Vehicle driving area detection method, program product and device

Abstract

The invention discloses a vehicle driving area detection method, a program product and a storage medium. The method comprises the following steps: constructing an original map corresponding to a current frame indicating a driving area; obtaining current point cloud data collected in a current frame, obtaining current driving data, and predicting a driving trajectory based on the current driving data; forming a reference line based on the driving trajectory, and calculating a left barrier projection peak value and a right barrier projection peak value of each map unit based on the current point cloud data and the reference line; determining a left side obstacle point cloud set of each map unit based on the left side obstacle projection peak value of each map unit, and determining a right side obstacle point cloud set of each map unit based on the right side obstacle projection peak value of each map unit; and based on the left side obstacle point cloud set and the right side obstacle point cloud set of each map unit, obtaining a left side driving area boundary and a right side driving area boundary of each map unit and a current map corresponding to the current frame.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a vehicle driving area detection method, a computer program product and a vehicle device. Background Art

[0002] In the field of vehicle driving technology, autonomous driving perception applications play an increasingly important role. For example, visual perception and laser radar can perceive the environment in front of the vehicle. Millimeter-wave radar is also used in the prior art to perceive the environment. Compared with visual perception and laser radar, millimeter-wave radar has the advantages of low hardware cost and ignoring the influence of weather and environmental factors. However, the electromagnetic waves emitted by millimeter-wave radar are easily reflected by strong reflective targets in the environment, thereby forming a large number of false point clouds. In order to remove these false point clouds, the drivable area is usually identified through the point cloud of the millimeter-wave radar, and the detected point cloud is marked whether it falls outside the drivable area, so as to identify the false point cloud to reduce the formation of false ghost targets. However, the drivable area identification scheme in the prior art is prone to the target being mislabeled, thereby affecting the accuracy of the drivable area identification, and the drivable area identification scheme in the prior art is prone to large space and time consumption as the area of ​​interest in front of the vehicle expands. Summary of the invention

[0003] In order to solve the existing technical problems, the present invention provides a vehicle driving area detection method, a computer program product and a vehicle device, which can reduce the amount of calculation, improve the real-time calculation speed, and update the original map constructed according to the current driving data in real time, which can improve the accuracy of drivable area identification.

[0004] In a first aspect, a vehicle driving area detection method is provided, comprising: constructing an original map corresponding to a current frame indicating a driving area, wherein the original map comprises a plurality of map units; acquiring current point cloud data collected in the current frame, and acquiring current driving data of the vehicle device, and predicting a driving trajectory of the vehicle device based on the current driving data; forming a reference baseline based on the driving trajectory, and calculating a left obstacle projection peak value of each map unit and a right obstacle projection peak value of each map unit based on the current point cloud data and the reference baseline; determining a left obstacle point cloud set of each map unit based on the left obstacle projection peak value of each map unit, and determining a right obstacle point cloud set of each map unit based on the right obstacle projection peak value of each map unit; updating the original map based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit to obtain a left driving area boundary and a right driving area boundary of each map unit, and generating a current map corresponding to the current frame based on the left driving area boundary and the right driving area boundary of each map unit.

[0005] In a second aspect, a computer program product is provided, including a computer program, characterized in that when the computer program is executed by a processor, the vehicle driving area detection method as described in any embodiment of the present application is implemented.

[0006] In a third aspect, a vehicle device is provided, characterized in that it includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the vehicle driving area detection method as described in any embodiment of the present application.

[0007] In the embodiment of the present application, during the driving process of the vehicle equipment, an original map indicating the driving area is first constructed, the original map includes a plurality of map units, and the driving trajectory of the vehicle equipment is predicted based on the current driving data obtained, and a reference baseline is formed based on the driving trajectory, so that the front area of ​​the vehicle equipment can be accurately divided into a left area and a right area, and then the left obstacle projection peak value of each map unit and the right obstacle projection peak value of each map unit are calculated, and then the left obstacle point cloud set is determined based on the point cloud corresponding to the horizontal coordinate near the left obstacle projection peak value, and the right obstacle point cloud set is determined based on the point cloud corresponding to the horizontal coordinate near the right obstacle projection peak value; then, based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit, the left driving area boundary and the right driving area boundary of each map unit are obtained. In the entire calculation process, there is no need to perform complex vector operations, thereby reducing the amount of operations and improving the real-time calculation speed; and the original map constructed by updating the current driving data in real time can improve the accuracy of drivable area recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 1. is an application environment diagram of a vehicle driving area detection method in an embodiment;

[0009] Figure 2 is a schematic diagram of a vehicle coordinate system in one embodiment;

[0010] Figure 3 is a schematic diagram of a vehicle coordinate system in another embodiment;

[0011] Figure 4 is a flow chart of a vehicle driving area detection method in one embodiment;

[0012] Figure 5 is a schematic diagram of a drivable area in an application scenario in an application scenario in an embodiment;

[0013] Figure 6 It is a flow chart of calculating the left obstacle projection peak value and the right obstacle projection peak value in the vehicle driving area detection method in one embodiment;

[0014] Figure 7 is a schematic diagram of lateral coordinate distribution data of a point cloud in one embodiment;

[0015] Figure 8 A flowchart of determining a right obstacle point cloud set of each map unit in a vehicle driving area detection method in one embodiment;

[0016] Fig. 9 is a schematic diagram of a vehicle driving area detection device in one embodiment;

[0017] Fig.10 Schematic diagram of vehicle equipment in one embodiment. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art of the technical field of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the scope of protection of the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0020] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it should be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0021] See also Figure 1 , is an application environment diagram of a vehicle driving area detection method in an embodiment. The vehicle driving area detection method is applied to a vehicle device 10, and the vehicle device 10 includes a display terminal 11, a radar device 12, and a controller 13. The radar device 12 is used to collect point cloud data of the front environment during the driving of the vehicle device 10, and the controller 13 generates a map indicating the drivable area according to the point cloud data collected by the radar device 12, so that the vehicle device 10 can accurately drive according to the map.

