information quantity of the target point cloud, wherein the point cloud information quantity is used to measure the richness of image information contained in the target point cloud.

[0014] The projection image determining module is configured to determine a two-dimensional projection image corresponding to the reference point cloud and a two-dimensional projection image corresponding to the to-be-matched point cloud as a reference projection image and a to-be-matched projection image, respectively.

[0015] The pixel point translation amount determining module is configured to perform matching filtering according to the reference projection image and the to-be-matched projection image, determine a matching filtering result, and determine a pixel point translation amount between the reference projection image and the to-be-matched projection image according to the matching filtering result.

[0016] The moving speed determining module is configured to determine the moving speed of the target obstacle according to the pixel point translation amount.

[0017] According to another aspect of the present disclosure, an electronic device is provided, comprising:

[0018] at least one processor; and

[0019] a memory connected with the at least one processor in communication; wherein,

[0020] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of the present disclosure.

[0021] According to another aspect of the present disclosure, a non-transitory computer readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to perform the method of any one of the present disclosure.

[0022] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, and the computer program is used to enable a processor to perform the method of any one of the present disclosure.

[0023] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings are used to better understand the present scheme, and do not constitute a limitation on the present disclosure. Among them:

[0025] Figure 1is a flow chart of some obstacle moving speed determination methods disclosed according to embodiments of the present disclosure;

[0026] Figure 2A is a flow chart of some other obstacle moving speed determination methods disclosed according to embodiments of the present disclosure;

[0027] Figure 2B is a schematic diagram of some target point clouds disclosed according to embodiments of the present disclosure;

[0028] Figure 2C is an interface schematic diagram of some grid divisions disclosed according to embodiments of the present disclosure;

[0029] Figure 2D is an interface schematic diagram of some auxiliary grid images disclosed according to embodiments of the present disclosure;

[0030] Figure 2E is a schematic diagram of some envelope maps disclosed according to embodiments of the present disclosure;

[0031] Figure 2F is a schematic diagram of some fitted line segments disclosed according to embodiments of the present disclosure;

[0032] Figure 3 is a structural schematic diagram of some obstacle moving speed determination apparatuses disclosed according to embodiments of the present disclosure;

[0033] Figure 4 is a block diagram of an electronic device for implementing the obstacle moving speed determination methods disclosed according to embodiments of the present disclosure. DETAILED DESCRIPTION

[0034] Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to help in understanding, and should be considered as merely exemplary. Thus, those of ordinary skill in the art should recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, for the sake of brevity and clarity, descriptions of well-known functions and constructions are omitted from the following description.

[0035] Speed measurement of obstacles using point cloud data is also referred to as point cloud speed measurement. Currently, the point cloud speed measurement methods commonly used in the industry generally include the following two methods: 1. Directly estimating the obstacle positioning result using point cloud, and then combining some filters to obtain the estimated result of the obstacle speed. 2. Projecting the point cloud onto the obstacle movement direction and the vertical direction of the obstacle movement direction, and then obtaining the speed estimation result according to the changes in the projected results between the point clouds.

[0036] However, the above method 1 generally has a lag, and the speed measurement result is unstable for obstacles with high maneuverability. The above method 2 has the characteristic of strong speed measurement maneuverability, but depends on the selection of the point cloud edge, which leads to the problem of insufficient speed measurement accuracy.

[0037] Figure 1 is a flowchart of some obstacle moving speed determination methods disclosed according to embodiments of the present disclosure. The embodiments can be applicable to the case of using point clouds to measure the speed of obstacles. The embodiment method can be executed by the obstacle moving speed determination device disclosed by the embodiments of the present disclosure, which can be implemented by software and / or hardware, and can be integrated on any electronic device with computing power, such as a vehicle with automatic driving or assisted driving function, or a smart signal light or road speedometer.

[0038] As shown in Figure 1 , the obstacle moving speed determination method disclosed by the embodiments can include:

[0039] S101, determining at least two target point clouds from candidate point clouds.

[0040] The point cloud is a data storage form of three-dimensional scene information obtained by a point cloud acquisition device (such as a laser radar), which is different from image data. The point cloud is composed of a series of discrete points in space and records the three-dimensional coordinates and reflection intensity of these points relative to the coordinate system of the point cloud acquisition device. The point cloud acquisition device obtains a frame of point cloud for each complete scanning of the surrounding environment.

[0041] The candidate point cloud represents the point cloud collected by the point cloud acquisition device for the target obstacle. The target obstacle represents any obstacle with mobility, including but not limited to vehicles, pedestrians, or animals, etc. The embodiments do not limit the specific type of target obstacle. The point cloud

[0042] The point cloud acquisition device represents any electronic device with point cloud acquisition function, including but not limited to laser radar, millimeter wave radar, 3D structured light camera, or depth image camera, etc. The embodiments do not limit the specific type of point cloud acquisition device.

[0043] When the point cloud acquisition device is mounted on a vehicle with automatic driving or assisted driving function, the obstacle moving speed determination method disclosed by the embodiments can be applied to the field of automatic driving technology or assisted driving technology. That is, the moving speed of the target obstacle is determined by using the point cloud collected for the target obstacle, so that the automatic driving module or the assisted driving module of the vehicle can plan the driving path of the vehicle or adjust the driving parameters of the vehicle according to the moving speed of the target obstacle.

[0044] When the point cloud collection device is mounted on an intelligent signal light or a road speed measurement instrument, the obstacle moving speed determination method disclosed in the embodiment can be applied to the field of traffic violation detection. That is, the moving speed of the target vehicle is determined by using the point cloud collected for the target vehicle, so as to determine whether the target vehicle has a speeding violation behavior according to the moving speed of the target vehicle.

[0045] In an implementation, according to the collection time of the candidate point cloud, one frame of candidate point cloud collected at the current time is taken as the current point cloud, and N frames of candidate point cloud collected at the historical time are taken as the historical point cloud, and the target point cloud is determined according to the current point cloud and the historical point cloud.

[0046] The historical time represents at least one collection time before the current time, and the historical time can be a continuous collection time or an interval collection time. In order to ensure the accuracy and reliability of the speed measurement, the historical time is optionally set as a continuous collection time.

[0047] The number N of the historical point cloud can be set and adjusted according to actual business requirements. Optionally, when it is necessary to improve the accuracy of the target obstacle speed measurement, the number N of the historical point cloud can be increased. Optionally, when it is necessary to improve the efficiency of the target obstacle speed measurement, the number N of the historical point cloud can be reduced.