[0022] The radar device 12 includes but is not limited to a combination of one or more of the following sensors: 4D millimeter wave radar, 3D millimeter wave radar, other types of three-dimensional or four-dimensional radar devices. The radar device 12 is installed at the front center of the vehicle equipment 10 and faces the front of the vehicle equipment 10 to collect environmental data in front of the vehicle equipment 10. For example, the radar device 12 is installed at the center of the front bumper of the vehicle equipment 10. It is understandable that the vehicle equipment 10 may also include other types of sensors, including but not limited to multi-spectral vision sensors, environmental perception sensors, and motion posture sensors. The multi-spectral sensor includes but is not limited to a combination of one or more of the following sensors: thermal imaging sensors, visible light image sensors, millimeter wave sensors, lidar sensors, infrared thermal imaging sensors, and depth sensors. Environmental perception sensors include but are not limited to one or more combinations of the following sensors, such as brightness sensors, temperature sensors, and haze sensors. Motion posture sensors include but are not limited to one or more of the following combinations: inertial sensor (Inertial Measurement Unit, IMU), velocity sensor, acceleration sensor, gyroscope sensor, geomagnetic sensor, rotation vector sensor, steering wheel angle sensor, level sensor, tilt sensor, vibration sensor, displacement sensor and gravity sensor, etc.

[0023] The screen is the surface on which the display terminal 11 displays images. The screen is arranged in front of the driver so that the driver can view the images. The display terminal 11 includes but is not limited to displays such as liquid crystal displays, projectors with projector screens, etc. The display terminal 11 can be a display located on the central control of the vehicle device 10, or a projector arranged at the rear of the vehicle device 10 to project images onto the projector screen.

[0024] The controller 13 may be one or more. When there are multiple controllers 13, the multiple controllers may be integrated on one chip or independently set on each chip. The vehicle device 10 is a device installed on any type of mobile body, such as a vehicle, an electric vehicle, a hybrid electric vehicle, a motorcycle, a bicycle, a personal mobile device, an airplane, a drone, a ship or a robot, etc. In some embodiments, the controller 13 and the display terminal 11 may be integrated.

[0025] The prior art scheme for dividing the drivable area based on millimeter-wave radar is to sort the point cloud collected by the millimeter-wave radar according to the azimuth angle to divide a fan-shaped drivable area, or to extract the map boundary based on the M*N grid map and divide the drivable area according to the map boundary. Among them, the angle division scheme has the problem that the longer the distance at the same angle, the longer the arc length, and the target crossing the intersection is prone to occlusion and mismarking. The M*N grid map scheme will consume a lot of space and time as the constructed region of interest (ROI) expands. Therefore, the vehicle driving area detection method provided in the embodiment of the present application solves the above problems.

[0026] In order to facilitate the subsequent description of the embodiments of the present application, the vehicle coordinate system is first described. Figure 2 As shown, it is a schematic diagram of the vehicle coordinate system in one embodiment, the x-axis direction is the longitudinal direction, which is along the center line of the vehicle device 10 and points to the forward direction of the vehicle device. The y-axis direction is the transverse direction, which is perpendicular to the longitudinal direction and located on the same plane, and the y-axis direction points from the right side of the vehicle device to the left side. The z-axis direction is the pitch direction, that is, the direction perpendicular to the x-axis and the y-axis, and points from the bottom side of the vehicle device to the top side. It can be understood that the coordinate origin position can be set at different positions as needed. Figure 3 As shown, Figure 3 Schematic diagram of a vehicle coordinate system in another embodiment, the radar device is arranged at the center of the front bumper of the vehicle equipment, and the coordinate origin position is arranged at the center of the front bumper.

[0027] See also Figure 4 , is a flow chart of a vehicle driving area detection method provided in an embodiment of the present application. The vehicle driving area detection method is applied to a vehicle device, and the vehicle driving area detection method includes the following steps:

[0028] S11. Construct an original map corresponding to a current frame indicating a driving area, wherein the original map includes a plurality of map units.

[0029] In this embodiment, during the driving of the vehicle device, there are roads and obstacles in the drivable area in the area in front of the vehicle device. Obstacles include obstacles on both sides of the road and obstacles in the road, including but not limited to guardrails, trees, stationary vehicles, etc. The original map indicates a map containing a drivable area initially created based on the area in front of the current frame. In one implementation, the original map is divided into a plurality of map units in the longitudinal direction, and each map unit corresponds to a left initial driving area boundary and a right initial driving area boundary. However, since the environment of the area in front changes rapidly during the real-time driving of the vehicle device, the created original map needs to be updated based on the point cloud data in the current frame and the point cloud data in the historical frame to obtain the current map of the current frame.

[0030] S12, obtaining the current point cloud data collected in the current frame, and obtaining the current driving data of the vehicle equipment, and predicting the driving trajectory of the vehicle equipment based on the current driving data.

[0031] In this embodiment, during the driving process of the vehicle device, the radar device transmits electromagnetic waves forward, which are reflected after encountering obstacles to form echo signals, and the radar device receives the echo signals and generates point cloud data based on the echo signals. The current point cloud data indicates the point cloud data generated based on the echo signals in the current frame.

[0032] The current driving data indicates data directly collected by one or more sensors, or calculated based on the collected data, including but not limited to the current driving speed and the current yaw rate. The current driving speed indicates the driving speed of the vehicle device in the current frame. The current yaw rate indicates the yaw angular velocity in the current frame. The steering data of the vehicle device can be determined by the current yaw rate, such as whether the vehicle device has oversteering or understeering, etc. Therefore, the driving trajectory of the vehicle device can be accurately predicted based on the current driving data, so that the left and right sides of the vehicle device can be accurately divided later, thereby improving the accuracy of the map of the drivable area.

[0033] S13, forming a reference baseline based on the driving trajectory line, and calculating a left obstacle projection peak value of each map unit and a right obstacle projection peak value of each map unit based on the current point cloud data and the reference baseline.