[0048] In another implementation, M frames of candidate point cloud are randomly selected as the target point cloud. The M frames of candidate point cloud can be continuous candidate point cloud in collection time or interval candidate point cloud in collection time. In order to ensure the accuracy and reliability of the speed measurement, the M frames of continuous candidate point cloud in collection time are optionally selected as the target point cloud. The image number M of the randomly selected candidate point cloud can also be set and adjusted according to actual business requirements, which will not be described here.

[0049] By determining at least two frames of target point cloud from the candidate point cloud, the subsequent speed measurement calculation is used, so as to avoid using all candidate point clouds for speed measurement calculation, and the calculation amount is reduced.

[0050] S102, determining a reference point cloud and at least one frame of to-be-matched point cloud from the at least two frames of target point cloud according to the point cloud information amount of the target point cloud.

[0051] The point cloud information amount is used to measure the richness of the point cloud information contained in the target point cloud. That is, the greater the point cloud information amount, the richer the point cloud information contained in the target point cloud, and correspondingly, the smaller the point cloud information amount, the more scarce the point cloud information contained in the target point cloud.

[0052] The number of the reference point cloud is unique, indicating the target point cloud that plays a reference role in the subsequent matching filtering process, and is used for matching filtering with at least one frame of to-be-matched point cloud.

[0053] In an embodiment, the number of sub-points included in each frame of target point cloud is determined, and the point cloud information quantity of each frame of target point cloud is determined according to the number of sub-points.

[0054] In another embodiment, the difference value of reflection intensity of the collection points in each frame of target point cloud is determined, and the point cloud information quantity of each frame of target point cloud is determined according to the difference value of reflection intensity.

[0055] According to the numerical size of the point cloud information quantity of each frame of target point cloud, the reference point cloud and at least one frame of point cloud to be matched are determined from at least two frames of target point cloud. It can be understood that, in order to ensure the quality of matching filtering, the target point cloud with larger point cloud information quantity is selected as the reference point cloud, and the remaining target point cloud is selected as the point cloud to be matched.

[0056] By determining the reference point cloud and at least one frame of point cloud to be matched from at least two frames of target point cloud according to the point cloud information quantity of the target point cloud, the richness of the point cloud information of the reference point cloud is ensured, and the quality of the subsequent matching filtering result is further ensured.

[0057] S103, the two-dimensional projection image corresponding to the reference point cloud and the two-dimensional projection image corresponding to the point cloud to be matched are respectively taken as the reference projection image and the projection image to be matched.

[0058] The reference projection image represents a two-dimensional projection image obtained by projecting the collection points with three-dimensional coordinates in the reference point cloud into two-dimensional coordinates. The projection image to be matched represents a two-dimensional projection image obtained by projecting the collection points with three-dimensional coordinates in the point cloud to be matched into two-dimensional coordinates.

[0059] In an embodiment, a first transformation matrix between the world coordinate system and the device coordinate system is determined according to the pose of the point cloud collection device in the world coordinate system. The first coordinates of the collection points in the world coordinate system in the reference point cloud are transformed into the second coordinates in the device coordinate system according to the first transformation matrix. The second coordinates in the device coordinate system are transformed into the third coordinates in the image coordinate system according to the intrinsic parameters of the point cloud collection device, so as to obtain the reference projection image.

[0060] Correspondingly, the fourth coordinates of the collection points in the world coordinate system in the point cloud to be matched are transformed into the fifth coordinates in the device coordinate system according to the first transformation matrix. The fifth coordinates in the device coordinate system are transformed into the sixth coordinates in the image coordinate system according to the intrinsic parameters of the point cloud collection device, so as to obtain the projection image to be matched.

[0061] In another embodiment, a mapping relationship between two-dimensional coordinates and three-dimensional coordinate intervals is established in advance, that is, the three-dimensional coordinates belonging to a certain three-dimensional coordinate interval are projected onto the two-dimensional coordinates having a mapping relationship with the three-dimensional coordinate interval.

[0062] The first coordinates of each collection point in the world coordinate system in the reference point cloud are obtained, and a three-dimensional coordinate interval to which each first coordinate belongs is determined, and then according to the mapping relationship between the three-dimensional coordinate interval and the two-dimensional coordinate, the first coordinates of each collection point are projected onto the two-dimensional coordinate having the mapping relationship with the three-dimensional coordinate interval to which the collection point belongs, so as to obtain a reference projection image.

[0063] Correspondingly, the fourth coordinates of each collection point in the world coordinate system in the to-be-matched point cloud are obtained, and a three-dimensional coordinate interval to which each fourth coordinate belongs is determined, and then according to the mapping relationship between the three-dimensional coordinate interval and the two-dimensional coordinate, the fourth coordinates of each collection point are projected onto the two-dimensional coordinate having the mapping relationship with the three-dimensional coordinate interval to which the collection point belongs, so as to obtain a to-be-matched projection image.

[0064] This embodiment can quickly realize two-dimensional coordinate projection of the point cloud, improve the efficiency of the overall processing flow, and also has the effect of downsampling because the collection points belonging to the same three-dimensional coordinate interval are projected onto the same two-dimensional coordinate in the projection process, thereby reducing the subsequent calculation amount.

[0065] By taking the two-dimensional projection images corresponding to the reference point cloud and the to-be-matched point cloud as the reference projection image and the to-be-matched projection image respectively, the effect of image dimension reduction is realized, and a data foundation is laid for subsequent two-dimensional matching filtering according to the reference projection image and the to-be-matched projection image.

[0066] S104, matching filtering is performed according to the reference projection image and the to-be-matched projection image, a matching filtering result is determined, and a pixel point translation amount between the reference projection image and the to-be-matched projection image is determined according to the matching filtering result.

[0067] The matching filtering is a basic operation in image processing, and its purpose is to selectively extract certain aspects of the image, which convey important information in a specific application environment. In this embodiment, each to-be-matched projection image is matched with the reference projection image for matching filtering, and the extracted content (matching filtering result) can convey the pixel point translation amount between each to-be-matched projection image and the reference projection image.

[0068] The pixel point translation amount represents the coordinate translation amount of the pixel point in the to-be-matched projection image relative to the pixel point in the reference projection image due to the movement of the target obstacle.

[0069] In one embodiment, since the essence of matching filtering is a spatial convolution process, the image matrix corresponding to each to-be-matched projection image is regarded as a convolution kernel, and the image matrix of the reference projection image is regarded as a to-be-convolved matrix, convolution calculation is performed on each convolution kernel and the to-be-convolved matrix respectively, and the obtained convolution result matrix is regarded as the matching filtering result.