[0034] In this embodiment, the driving trajectory line indicates the driving route of the vehicle device, and the driving trajectory line can be used as a reference baseline, or the driving trajectory line can be preprocessed to obtain the reference baseline, wherein the preprocessing includes but is not limited to smoothing the driving trajectory line, etc. Therefore, based on the reference baseline, the current point cloud in any map unit can be divided into a left area point cloud and a right area point cloud, wherein the left area point cloud is the point cloud located on the left side of the reference baseline, and the right area point cloud is the point cloud located on the right side of the reference baseline.

[0035] The left obstacle projection peak indicates the position of the left obstacle point cloud in the lateral coordinate concentration, and the right obstacle projection peak indicates the position of the right obstacle point cloud in the lateral coordinate concentration. Based on the point cloud data of the left area point cloud, the left obstacle projection peak can be calculated. Based on the point cloud data of the right area point cloud, the right obstacle projection peak can be calculated.

[0036] S14. Determine a left obstacle point cloud set of each map unit based on a left obstacle projection peak value of each map unit, and determine a right obstacle point cloud set of each map unit based on a right obstacle projection peak value of each map unit.

[0037] In this embodiment, for a map unit, after determining the left obstacle projection peak, the left obstacle point cloud set can be determined based on the point cloud corresponding to the horizontal coordinate near the left obstacle projection peak. The right obstacle point cloud set can be determined based on the point cloud corresponding to the horizontal coordinate near the right obstacle projection peak. After determining the left obstacle point cloud set and the right obstacle point cloud set of each map unit, the left driving area boundary and the right driving area boundary can be determined later.

[0038] S15. Based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit, the original map is updated to obtain the left driving area boundary and the right driving area boundary of each map unit, and based on the left driving area boundary and the right driving area boundary of each map unit, a current map corresponding to the current frame is generated.

[0039] In this embodiment, due to the speed or road during the driving of the vehicle equipment, the environment in the area ahead of the vehicle equipment changes rapidly, so the constructed original map needs to be updated. The left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit are obtained by partitioning the front area according to the predicted driving trajectory line, and can accurately represent the left obstacle area and the right obstacle area in the front area. Therefore, for a map unit, based on the left obstacle point cloud set of the map unit and the right obstacle point cloud set of the map unit, the left driving area boundary and the right driving area boundary of the map unit can be traversed. After obtaining the left driving area boundary and the right driving area boundary of each map unit, the drivable area can be accurately divided to obtain a more accurate current map, thereby improving driving safety.

[0040] like Figure 5 As shown, Figure 5 is a schematic diagram of a drivable area in an application scenario in an embodiment, Figure 5 The lower right picture in the figure is an application scenario picture. Figure 5 The left side of is a map obtained by using the vehicle driving area detection method of the embodiment of the present application for the application scene picture, and the map includes the left driving area boundary and the right driving area boundary. The upper side of the application scene picture is the point cloud data corresponding to the application scene picture.

[0041] In the above embodiment, during the driving process of the vehicle device, an original map indicating the driving area is first constructed, and the original map includes a plurality of map units. Based on the current driving data obtained, the driving trajectory of the vehicle device is predicted, and a reference baseline is formed based on the driving trajectory, so that the front area of ​​the vehicle device can be accurately divided into a left area and a right area. Then, the left obstacle projection peak value of each map unit and the right obstacle projection peak value of each map unit are calculated, and then the left obstacle point cloud set is determined based on the point cloud corresponding to the horizontal coordinate near the left obstacle projection peak value, and the right obstacle point cloud set is determined based on the point cloud corresponding to the horizontal coordinate near the right obstacle projection peak value; then, based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit, the left driving area boundary and the right driving area boundary of each map unit are obtained. In the entire calculation process, there is no need to perform complex vector operations, so that the amount of operations can be reduced and the real-time calculation speed can be improved; and the original map constructed by updating the real-time according to the current driving data can improve the accuracy of the drivable area recognition.

[0042] In some embodiments, constructing an original map corresponding to the current frame indicating the driving area includes:

[0043] Acquire a region of interest in front of the vehicle equipment;

[0044] Based on the division step, the region of interest is divided into a plurality of map units in a longitudinal direction, wherein the longitudinal direction is along a center line of the vehicle equipment and points to a forward direction of the vehicle equipment, and each map unit includes a left initial driving area boundary and a right initial driving area boundary.

[0045] In this embodiment, the region of interest is a sector area set according to the detection range of the radar device. For example, the radar device can detect a sector area with a radius of 300m. For autonomous driving, the area within 20m on both sides of the vehicle equipment in the lateral direction and within 150m in the longitudinal direction is the region of interest. The range of the region of interest can be set and adjusted as needed.

[0046] Based on the division step length, the region of interest is divided into multiple map units in the longitudinal direction. Each map unit has a left initial driving area boundary and a right initial driving area boundary, wherein the division step length can be set according to the resolution, that is, a fixed step length. For example, the range of 150M in front of the vehicle equipment can be divided into 150 longitudinal regions according to the resolution of 1m, and the longitudinal length of each region is 1M, and the horizontal length is within 20m on both sides. In this way, the region of interest is divided into 150*1 map units. Of course, this division step length can be other values, such as 2m, 3m, 5m, and the division step length can be a uniform resolution or a non-uniform resolution. For example, autonomous driving has higher requirements for close-range targets, and is divided according to a resolution of 1m within 100m. The long-distance requirements are not so high, and can be divided according to a resolution of 5m. In this way, the longitudinal lengths of each map unit of the original map obtained can be different. From the map units constructed above, it can be seen that each map unit corresponds to a longitudinal position and the regional range of the map unit. The corresponding map unit can be queried according to the longitudinal position without traversing the entire map, thereby reducing the amount of calculation.

[0047] In the above embodiment, this method of dividing the map in the longitudinal direction can effectively solve the problem of obstruction of crossing targets in complex environments (such as intersections); and the resolution of each map unit in the front area of ​​the vehicle equipment is not much different; each map unit corresponds to a longitudinal position and the area range of the map unit, and the corresponding map unit can be queried according to the longitudinal position without traversing the entire map, thereby reducing the amount of calculation.

[0048] In some embodiments, the current driving data includes a current driving speed and a current yaw rate. Before predicting a driving trajectory of the vehicle device based on the current driving data, the method further includes:

[0049] Determining whether current conditions meet map update conditions based on the current driving data;

[0050] When it is determined based on the current driving data that the current condition satisfies the map update condition, predicting a driving trajectory of the vehicle device based on the current driving data;

[0051] When the current driving speed is greater than a preset speed and the current yaw rate is less than a first preset yaw rate, determining that the current driving data determines that the original map meets a map update condition;

[0052] When the current driving speed is less than or equal to the preset speed and the status indicator of the original map indicates a valid state, continue to use the original map;

[0053] When the current yaw rate is less than a first preset yaw rate and the state indicator of the original map indicates a valid state, continue to use the original map;

[0054] When the current yaw rate is greater than or equal to the first preset yaw rate, the state mark of updating the original map is invalid.

[0055] In this embodiment, the map update condition includes that the current driving speed is greater than the preset speed and the current yaw rate is less than the first preset yaw rate. When it is determined based on the current driving data that the current condition meets the map update condition, it means that the original map cannot provide an accurate drivable area, and the original map needs to be updated. The current driving data of the vehicle equipment is continued to be obtained, and the driving trajectory of the vehicle equipment is predicted based on the current driving data.

[0056] The map corresponds to a map status indicator, which is used to indicate whether the map can be used. The map status indicator is updated in real time based on the current driving data and the current point cloud data collected. When the current driving speed is less than or equal to the preset speed, and the status indicator of the original map indicates a valid state, or when the current yaw rate is less than the first preset yaw rate, and the status indicator of the original map indicates a valid state, it means that the environment in front of the vehicle equipment here has not changed much, and the original map can continue to be used. When the current yaw rate is greater than or equal to the first preset yaw rate, it means that the vehicle equipment here has turned a lot, and the environment in front of the vehicle equipment has changed a lot, and it is necessary to re-acquire the map, and update the status indicator of the original map to an invalid state.

[0057] In the above embodiment, during the vehicle driving process, it is determined in real time whether the current conditions meet the map update conditions. If the map update conditions are met, the original map is updated, and the map status mark can be updated in real time based on the current driving data to improve the accuracy of subsequent drivable area identification.

[0058] In some embodiments, predicting the driving trajectory of the vehicle device based on the current driving data includes:

[0059] Calculating a reference benchmark factor based on the current driving speed and the current yaw rate;

[0060] Based on the reference benchmark factor, the driving trajectory line is calculated using a trajectory prediction formula;

[0061] The calculation formula of the reference benchmark factor is as follows:

[0062]

[0063] The trajectory prediction formula is as follows:

[0064] y=0.5*C*x 2 ,

[0065] Among them, yawRate represents the current yaw rate, thre represents the second preset yaw rate, C represents the road curvature, CC represents the reference benchmark factor, vehicleSpeed ​​represents the current driving speed, x represents the distance that the vehicle device has traveled forward in the current frame relative to the previous preset frame of the current frame, and y represents the lateral offset distance of the vehicle device in the current frame relative to the previous preset frame of the current frame.

[0066] In this embodiment, the driving trajectory is approximately predicted based on the road curve in front of the vehicle equipment, wherein the road curve is an approximate parabolic curve. Therefore, by discarding the higher-order terms of the parabolic curve, a trajectory prediction formula can be obtained, which can approximately represent the driving trajectory within a certain distance range in front of the vehicle equipment.

[0067] In the above embodiment, the driving trajectory within a certain distance range in front of the vehicle equipment is predicted in real time based on the current driving data, which can facilitate more accurate acquisition of the left area and the right area of ​​the vehicle equipment in the subsequent process, so as to accurately obtain the left driving area boundary and the right driving area boundary.

[0068] In some embodiments, Figure 6 This is a flow chart of calculating the left obstacle projection peak value and the right obstacle projection peak value in the vehicle driving area detection method in one embodiment, and step S13 specifically includes:

[0069] S131. For any map unit, based on the current point cloud data, obtain a left stationary point cloud and a right stationary point cloud.

[0070] In this embodiment, after the current point cloud data is collected, the historical point cloud data and the identification of the point cloud can be combined to determine whether the current point cloud is a static point cloud. If the identification of a current point cloud indicates a static point cloud, then the current point cloud is a static point cloud. Static targets in the road environment, such as guardrails, logs, and stationary vehicles, do not have a moving speed. Therefore, in one implementation, the radial velocity of the point cloud to the ground can be calculated to be less than a preset radial velocity threshold. If the radial velocity of a point cloud to the ground is less than the preset radial velocity threshold, it can be considered to be a static point cloud. A static point cloud indicates a point cloud corresponding to a static target.

[0071] Based on the reference baseline, the left area point cloud and the right area point cloud can be obtained, the left static point cloud is obtained from the left area point cloud, and the right static point cloud is obtained from the right area point cloud.

[0072] S132, projecting the left stationary point cloud and the right stationary point cloud to the lateral direction respectively, to obtain lateral coordinate distribution data of the left obstacle and lateral coordinate distribution data of the right obstacle.

[0073] In this embodiment, the lateral coordinate distribution data indicates the distribution of the lateral coordinates of the point cloud. Figure 7 As shown, Figure 7 Schematic diagram of the lateral coordinate distribution data of the point cloud in one embodiment. It can be seen from the lateral coordinate distribution data that there may be a projection peak on the left or right side where obstacles are concentrated.

[0074] S133. Determine a left projection peak based on the lateral coordinate distribution data of the left obstacle, and determine a right projection peak based on the lateral coordinate distribution data of the right obstacle.

[0075] In this embodiment, for the lateral distribution data of the left obstacle, a grid is formed with a preset length, the lateral coordinates of the left area are divided into multiple grids, and the left projection peak is formed based on the horizontal coordinate corresponding to the grid with the most point clouds. For the lateral distribution data of the right obstacle, a grid is formed with a preset length, the lateral coordinates of the right area are divided into multiple grids, and the right projection peak is formed based on the horizontal coordinate corresponding to the grid with the most point clouds.

[0076] S134, determining whether the left projection peak satisfies the left peak condition.