[0070] The envelope diagram corresponding to each matched filtering result (convolution result matrix) is generated, and the peak points in each envelope diagram are determined, and then the pixel point translation amount between each to-be-matched projection image and the reference projection image is determined according to the abscissa values corresponding to the peak points.

[0071] The principle of the above embodiment is explained as follows:

[0072] Since each to-be-matched projection image and the reference projection image are both obtained by projecting the point cloud collected from the same target obstacle, theoretically, with the movement of the target obstacle, a certain pixel point (projection point) in the reference projection image will only produce pixel point translation compared with the corresponding pixel point in each to-be-matched projection image, and will not change the pixel value (reflection intensity) of the pixel point. That is, the reference projection image and any to-be-matched projection image have the following relationship:

[0073] S b (t)=S a (t-t0)

[0074] Wherein, S b (t) represents the functional relationship between the image coordinate t and the pixel value S of any to-be-matched projection image, S a (t) represents the functional relationship between the image coordinate t and the pixel value S of the reference projection image, t represents the image coordinate (x, y), and t0 represents the pixel point translation amount (x0, y0) between the reference projection image and the to-be-matched projection image. It can be seen from the above formula that even if the image coordinate t produces the translation amount t0, the to-be-matched projection image and the reference projection image still have the same pixel value.

[0075] The matched filtering result h(t) can be calculated by the following formula:

[0076] h(t)=S a (t)*S b (t)=S a (t)*S a (t-t0)=sinc(A(t-t0))

[0077] Wherein, A represents a fixed constant.

[0078] It can be seen that the matched filtering result h(t) represents a functional relationship between the image coordinate t and the pixel value h. According to the functional expression of the matched filtering result h(t) and the envelope graph characteristic of the sinc function, it can be known that when t=t0, the peak point of the envelope graph is reached. Therefore, the peak point in the envelope graph is determined, and then the pixel point translation t0 between the to-be-matched projection image and the reference projection image is determined according to the abscissa value t corresponding to each peak point.

[0079] By performing matched filtering according to the reference projection image and the to-be-matched projection image, the matched filtering result is determined, and the pixel point translation between the reference projection image and the to-be-matched projection image is determined according to the matched filtering result, so that the effect of determining the pixel point translation through two-dimensional matched filtering is realized, which is easy to implement and has a relatively low calculation amount.

[0080] S105, determining the moving speed of the target obstacle according to the pixel point translation.

[0081] In an implementation manner, the pixel point translation between the reference projection image and each to-be-matched projection image is acquired, and the acquisition time interval between the reference projection image and each to-be-matched projection image is also acquired, and then the moving speed of the target obstacle is determined according to the pixel point translation and the acquisition time interval.

[0082] The disclosure determines the moving speed of the obstacle by using two-dimensional matched filtering of the two-dimensional projection image of the point cloud. Since the two-dimensional matched filtering process involves the overall image features of the two-dimensional projection image, the accuracy of determining the pixel point translation according to the matched filtering result can be ensured, the accuracy of the obstacle speed measurement is further ensured, and the adaptation range of the point cloud speed measurement is improved without relying on the selection of the point cloud edge. Moreover, since the speed is measured by using the point cloud acquired in multiple frames, the maneuverability of the speed measurement can be ensured, and the speed measurement scene of the obstacle with high maneuverability can be applied.

[0083] Figure 2A The flowchart of another obstacle moving speed determination method disclosed according to an embodiment of the disclosure is based on the above technical solution and is further optimized and expanded, and can be combined with each of the optional implementation manners.

[0084] As shown in Figure 2A the obstacle moving speed determination method disclosed in the embodiment can include:

[0085] S201, determining a current point cloud acquired at a current time and a preset number of historical point clouds acquired at historical times from candidate point clouds, and determining a target point cloud according to the current point cloud and the historical point clouds.

[0086] Here, "current moment" refers to the current point cloud acquisition time, and "current point cloud" refers to the candidate point cloud acquired at the current point cloud acquisition time. Similarly, "historical moment" refers to the historical point cloud acquisition time, and "historical point cloud" refers to the candidate point cloud acquired at the historical point cloud acquisition time.

[0087] In one implementation, the candidate point cloud captured at the current moment is taken as the current point cloud, and a preset number of candidate point clouds captured at historical moments are taken as historical point clouds. The image set of the current point cloud and the historical point clouds is taken as the target point cloud.

[0088] Figure 2B These are schematic diagrams of some target point clouds disclosed in embodiments of this disclosure, such as... Figure 2B As shown, the timeline reflects the chronological order of point cloud acquisition, with different candidate point clouds acquired as time progresses. The candidate point cloud 200, located on the far right of the timeline, is the candidate point cloud acquired at the current moment and serves as the current point cloud. A predetermined number of candidate point clouds 201, acquired at previous historical moments before candidate point cloud 200, are designated as historical point clouds. The image set 202 of historical point clouds 201 and the current point cloud 200 is used as the target point cloud.

[0089] By determining the current point cloud collected at the current moment from the candidate point cloud and a preset number of historical point clouds collected at historical moments, the target point cloud is determined based on the current point cloud and the historical point cloud. This ensures that the target point cloud contains the current point cloud collected at the latest collection moment (the current collection moment), so that the movement speed of the target obstacle determined based on the target point cloud can better reflect the current movement state of the target obstacle. Thus, the obstacle movement speed determination method provided in this embodiment can be adapted to scenarios where the instantaneous speed of the obstacle needs to be determined, such as autonomous driving scenarios or traffic violation detection scenarios.

[0090] Optionally, the preset number is negatively correlated with the point cloud acquisition distance and the number of point cloud acquisitions.

[0091] Among them, the point cloud acquisition distance is the distance between the point cloud acquisition device and the area where the target obstacle is located, and the point cloud acquisition quantity is the total number of acquisition points in the candidate point cloud.

[0092] In other words, the greater the distance between the point cloud acquisition device and the area where the target obstacle is located, the fewer historical point clouds are selected as the target point cloud; the closer the distance between the point cloud acquisition device and the area where the target obstacle is located, the more historical point clouds are selected as the target point cloud.