[0077] When the left projection peak value meets the left peak value condition, S135 is executed to obtain the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and the left projection peak value and the left obstacle projection peak value corresponding to the previous preset frame are weighted to obtain the left obstacle projection peak value of any map unit. When the left projection peak value does not meet the left peak value condition, S136 is executed to obtain the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and based on the left obstacle projection peak value corresponding to the previous preset frame, the left obstacle projection peak value of any map unit is obtained.

[0078] S137, determining whether the right projection peak satisfies the right peak condition.

[0079] When the right projection peak value meets the right peak value condition, S138 is executed to obtain the right obstacle projection peak value corresponding to the previous preset frame of the current frame, and the right obstacle projection peak value and the right obstacle projection peak value corresponding to the previous preset frame are weighted to obtain the right obstacle projection peak value of any map unit. When the right projection peak value does not meet the left peak value condition, S139 is executed to obtain the right obstacle projection peak value corresponding to the previous preset frame of the current frame, and based on the right obstacle projection peak value corresponding to the previous preset frame, the right obstacle projection peak value of any map unit is obtained.

[0080] It is understandable that, in the above steps, S134 and S137 are executed in no particular order and can be executed simultaneously. For any map unit, by executing the above S131-S139, the corresponding left obstacle projection peak value and right obstacle projection peak value can be obtained.

[0081] In the above embodiment, the point cloud is projected in the lateral direction, and for the left or right side, the position with the most point clouds is statistically obtained, and the projection peak value of this side is obtained based on this position. In this calculation process, no complicated vector calculation is required, thereby reducing the amount of calculation and improving real-time performance; and based on the predicted driving trajectory line, a base reference baseline is formed to obtain more accurate point clouds in the left area and the right area, so as to obtain more accurate obstacle projection peaks on both sides.

[0082] In some embodiments, Figure 8 As shown, Figure 8 This is a flow chart of determining the right obstacle point cloud of each map unit in the vehicle driving area detection method in one embodiment, and step S14 specifically includes:

[0083] S141, obtaining point cloud data corresponding to a previous preset frame of the current frame, and based on the point cloud data corresponding to the previous preset frame, performing compensation correction on the current point cloud data to obtain corrected point cloud data.

[0084] In this embodiment, the following point cloud compensation correction formula is used to obtain the corrected point cloud data. The point cloud compensation correction formula is:

[0085]

[0086] xpos′, ypos′ respectively represent the longitudinal coordinate and transverse coordinate of the point cloud in the previous preset frame coordinate system, xpos, ypos respectively represent the longitudinal coordinate and transverse coordinate of the point cloud in the current frame coordinate system, and angle represents the angular rotation of the vehicle device due to displacement from the previous preset frame moment to the current frame moment. Since the vehicle device is in motion, there will be a displacement and rotation in the coordinate system at the current moment compared to the previous moment, which is equivalent to compensating the position of the point cloud in the coordinate system of the previous frame moment to the corresponding position in the current coordinate system.

[0087] S142. Determine a left obstacle point cloud candidate set for each map unit based on the corrected point cloud data and the left obstacle projection peak value of each map unit, and determine a right obstacle point cloud candidate set for each map unit based on the corrected point cloud data and the right obstacle projection peak value of each map unit.

[0088] In an optional implementation, the determining a left obstacle point cloud candidate set of each map unit based on the corrected point cloud data and a left obstacle projection peak value of each map unit, and the determining a right obstacle point cloud candidate set of each map unit based on the corrected point cloud data and a right obstacle projection peak value of each map unit, comprises:

[0089] For any map unit, based on the corrected point cloud data and the left obstacle projection peak value of the any map unit, a left stationary point cloud in the any map unit that meets the left screening condition is determined as a left obstacle point cloud candidate set of the any map unit, wherein the left screening condition includes that the pitch position of the point cloud is within a preset pitch interval range, and the lateral coordinates of the point cloud and the left obstacle projection peak value of the any map unit are within a preset left lateral error range;

[0090] Based on the corrected point cloud data and the right obstacle projection peak value of any map unit, the right stationary point cloud in any map unit that meets the right screening condition is determined as the right obstacle point cloud candidate set of any map unit, wherein the right screening condition includes that the pitch position of the point cloud is within a preset pitch interval, and the lateral coordinates of the point cloud and the right obstacle projection peak value of any map unit are within a preset right lateral error range.

[0091] In this embodiment, the pitch position of the point cloud is used to determine whether the point cloud can block the passage of the vehicle equipment. When the pitch position of the point cloud is within the preset pitch range, it means that the point cloud can block the passage of the vehicle equipment. Therefore, the candidate obstacle point cloud sets on both sides can be obtained by the pitch position of the point cloud and the lateral distance from the peak value of the obstacle projection on the left side of the point cloud or the peak value of the obstacle projection on the right side.

[0092] In an optional implementation, the determining a left obstacle point cloud candidate set of each map unit based on the corrected point cloud data and a left obstacle projection peak value of each map unit, and the determining a right obstacle point cloud candidate set of each map unit based on the corrected point cloud data and a right obstacle projection peak value of each map unit, comprises:

[0093] For any map unit, fitting the point cloud corresponding to the left obstacle projection peak of the any map unit to obtain a left obstacle fitting curve; determining the point cloud whose distance to the left obstacle fitting curve is within a preset left distance range as the left obstacle point cloud candidate set of the any map unit;

[0094] The point cloud corresponding to the right obstacle projection peak of any map unit is fitted to obtain a right obstacle fitting curve; and the point cloud whose distance to the right obstacle fitting curve is within a preset right distance range is determined as a right obstacle point cloud candidate set of any map unit.

[0095] S143: determining a left obstacle point cloud set of each map unit based on the left obstacle point cloud candidate set of each map unit, and determining a right obstacle point cloud set of each map unit based on the right obstacle point cloud candidate set of each map unit.