[0093] The reason is that the farther the distance between the point cloud collection device and the area where the target obstacle is located, the smaller the influence of the target obstacle on the point cloud collection device, and the closer the distance between the point cloud collection device and the area where the target obstacle is located, the greater the influence of the target obstacle on the point cloud collection device. For example, when the target obstacle is an obstacle vehicle and the point cloud collection device is installed in the current driving vehicle, the farther the distance between the current driving vehicle and the area where the obstacle vehicle is located, the lower the safety threat of the obstacle vehicle to the current driving vehicle, and the less the number of historical point clouds selected as target point clouds should be; the closer the distance between the current driving vehicle and the area where the obstacle vehicle is located, the greater the safety threat of the obstacle vehicle to the current driving vehicle, and the greater the number of historical point clouds selected as target point clouds should be.

[0094] The more the total number of collection points in the candidate point cloud, the less the number of historical point clouds selected as target point clouds; the less the total number of collection points in the candidate point cloud, the greater the number of historical point clouds selected as target point clouds.

[0095] The reason is that the more the total number of collection points in the candidate point cloud, the smaller the area of the target obstacle surface blocked by the obstacle or weather, and in order to improve the efficiency of speed measurement, the number of historical point clouds selected as target point clouds should be reduced; the less the total number of collection points in the candidate point cloud, the greater the area of the target obstacle surface blocked by the obstacle or weather, and in order to improve the accuracy of speed measurement, the number of historical point clouds selected as target point clouds should be increased.

[0096] By setting the negative correlation between the preset number and the point cloud collection distance and the point cloud collection number, the balance between the speed measurement efficiency and the speed measurement accuracy is ensured.

[0097] S202, determine the point cloud information amount of the target point cloud, and determine the reference point cloud and at least one frame of to-be-matched point cloud from at least two frames of target point clouds according to the point cloud information amount of the target point cloud.

[0098] Optionally, determining the point cloud information amount of the target point cloud comprises the following steps A1 and B1:

[0099] A1, taking the two-dimensional projection image corresponding to the target point cloud as a target projection image, and performing grid division on the target projection image to determine at least one target grid image included in the target projection image.

[0100] In an implementation, a two-dimensional coordinate projection is performed on each target point cloud to obtain a two-dimensional projection image as a target projection image corresponding to each target point cloud. Each target projection image is equally sized grid partitioned to obtain at least one target grid image in each target projection image.

[0101] Figure 2C Fig. 1 is an interface schematic diagram of some grid partitioning according to an embodiment of the present disclosure, as shown, a target projection image 203 is grid partitioned according to a grid size of 20x20, and each grid region 204 obtained is taken as a target grid image. Figure 2C

[0102] B1, taking the target grid image containing the projection point as an auxiliary grid image, and determining the point cloud information quantity according to the image quantity of the auxiliary grid image.

[0103] The image quantity of the auxiliary grid image reflects the distribution area of the projection point of the target point cloud, and the distribution area of the projection point reflects the richness of the point cloud information contained in the target point cloud, i.e., the point cloud information quantity of the target point cloud. The more the image quantity of the auxiliary grid image is, the greater the distribution area of the projection point of the target point cloud is, i.e., the greater the point cloud information quantity of the target point cloud is; correspondingly, the less the image quantity of the auxiliary grid image is, the smaller the distribution area of the projection point of the target point cloud is, i.e., the smaller the point cloud information quantity of the target point cloud is.

[0104] Figure 2D Fig. 2 is an interface schematic diagram of some auxiliary grid images according to an embodiment of the present disclosure, as shown, continuing to take the target projection image 203 as an example, it can be seen that there are seven target grid images containing projection points in the target projection image 203, and the image quantity of the auxiliary grid image contained in the target projection image 203 is 7. Figure 2D

[0105] By taking the two-dimensional projection image corresponding to the target point cloud as the target projection image, grid partitioning the target projection image, determining at least one target grid image included in the target projection image, taking the target grid image containing the projection point as the auxiliary grid image, and determining the point cloud information quantity according to the image quantity of the auxiliary grid image, since the image quantity of the auxiliary grid image reflects the distribution area of the projection point of the target point cloud, the effect of determining the point cloud information quantity from the dimension of the distribution area of the projection point is achieved, and the means of determining the point cloud information quantity is enriched.

[0106] Optionally, the grid partitioning of the target projection image includes the following steps A11, A12 and A13:

[0107] A11, obtaining a moving speed accuracy threshold, a distance resolution of a point cloud acquisition device, and a first acquisition time interval between candidate point clouds.

[0108] ​​The moving speed accuracy threshold value represents an accuracy value required for speed measurement. The distance resolution reflects the ability of the point cloud collection device to distinguish two obstacles. When the two obstacles are located at the same azimuth angle, but have different distances from the point cloud collection device, the minimum distance at which the two obstacles are distinguished by the point cloud collection device is called the distance resolution. The first collection time interval represents the interval value of the collection time between each candidate point cloud.

[0109] A12, determining the first parameter value according to the product result between the distance resolution and the constant factor, and determining the second parameter value according to the product result between the moving speed accuracy threshold value, the first collection time interval and the constant factor.

[0110] The constant factor represents a preset fixed constant, and the value of the constant factor can be selected from 0.4 to 0.8.

[0111] For example, assuming that the distance resolution of the point cloud collection device is represented by δ lidar , the constant factor is represented by α, the moving speed accuracy threshold value is represented by δ velocity , and the first collection time interval is represented by Δt, the first parameter value is α·δ lidar , and the second parameter value is α·δ velocity ·Δt.

[0112] A13, determining the target grid size according to the first parameter value and the second parameter value, and performing grid division on the target projection image using the target grid size.

[0113] The target grid size represents the image side length of the target grid image obtained by the division. For example, when the image side length of the target grid image is 20×20, the target grid size is 20.

[0114] In an embodiment, when the first parameter value is greater than the second parameter value, the first parameter value is taken as the target grid size; when the second parameter value is greater than the first parameter value, the second parameter value is taken as the target grid size.

[0115] Optionally, the target grid size is determined by the following formula:

[0116] size grid =max(α·δ lidar , α·δ velocity ·Δt)

[0117] The size grid represents the target grid size, α·δ lidar represents the first parameter value, and α·δ velocity ·Δt represents the second parameter value.

[0118] The moving speed precision threshold, the distance resolution of the point cloud acquisition device, and the first acquisition time interval between the candidate point clouds are acquired, a first parameter value is determined according to a product result between the distance resolution and a constant factor, a second parameter value is determined according to a product result between the moving speed precision threshold, the first acquisition time interval and the constant factor, a target grid size is determined from the first parameter value and the second parameter value according to a size relationship between the first parameter value and the second parameter value, and the target projection image is grid-divided by using the target grid size, so that the problem that the size of the target grid image obtained after the grid division is too large and the accuracy of the auxiliary grid image is insufficient is avoided, and the problem that the size of the target grid image obtained after the grid division is too small and the calculation amount of the auxiliary grid image is too large is avoided.

[0119] Optionally, the point cloud information amount of the target point cloud further includes the following steps A2 and B2:

[0120] A2, determining the maximum reflection intensity and the minimum reflection intensity of the acquisition points in the target point cloud.

[0121] The reflection intensity of the different regions of the obstacle surface to the acquisition signal of the point cloud acquisition device is different, that is, the reflection intensity of each acquisition point in the target point cloud is not the same. The maximum reflection intensity represents the maximum reflection intensity of each acquisition point in the target point cloud, and the minimum reflection intensity represents the minimum reflection intensity of each acquisition point in the target point cloud.

[0122] B2, determining the reflection intensity difference value according to the maximum reflection intensity and the minimum reflection intensity, and determining the point cloud information amount according to the reflection intensity difference value.

[0123] The reflection intensity difference value reflects the reflection intensity dynamic range of the acquisition points in the target point cloud, and the reflection intensity dynamic range reflects the richness of the point cloud information contained in the target point cloud, that is, the point cloud information amount of the target point cloud. The greater the reflection intensity difference value is, the greater the reflection intensity dynamic range of the acquisition points in the target point cloud is, that is, the greater the point cloud information amount of the target point cloud is. Correspondingly, the smaller the reflection intensity difference value is, the smaller the reflection intensity dynamic range of the acquisition points in the target point cloud is, that is, the smaller the point cloud information amount of the target point cloud is.

[0124] By determining the maximum reflection intensity and the minimum reflection intensity of the acquisition points in the target point cloud, determining the reflection intensity difference value according to the maximum reflection intensity and the minimum reflection intensity, and determining the point cloud information amount according to the reflection intensity difference value, since the reflection intensity difference value reflects the reflection intensity dynamic range of the acquisition points in the target point cloud, the effect of determining the point cloud information amount from the dimension of the reflection intensity dynamic range of the acquisition points is achieved, and the means of determining the point cloud information amount is enriched.

[0125] Optionally, determining the point cloud information amount according to the reflection intensity difference value includes the following steps B21 and B22:

[0126] B21, determine the sub-quantity of the collection points in the target point cloud.

[0127] wherein the number of the collection points included in each frame of the target point cloud is taken as the sub-quantity.

[0128] B22, determine the point cloud information quantity according to the weighted result between the sub-quantity, the image quantity and the reflection intensity difference value.

[0129] In an implementation, a first weight corresponding to the sub-quantity, a second weight corresponding to the image quantity, and a third weight corresponding to the reflection intensity difference value are obtained. The weighted sum calculation is performed according to the sub-quantity and the first weight, the image quantity and the second weight, and the reflection intensity difference value and the third weight to determine the weighted result, and the point cloud information quantity is determined according to the weighted result.

[0130] By determining the sub-quantity of the collection points in the target point cloud, the point cloud information quantity is determined according to the weighted result between the sub-quantity, the image quantity and the reflection intensity difference value, thereby achieving the effect of determining the point cloud information quantity from three dimensions together, and making the determination of the point cloud information quantity more reliable.

[0131] Optionally, according to the point cloud information quantity of the target point cloud, a reference point cloud and at least one to-be-matched point cloud are determined from the target point cloud, comprising:

[0132] determining the sorting result of the point cloud information quantity, and determining the reference point cloud from the target point cloud according to the sorting result; and taking the target point cloud other than the reference point cloud as the to-be-matched point cloud.

[0133] In an implementation, according to the sorting result of the point cloud information quantity, the target point cloud with the largest point cloud information quantity is taken as the reference point cloud, and the other target point clouds other than the reference point cloud are taken as the to-be-matched point cloud.

[0134] By determining the sorting result of the point cloud information quantity, and determining the reference point cloud from the target point cloud according to the sorting result, and taking the target point cloud other than the reference point cloud as the to-be-matched point cloud, the richness of the point cloud information quantity of the reference point cloud is ensured, so that subsequent matching filtering can be better performed according to the reference projection image and the to-be-matched projection image, and a better matching filtering result is obtained.

[0135] S203, taking the two-dimensional projection images corresponding to the reference point cloud and the to-be-matched point cloud as the reference projection image and the to-be-matched projection image respectively.

[0136] S204, performing Fourier transform on the reference projection image to determine the reference frequency spectrum image, and performing Fourier transform on the to-be-matched projection image to determine the to-be-matched frequency spectrum image.

[0137] Wherein, since the essence of the matched filtering is a spatial convolution process, and the spatial convolution between two images can be realized by the way of multiplying the frequency spectrum images.

[0138] In an embodiment, the reference frequency spectrum image and the to-be-matched frequency spectrum image are determined by the following formula:

[0139] S a (f) = FT(S a (t))

[0140] S b (f) = FT(S b (t))

[0141] Wherein, S a (f) represents the reference frequency spectrum image, S a (t) represents the reference projection image, S b (f) represents the to-be-matched frequency spectrum image, S b (t) represents the to-be-matched projection image, and FT() represents the Fourier transform.

[0142] S205, determining the frequency spectrum image product result between the reference frequency spectrum image and the to-be-matched frequency spectrum image, performing inverse Fourier transform on the frequency spectrum image product result, and determining the matched filtering result according to the inverse Fourier transform result.

[0143] In an embodiment, the matched filtering result is determined by the following formula:

[0144] H(f) = S a (f) x S b (f)

[0145] h(t) = IFT{H(f)}

[0146] Wherein, H(f) represents the frequency spectrum image product result, h(t) represents the matched filtering result, and IFT{} represents the inverse Fourier transform.

[0147] The reference frequency spectrum image is determined by performing Fourier transform on the reference projection image, the to-be-matched frequency spectrum image is determined by performing Fourier transform on the to-be-matched projection image, the frequency spectrum image product result between the reference frequency spectrum image and the to-be-matched frequency spectrum image is determined, the inverse Fourier transform is performed on the frequency spectrum image product result, and the matched filtering result is determined according to the inverse Fourier transform result, so that the spatial convolution process can be realized by multiplying the frequency spectrum images according to the relationship between the spatial domain and the frequency domain, the matched filtering result can be determined quickly and efficiently, and the efficiency of the entire speed measurement process is improved.

[0148] S206, an envelope graph corresponding to the matching filter result is acquired, a peak point in the envelope graph is determined, and a pixel point translation amount is determined according to a horizontal coordinate value of the peak point in the envelope graph.