[0096] In an optional implementation, S143 specifically includes:

[0097] For any map unit, if the number of point clouds in the left obstacle point cloud candidate set of the any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the left obstacle point cloud candidate set of the any map unit to obtain the left obstacle point cloud set of the any map unit;

[0098] When the number of point clouds in the right obstacle point cloud candidate set of any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the right obstacle point cloud candidate set of any map unit to obtain the right obstacle point cloud set of any map unit, wherein the deletion condition includes that the distance between point clouds is less than the preset interval distance.

[0099] In this embodiment, the deletion condition can be used to delete the point clouds that are very close to each other in the left obstacle point cloud candidate set and the right obstacle point cloud candidate set, thereby reducing the subsequent calculation amount and avoiding the influence of too many close points on the subsequent calculation accuracy to improve the recognition accuracy of the driving area.

[0100] In an optional implementation, before executing S143, it is determined whether the left obstacle point cloud candidate set and / or the right obstacle point cloud candidate set are valid, by counting whether the number of point clouds in the left obstacle point cloud candidate set and / or the right obstacle point cloud candidate set exceeds the preset point cloud number threshold. If the left obstacle point cloud candidate set exceeds the preset point cloud number threshold, it means that the left obstacle point cloud candidate set is valid, and if the right obstacle point cloud candidate set exceeds the preset point cloud number threshold, it means that the right obstacle point cloud candidate set is valid. When the left obstacle point cloud candidate set and / or the right obstacle point cloud candidate set are valid, S143 is executed. When the left obstacle point cloud candidate set and the right obstacle point cloud candidate set are invalid, the process ends.

[0101] In the above embodiment, the position data of the point cloud in the current frame is corrected according to the point cloud of the historical frame, so as to obtain accurate point cloud position data, and then the left obstacle point cloud candidate set of each map unit is determined based on the corrected position data and the left obstacle projection peak of each map unit, and the right obstacle point cloud candidate set is determined based on the corrected point cloud data and the right obstacle projection peak of each map unit. Since the corrected point cloud position data is used and the base reference baseline is formed based on the predicted driving trajectory line, more accurate left area point cloud and right area point cloud can be obtained, which facilitates obtaining more accurate obstacle point cloud candidate sets on both sides, thereby improving the recognition accuracy of the drivable area.

[0102] In some embodiments, step S15 specifically includes:

[0103] Sort the point clouds in the left obstacle point cloud set of each map unit according to the longitudinal coordinate to obtain the left sorted point cloud set of each map unit, and determine the position of the point cloud with the smallest transverse coordinate from the left sorted point cloud set of each map unit as the left driving area boundary of each map unit, wherein the longitudinal coordinate is the coordinate in the longitudinal direction, the transverse coordinate is the coordinate in the transverse direction, the transverse direction is perpendicular to the longitudinal direction and located on the same plane, the positive direction of the transverse direction is toward the left side of the vehicle device, and the longitudinal direction is along the center line of the vehicle device and points to the forward direction of the vehicle device;

[0104] The point clouds in the right obstacle point cloud set of each map unit are sorted according to the longitudinal coordinates to obtain the right sorted point cloud set of each map unit, and the position of the point cloud with the largest transverse coordinate is determined from the right sorted point cloud set of each map unit as the right driving area boundary of each map unit;

[0105] The current map is generated based on the left driving area boundary of each map unit and the right driving area boundary of each map unit.

[0106] In this embodiment, for any map unit, by traversing the corresponding left obstacle point cloud set, the position of the point cloud with the smallest horizontal coordinate can be obtained, thereby obtaining the boundary of the left driving area; by traversing the corresponding right obstacle point cloud set, the position of the point cloud with the largest horizontal coordinate can be obtained, thereby obtaining the boundary of the right driving area. The specific traversal adopts the following formula:

[0107] boundaryYpos1=min(boundryYpos,detYpos),

[0108] boundaryYpos2=max(boundaryYpos,detYpos),

[0109] Wherein boundaryYpos1 represents the boundary of the left driving area, boundaryYpos represents the current boundary of the left driving area during the traversal process, detYpos represents the lateral coordinate of the traversed point cloud, and boundaryYpos2 represents the boundary of the right driving area.

[0110] In the above embodiment, for any map unit, after obtaining the left obstacle point cloud set and the right obstacle point cloud set, the left driving area boundary can be traversed based on the left obstacle point cloud set, and the right driving area boundary can be traversed based on the right obstacle point cloud set. In the whole process, no complicated vector calculation process is required, which reduces the calculation time. Moreover, the left obstacle point cloud set and the right obstacle point cloud set are corrected based on the current driving data and historical data, thereby improving the recognition accuracy of the drivable area.

[0111] In some embodiments, the method further comprises:

[0112] For any map unit, if there is a gap unit at the driving area boundary of any map unit and the gap unit meets the repair condition, the gap unit is interpolated based on the driving area boundary of the adjacent map unit on the same side as the gap unit to obtain a repaired gap unit.

[0113] In this embodiment, the repair conditions include but are not limited to at least one of the following: the width of the gap unit is greater than a preset width value, and the horizontal coordinates of the gap unit and the horizontal coordinates of the driving area boundary of the adjacent map unit are within a preset horizontal distance range. Then, the gap unit can be filled with the driving area boundary on the same side as the gap unit in the adjacent map unit. For example, a small gap will appear at some positions of the guardrail on the road for vehicles to turn around or pedestrians to pass, but for the radar device, the point cloud may not be collected at the gap position, so a gap unit will appear in the corresponding map unit.

[0114] In the above embodiment, by filling the gap units in the map units that meet the repair conditions, the accuracy of the drivable area map can be improved, thereby improving driving safety.

[0115] In some embodiments, the method further comprises:

[0116] A map unit to be cleaned that meets a drawing error condition is obtained, and the map unit to be cleaned is cleaned, wherein the drawing error condition includes at least one of the following: the map unit to be cleaned does not have driving area boundaries on both sides, the left driving area boundary or the right driving area boundary is smaller than a preset size value, and the number of point clouds in the map unit to be cleaned is smaller than a preset total number of point clouds.

[0117] In this embodiment, when the left driving area boundary or the right driving area boundary is smaller than the preset size value and the number of point clouds in the map unit to be cleaned is less than the preset total number of point clouds, it indicates that it may be a noise point. At this time, the map unit needs to be cleaned. For example, the map unit to be cleaned can be restored to the initial driving area boundary.