[0149] The matching filter result h(t) between the reference projection image and the to-be-matched projection image represents a functional relationship between an image coordinate t and a pixel value h, and can be represented by sinc(A(t-t0)). According to the envelope graph characteristic of the sinc function, it can be known that the peak point is reached when t=t0. Therefore, the peak point is determined in the envelope graph, and the pixel point translation amount is determined according to the horizontal coordinate value of the peak point in the envelope graph.

[0150] Figure 2E Some envelope graphs according to the embodiments of the present disclosure are shown in the schematic diagram as shown in Figure 2E It can be seen that the peak point is reached when the independent variable t of the sinc function takes the value of zero, that is, the peak point is reached when t-t0=0, that is, t=t0 in the present embodiment. Therefore, as long as the peak point in the envelope graph is found, and the horizontal coordinate value t of the peak point is found, the pixel point translation amount t0 can be determined.

[0151] The envelope graph corresponding to the matching filter result is acquired, the peak point in the envelope graph is determined, and the pixel point translation amount is determined according to the horizontal coordinate value of the peak point in the envelope graph, so that the effect of quickly determining the pixel point translation amount according to the characteristic of the envelope graph corresponding to the matching filter result is realized, and the efficiency of the entire speed measurement process is improved.

[0152] S207, a second collection time interval between the reference point cloud and the to-be-matched point cloud is determined, a fitting line segment is generated through linear fitting according to the pixel point translation amount and the second collection time interval, and a moving speed is determined according to a first-order coefficient of the fitting line segment.

[0153] The second collection time interval represents the interval value of the collection time between each to-be-matched point cloud and the reference point cloud.

[0154] In an implementation manner, the pixel point translation amounts between each to-be-matched point cloud and the reference point cloud are sorted according to the collection time sequence of each to-be-matched point cloud, the second collection time interval corresponding to each pixel point translation amount is determined, and then the fitting line segment is generated through linear fitting according to the sorted pixel point translation amounts and the corresponding second collection time intervals, the first-order coefficient of the fitting line segment is determined, and the first-order coefficient is taken as the moving speed of the target obstacle.

[0155] Figure 2F Some fitting line segments according to the embodiments of the present disclosure are shown in the schematic diagram, and it is assumed that the pixel point translation amount t=(x0, y0), that is, the pixel point translation amount t is divided into the pixel point translation amount x0 in the image x-axis direction and the pixel point translation amount y0 in the image y-axis direction, as shown inFigure 2F As shown, 205 represents a first fitting line segment obtained by performing linear fitting according to the pixel translation x0 and the corresponding second acquisition time interval, and the first order coefficient of the first fitting line segment represents the moving sub-velocity of the target obstacle in the x-axis direction; 206 represents a second fitting line segment obtained by performing linear fitting according to the pixel translation y0 and the corresponding second acquisition time interval, and the first order coefficient of the second fitting line segment represents the moving sub-velocity of the target obstacle in the y-axis direction.

[0156] By determining the second acquisition time interval between the reference point cloud and the to-be-matched point cloud, performing linear fitting according to the pixel translation and the second acquisition time interval to generate a fitting line segment, and determining the moving velocity according to the first order coefficient of the fitting line segment, since the fitting line segment can express the change trend of the pixel translation between each to-be-matched point cloud and the reference point cloud, the accuracy and reliability of the determined moving velocity can be ensured.

[0157] Figure 3 is a structural schematic diagram of some obstacle moving velocity determination apparatus disclosed according to an embodiment of the present disclosure, which can be applied to the case of measuring the velocity of an obstacle by using a point cloud. The apparatus of the present embodiment can be implemented by software and / or hardware, and can be integrated on any electronic device with computing capability.

[0158] As shown in Figure 3 , the obstacle moving velocity determination apparatus 30 disclosed in the present embodiment can include a first point cloud determination module 31, a second point cloud determination module 32, a projection image determination module 33, a pixel translation determination module 34, and a moving velocity determination module 35, wherein:

[0159] The first point cloud determination module 31 is configured to determine at least two frames of target point clouds from candidate point clouds; wherein the candidate point clouds represent point clouds collected for a target obstacle;

[0160] The second point cloud determination module 32 is configured to determine a reference point cloud and at least one frame of to-be-matched point clouds from the at least two frames of target point clouds according to the point cloud information quantity of the target point clouds; wherein the point cloud information quantity is used to measure the richness of image information contained in the target point clouds;

[0161] The projection image determination module 33 is configured to take the two-dimensional projection image corresponding to the reference point cloud and the two-dimensional projection image corresponding to the to-be-matched point cloud as a reference projection image and a to-be-matched projection image, respectively;

[0162] The pixel translation determination module 34 is configured to perform matched filtering on the reference projection image and the to-be-matched projection image, determine a matched filtering result, and determine the pixel translation between the reference projection image and the to-be-matched projection image according to the matched filtering result;

[0163] The moving speed determination module 35 is configured to determine the moving speed of the target obstacle according to the pixel point translation amount.

[0164] Optionally, the first point cloud determination module 31 is specifically configured to:

[0165] determine a current point cloud collected at a current time and a preset number of historical point clouds collected at historical times from the candidate point cloud;

[0166] determine the target point cloud according to the current point cloud and the historical point clouds.

[0167] Optionally, the preset number is negatively correlated with a point cloud collection distance and a point cloud collection quantity, the point cloud collection distance is a distance between a point cloud collection device and a region where the target obstacle is located, and the point cloud collection quantity is a total number of collected points in the candidate point cloud.

[0168] Optionally, the device further includes a first point cloud information quantity determination module, which is specifically configured to:

[0169] take a two-dimensional projection image corresponding to the target point cloud as a target projection image, perform grid division on the target projection image, and determine at least one target grid image included in the target projection image;

[0170] take the target grid image containing the projection point as an auxiliary grid image, and determine the point cloud information quantity according to an image quantity of the auxiliary grid image.

[0171] Optionally, the first point cloud information quantity determination module is further configured to:

[0172] obtain a moving speed precision threshold, a distance resolution of the point cloud collection device, and a first collection time interval between the candidate point clouds;

[0173] determine a first parameter value according to a product result between the distance resolution and a constant factor, and determine a second parameter value according to a product result between the moving speed precision threshold, the first collection time interval and the constant factor;

[0174] determine a target grid size from the first parameter value and the second parameter value according to a size relationship between the first parameter value and the second parameter value, and perform grid division on the target projection image by using the target grid size.

[0175] Optionally, the device further includes a second point cloud information quantity determination module, which is specifically configured to:

[0176] determine a maximum reflection intensity and a minimum reflection intensity of the collected points in the target point cloud;

[0177] determine a reflection intensity difference value according to the maximum reflection intensity and the minimum reflection intensity, and determine the point cloud information quantity according to the reflection intensity difference value.

[0178] Optionally, the second point cloud information quantity determination module is specifically configured to:

[0179] determine the sub quantity of the collection points in the target point cloud;

[0180] determine the point cloud information quantity according to a weighted result between the sub quantity, the image quantity, and the reflection intensity difference value.

[0181] Optionally, the second point cloud determination module 32 is specifically configured to:

[0182] determine a sorting result of the point cloud information quantity, and determine the reference point cloud from the target point cloud according to the sorting result;

[0183] take the target point cloud other than the reference point cloud as the to-be-matched point cloud.

[0184] Optionally, the pixel point translation quantity determination module 34 is specifically configured to:

[0185] perform Fourier transform on the reference projection image to determine a reference frequency spectrum image, and perform Fourier transform on the to-be-matched projection image to determine a to-be-matched frequency spectrum image;

[0186] determine a frequency spectrum image product result between the reference frequency spectrum image and the to-be-matched frequency spectrum image;

[0187] perform inverse Fourier transform on the frequency spectrum image product result, and determine a matched filtering result according to a result of the inverse Fourier transform.

[0188] Optionally, the pixel point translation quantity determination module 34 is specifically configured to:

[0189] obtain an envelope diagram corresponding to the matched filtering result, and determine a peak point in the envelope diagram;

[0190] determine the pixel point translation quantity according to an abscissa value of the peak point in the envelope diagram.

[0191] Optionally, the moving speed determination module 35 is specifically configured to:

[0192] determine a second collection time interval between the reference point cloud and the to-be-matched point cloud;

[0193] perform linear fitting to generate a fitting line segment according to the pixel point translation quantity and the second collection time interval, and determine the moving speed according to a first-order coefficient of the fitting line segment.

[0194] The obstacle moving speed determination apparatus 30 disclosed in the embodiments of the present disclosure can perform the obstacle moving speed determination method disclosed in the embodiments of the present disclosure, and has the function modules and beneficial effects corresponding to the performing method. The contents not described in detail in the embodiments can be referred to the description in the method embodiments of the present disclosure.

[0195] In the technical solutions of the present disclosure, the acquisition, storage and application of user personal information comply with relevant laws and regulations and do not violate public order and good customs.

[0196] According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0197] Figure 4 A schematic block diagram of an example electronic device 400 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present disclosure described and / or claimed in this document.

[0198] As shown in Figure 4 The device 400 includes a computing unit 401 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 402 or a computer program loaded into a random access memory (RAM) 403 from a storage unit 408. Various programs and data required for the operation of the device 400 can also be stored in the RAM 403. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0199] Various components in the device 400 are connected to the I / O interface 405, including an input unit 406, such as a keyboard, a mouse, etc., an output unit 407, such as various types of displays, speakers, etc., a storage unit 408, such as a magnetic disk, an optical disk, etc., and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0200] The computing unit 401 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs various methods and processes described above, such as the determination method of the obstacle moving speed. For example, in some embodiments, the determination method of the obstacle moving speed can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the determination method of the obstacle moving speed described above can be performed. Alternatively, in other embodiments, the computing unit 401 can be configured to perform the determination method of the obstacle moving speed by any other appropriate means, such as by means of firmware.

[0201] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0202] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces a means for implementing the functions / acts specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0203] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0204] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0205] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0206] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0207] It should be understood that the various forms of flow shown above can be used to reorder, add, or delete steps. For example, the steps described in the present disclosure can be performed in parallel, in series, or in a different order, as long as the desired results of the technology disclosed in the present disclosure can be achieved, which is not limited herein.

[0208] The above detailed description does not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for determining a moving speed of an obstacle, comprising: determining at least two target point clouds from candidate point clouds, wherein the candidate point clouds represent point clouds collected from a target obstacle; determining a reference point cloud and at least one to-be-matched point cloud from the at least two target point clouds according to a point cloud information quantity of the target point clouds, wherein the point cloud information quantity is used to measure a richness of point cloud information contained in the target point clouds; taking a two-dimensional projection image corresponding to the reference point cloud and a two-dimensional projection image corresponding to the to-be-matched point cloud as a reference projection image and a to-be-matched projection image, respectively; performing matching filtering according to the reference projection image and the to-be-matched projection image, determining a matching filtering result, and determining a pixel point translation amount between the reference projection image and the to-be-matched projection image according to the matching filtering result; determining the moving speed of the target obstacle according to the pixel point translation amount. Before the determining a reference point cloud and at least one to-be-matched point cloud from the at least two target point clouds according to a point cloud information quantity of the target point clouds, the method further comprises: taking a two-dimensional projection image corresponding to the target point cloud as a target projection image, and obtaining a moving speed accuracy threshold, a distance resolution of a point cloud collecting device, and a first collection time interval between the candidate point clouds; determining a first parameter value according to a product result between the distance resolution and a constant factor, and determining a second parameter value according to a product result between the moving speed accuracy threshold, the first collection time interval and the constant factor; determining a target grid size from the first parameter value and the second parameter value according to a size relationship between the first parameter value and the second parameter value, and performing grid division on the target projection image by using the target grid size to determine at least one target grid image included in the target projection image; taking the target grid image containing a projection point as an auxiliary grid image, and determining the point cloud information quantity according to an image quantity of the auxiliary grid image.

2. The method of claim 1, wherein, The determining at least two target point clouds from candidate point clouds comprises: determining a current point cloud collected at a current time and a preset number of historical point clouds collected at historical times from the candidate point clouds; determining the target point clouds according to the current point cloud and the historical point clouds.

3. The method of claim 2, wherein, The preset number is negatively correlated with a point cloud collecting distance and a point cloud collecting quantity, the point cloud collecting distance is a distance between a point cloud collecting device and a region where the target obstacle is located, and the point cloud collecting quantity is a total number of collected points in the candidate point clouds. 4.The method of claim 1, before the determining a reference point cloud and at least one to-be-matched point cloud from the at least two target point clouds according to a point cloud information quantity of the target point clouds, further comprising: determining a maximum reflection intensity and a minimum reflection intensity of collected points in the target point clouds; determining a reflection intensity difference value according to the maximum reflection intensity and the minimum reflection intensity, and determining the point cloud information quantity according to the reflection intensity difference value.