[0118] In the above embodiment, the map units meeting the drawing error condition are screened and cleaned, thereby improving the accuracy of the drivable area map and thus improving driving safety.

[0119] In some embodiments, the method further comprises:

[0120] Traversing the current point cloud data, if the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is smaller than the left driving area boundary of the map unit where the point cloud is located or larger than the right driving area boundary of the map unit where the point cloud is located, marking the point cloud as a point cloud within the non-driving area;

[0121] If the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is greater than the left driving area boundary of the map unit where the point cloud is located, and smaller than the right driving area boundary of the map unit where the point cloud is located, the point cloud is marked as a point cloud within the drivable area;

[0122] If the point cloud is within the region of interest of the current map, and the lateral coordinates of the point cloud and the left driving area boundary or the right driving area boundary of the map unit where the point cloud is located are within a preset boundary error range, the point cloud is marked as a boundary point cloud;

[0123] If the point cloud is not within the region of interest of the current map, the point cloud is marked as a point cloud in an unknown state.

[0124] In the above embodiment, the current point cloud is marked according to the area of ​​interest and the drivable area of ​​the current map, so that the status of the point cloud can be updated in real time later, the accuracy of the drivable area map is improved, and driving safety is improved.

[0125] On the other hand, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the vehicle driving area detection method described in any embodiment of the present application.

[0126] Among them, in the computer program product, an optional implementation form of the program module architecture of the computer program that implements each step of the target recognition method can be a vehicle driving area detection device.

[0127] See also Fig. 9 An embodiment of the present application provides a vehicle driving area detection device, comprising: a construction module 91, used to construct an original map corresponding to a current frame indicating a driving area, wherein the original map comprises a plurality of map units; a prediction module 92, used to obtain current point cloud data collected in the current frame, and obtain current driving data of the vehicle device, and predict a driving trajectory of the vehicle device based on the current driving data; a calculation module 93, used to form a reference baseline based on the driving trajectory, and calculate a left obstacle projection peak value of each map unit and a right obstacle projection peak value of each map unit based on the current point cloud data and the reference baseline. The calculation module 93 is also used to determine the left obstacle point cloud set of each map unit based on the left obstacle projection peak of each map unit, and to determine the right obstacle point cloud set of each map unit based on the right obstacle projection peak of each map unit; the calculation module 93 is also used to update the original map based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit to obtain the left driving area boundary and the right driving area boundary of each map unit, and generate the current map corresponding to the current frame based on the left driving area boundary and the right driving area boundary of each map unit.

[0128] Optionally, the building block 91 is further used for:

[0129] Acquire a region of interest in front of the vehicle equipment;

[0130] Based on the division step, the region of interest is divided into a plurality of map units in a longitudinal direction, wherein the longitudinal direction is along a center line of the vehicle equipment and points to a forward direction of the vehicle equipment, and each map unit includes a left initial driving area boundary and a right initial driving area boundary.

[0131] Optionally, the prediction module 92 is further configured to:

[0132] Determining whether current conditions meet map update conditions based on the current driving data;

[0133] When it is determined based on the current driving data that the current condition satisfies the map update condition, predicting a driving trajectory of the vehicle device based on the current driving data;

[0134] When the current driving speed is greater than a preset speed and the current yaw rate is less than a first preset yaw rate, determining that the current driving data determines that the original map meets a map update condition;

[0135] When the current driving speed is less than or equal to the preset speed and the status indicator of the original map indicates a valid state, continue to use the original map;

[0136] When the current yaw rate is less than a first preset yaw rate and the state indicator of the original map indicates a valid state, continue to use the original map;

[0137] When the current yaw rate is greater than or equal to the first preset yaw rate, the state mark of updating the original map is invalid.

[0138] Optionally, the prediction module 92 is further configured to:

[0139] Calculating a reference benchmark factor based on the current driving speed and the current yaw rate;

[0140] Based on the reference benchmark factor, the driving trajectory line is calculated using a trajectory prediction formula;

[0141] The calculation formula of the reference benchmark factor is as follows:

[0142]

[0143] The trajectory prediction formula is as follows:

[0144] y=0.5*C*x 2 ,

[0145] Among them, yawRate represents the current yaw rate, thre represents the second preset yaw rate, C represents the road curvature, CC represents the reference benchmark factor, vehicleSpeed ​​represents the current driving speed, x represents the distance that the vehicle device has traveled forward in the current frame relative to the previous preset frame of the current frame, and y represents the lateral offset distance of the vehicle device in the current frame relative to the previous preset frame of the current frame.

[0146] Optionally, the calculation module 93 is further used for:

[0147] For any map unit, based on the current point cloud data, obtain a left static point cloud and a right static point cloud;

[0148] Respectively projecting the left stationary point cloud and the right stationary point cloud to the lateral direction to obtain lateral coordinate distribution data of the left obstacle and lateral coordinate distribution data of the right obstacle, wherein the lateral direction is perpendicular to the longitudinal direction and located on the same plane, and the longitudinal direction is along the center line of the vehicle equipment and points to the forward direction of the vehicle equipment;

[0149] Determine a left projection peak value based on the lateral coordinate distribution data of the left obstacle, and determine a right projection peak value based on the lateral coordinate distribution data of the right obstacle;

[0150] When the left projection peak value satisfies the left peak value condition, obtaining the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and weighting the left projection peak value and the left obstacle projection peak value corresponding to the previous preset frame to obtain the left obstacle projection peak value of any map unit;

[0151] When the left projection peak value does not meet the left peak value condition, obtaining the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and obtaining the left obstacle projection peak value of any map unit based on the left obstacle projection peak value corresponding to the previous preset frame;

[0152] When it is determined that the right projection peak value satisfies the right peak value condition, obtaining the right obstacle projection peak value corresponding to the previous preset frame of the current frame, and weighting the right projection peak value and the right obstacle projection peak value corresponding to the previous preset frame to obtain the right obstacle projection peak value of any map unit;

[0153] When the right side projection peak value does not meet the right side peak condition, the right side obstacle projection peak value corresponding to the previous preset frame of the current frame is obtained, and the right side obstacle projection peak value of any map unit is obtained based on the right side obstacle projection peak value corresponding to the previous preset frame.