5. The method of claim 4, wherein, The determining the point cloud information quantity according to the reflection intensity difference value comprises: determining a sub-number of the collected points in the target point clouds. determine the point cloud information quantity according to a weighted result of the sub-quantity, the image quantity, and the reflection intensity difference value.

6. The method of claim 1, wherein, The determining the reference point cloud and the at least one frame of point cloud to be matched from the at least two frames of target point clouds according to the point cloud information quantity of the target point cloud comprises: determining a sorting result of the point cloud information quantity, and determining the reference point cloud from the target point clouds according to the sorting result; taking the target point clouds other than the reference point cloud as the point cloud to be matched.

7. The method of claim 1, wherein, The matching filtering according to the reference projection image and the projection image to be matched to determine a matching filtering result comprises: performing Fourier transform on the reference projection image to determine a reference frequency spectrum image, and performing Fourier transform on the projection image to be matched to determine a frequency spectrum image to be matched; determining a frequency spectrum image product result between the reference frequency spectrum image and the frequency spectrum image to be matched; performing inverse Fourier transform on the frequency spectrum image product result, and determining the matching filtering result according to the inverse Fourier transform result.

8. The method of claim 1, wherein, The determining the pixel point translation between the reference projection image and the projection image to be matched according to the matching filtering result comprises: obtaining an envelope graph corresponding to the matching filtering result, and determining a peak point in the envelope graph; determining the pixel point translation according to an abscissa value of the peak point in the envelope graph.

9. The method of claim 1, wherein, The determining the moving speed of the target obstacle according to the pixel point translation comprises: determining a second acquisition time interval between the reference point cloud and the point cloud to be matched; performing linear fitting to generate a fitting line segment according to the pixel point translation and the second acquisition time interval, and determining the moving speed according to a first-order coefficient of the fitting line segment.

10. An apparatus for determining a moving speed of an obstacle, comprising: a first point cloud determining module configured to determine at least two frames of target point clouds from candidate point clouds; wherein the candidate point clouds represent point clouds collected from a target obstacle; a second point cloud determining module configured to determine a reference point cloud and at least one frame of point cloud to be matched from the at least two frames of target point clouds according to a point cloud information quantity of the target point clouds; wherein the point cloud information quantity is used to measure the richness of image information contained in the target point clouds; a projection image determining module configured to take a two-dimensional projection image corresponding to the reference point cloud and a two-dimensional projection image corresponding to the point cloud to be matched as a reference projection image and a projection image to be matched, respectively; a pixel point translation determining module configured to perform matching filtering according to the reference projection image and the projection image to be matched to determine a matching filtering result, and determine a pixel point translation between the reference projection image and the projection image to be matched according to the matching filtering result; a moving speed determining module configured to determine a moving speed of the target obstacle according to the pixel point translation; a first point cloud information quantity determining module configured to: take a two-dimensional projection image corresponding to the target point clouds as a target projection image, and obtain a moving speed precision threshold, a distance resolution of a point cloud acquisition device, and a first acquisition time interval between the candidate point clouds; determine a first parameter value according to a product result between the distance resolution and a constant factor, and determine a second parameter value according to a product result between the moving speed precision threshold, the first acquisition time interval and a constant factor; determine a target grid size from the first parameter value and the second parameter value according to a size relationship between the first parameter value and the second parameter value, and perform grid division on the target projection image by using the target grid size to determine at least one target grid image included in the target projection image; determine the target grid image containing the projection point as an auxiliary grid image, and determine the point cloud information quantity according to an image quantity of the auxiliary grid image.

11. The apparatus of claim 10, wherein, The first point cloud determination module is specifically configured to: determine a current point cloud collected at a current time and a preset quantity of historical point clouds collected at historical times from the candidate point cloud; and determine the target point cloud according to the current point cloud and the historical point clouds.

12. The apparatus of claim 11, wherein, The preset quantity is negatively correlated with a point cloud acquisition distance and a point cloud acquisition quantity, the point cloud acquisition distance is a distance between a point cloud acquisition device and a region where the target obstacle is located, and the point cloud acquisition quantity is a total number of collected points in the candidate point cloud.

13. The apparatus of claim 10, wherein, The device further includes a second point cloud information quantity determination module, which is specifically configured to: determine a maximum reflection intensity and a minimum reflection intensity of collected points in the target point cloud; and determine a reflection intensity difference value according to the maximum reflection intensity and the minimum reflection intensity, and determine the point cloud information quantity according to the reflection intensity difference value.

14. The apparatus of claim 13, wherein, The second point cloud information quantity determination module is specifically further configured to: determine a sub quantity of the collected points in the target point cloud; and determine the point cloud information quantity according to a weighted result among the sub quantity, the image quantity and the reflection intensity difference value.

15. The apparatus of claim 10, wherein, The second point cloud determination module is specifically configured to: determine a sorting result of the point cloud information quantity, and determine a reference point cloud from the target point cloud according to the sorting result; and determine the target point cloud except the reference point cloud as the to-be-matched point cloud.

16. The apparatus of claim 10, wherein, The pixel point translation amount determination module is specifically configured to: perform Fourier transform on the reference projection image to determine a reference frequency spectrum image, and perform Fourier transform on the to-be-matched projection image to determine a to-be-matched frequency spectrum image; determine a frequency spectrum image product result between the reference frequency spectrum image and the to-be-matched frequency spectrum image; and perform inverse Fourier transform on the frequency spectrum image product result, and determine a matching filtering result according to an inverse Fourier transform result. The pixel point translation amount determination module is specifically further configured to: obtain an envelope graph corresponding to the matching filtering result, and determine a peak point in the envelope graph; and 17. The apparatus of claim 10, wherein, determine the pixel point translation amount according to an abscissa value of the peak point in the envelope graph. The moving speed determination module is specifically configured to: determine a second acquisition time interval between the reference point cloud and the to-be-matched point cloud; and 18. The apparatus of claim 10, wherein, perform straight line fitting to generate a fitting line segment according to the pixel point translation amount and the second acquisition time interval, and determine the moving speed according to a first order coefficient of the fitting line segment.

19. An electronic device, comprising: at least one processor; and ​ ​ ​ a memory in communication with the at least one processor; wherein the memory has stored instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

20. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, the computer instructions are for causing the computer to perform the method of any one of claims 1-9.

21. A computer program product comprising a computer program which, when executed by a processor, implements the method of any one of claims 1-9.

Images