[0154] Optionally, the calculation module 93 is further used for:

[0155] Acquire point cloud data corresponding to a previous preset frame of the current frame, and perform compensation correction on the current point cloud data based on the point cloud data corresponding to the previous preset frame to obtain corrected point cloud data;

[0156] Determine a left obstacle point cloud candidate set for each map unit based on the corrected point cloud data and the left obstacle projection peak value of each map unit, and determine a right obstacle point cloud candidate set based on the corrected point cloud data and the right obstacle projection peak value of each map unit;

[0157] Based on the left obstacle point cloud candidate set of each map unit, the left obstacle point cloud set of each map unit is determined, and based on the right obstacle point cloud candidate set of each map unit, the right obstacle point cloud set of each map unit is determined.

[0158] Optionally, the calculation module 93 is further used for:

[0159] For any map unit, based on the corrected point cloud data and the left obstacle projection peak value of the any map unit, a left stationary point cloud in the any map unit that meets the left screening condition is determined as a left obstacle point cloud candidate set of the any map unit, wherein the left screening condition includes that the pitch position of the point cloud is within a preset pitch interval range, and the lateral coordinates of the point cloud and the left obstacle projection peak value of the any map unit are within a preset left lateral error range;

[0160] Based on the corrected point cloud data and the right obstacle projection peak value of any map unit, a right stationary point cloud in any map unit that meets a right screening condition is determined as a right obstacle point cloud candidate set of any map unit, wherein the right screening condition includes that the pitch position of the point cloud is within a preset pitch interval, and the lateral coordinate of the point cloud and the right obstacle projection peak value of any map unit are within a preset right lateral error range;

[0161] The transverse coordinate is a coordinate in the transverse direction, the transverse direction is perpendicular to the longitudinal direction and is located on the same plane, and the longitudinal direction is along the center line of the vehicle equipment and points to the forward direction of the vehicle equipment.

[0162] Optionally, the calculation module 93 is further used for:

[0163] For any map unit, fitting the point cloud corresponding to the left obstacle projection peak of the any map unit to obtain a left obstacle fitting curve; determining the point cloud whose distance to the left obstacle fitting curve is within a preset left distance range as the left obstacle point cloud candidate set of the any map unit;

[0164] The point cloud corresponding to the right obstacle projection peak of any map unit is fitted to obtain a right obstacle fitting curve; and the point cloud whose distance to the right obstacle fitting curve is within a preset right distance range is determined as a right obstacle point cloud candidate set of any map unit.

[0165] Optionally, the calculation module 93 is further used for:

[0166] For any map unit, if the number of point clouds in the left obstacle point cloud candidate set of the any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the left obstacle point cloud candidate set of the any map unit to obtain the left obstacle point cloud set of the any map unit;

[0167] When the number of point clouds in the right obstacle point cloud candidate set of any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the right obstacle point cloud candidate set of any map unit to obtain the right obstacle point cloud set of any map unit, wherein the deletion condition includes that the distance between point clouds is less than the preset interval distance.

[0168] Optionally, the calculation module 93 is further used for:

[0169] Sort the point clouds in the left obstacle point cloud set of each map unit according to the longitudinal coordinate to obtain the left sorted point cloud set of each map unit, and determine the position of the point cloud with the smallest transverse coordinate from the left sorted point cloud set of each map unit as the left driving area boundary of each map unit, wherein the longitudinal coordinate is the coordinate in the longitudinal direction, the transverse coordinate is the coordinate in the transverse direction, the transverse direction is perpendicular to the longitudinal direction and located on the same plane, the positive direction of the transverse direction is toward the left side of the vehicle device, and the longitudinal direction is along the center line of the vehicle device and points to the forward direction of the vehicle device;

[0170] The point clouds in the right obstacle point cloud set of each map unit are sorted according to the longitudinal coordinates to obtain the right sorted point cloud set of each map unit, and the position of the point cloud with the largest transverse coordinate is determined from the right sorted point cloud set of each map unit as the right driving area boundary of each map unit;

[0171] The current map is generated based on the left driving area boundary of each map unit and the right driving area boundary of each map unit.

[0172] Optionally, the cleaning module 94 is used to:

[0173] For any map unit, if there is a gap unit at the driving area boundary of any map unit and the gap unit meets the repair condition, the gap unit is interpolated based on the driving area boundary of the adjacent map unit on the same side as the gap unit to obtain a repaired gap unit.

[0174] Optionally, the cleaning module 94 is also used for:

[0175] A map unit to be cleaned that meets a drawing error condition is obtained, and the map unit to be cleaned is cleaned, wherein the drawing error condition includes at least one of the following: the map unit to be cleaned does not have driving area boundaries on both sides, the left driving area boundary or the right driving area boundary is smaller than a preset size value, and the number of point clouds in the map unit to be cleaned is smaller than a preset total number of point clouds.

[0176] Optionally, the update module 95 is used to:

[0177] Traversing the current point cloud data, if the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is smaller than the left driving area boundary of the map unit where the point cloud is located or larger than the right driving area boundary of the map unit where the point cloud is located, marking the point cloud as a point cloud within the non-driving area;

[0178] If the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is greater than the left driving area boundary of the map unit where the point cloud is located, and smaller than the right driving area boundary of the map unit where the point cloud is located, the point cloud is marked as a point cloud within the drivable area;

[0179] If the point cloud is within the region of interest of the current map, and the lateral coordinates of the point cloud and the left driving area boundary or the right driving area boundary of the map unit where the point cloud is located are within a preset boundary error range, the point cloud is marked as a boundary point cloud;

[0180] If the point cloud is not within the region of interest of the current map, the point cloud is marked as a point cloud in an unknown state.

[0181] It can be understood by those skilled in the art that Fig. 9 The structure of the vehicle driving area detection device does not constitute a limitation of the vehicle driving area detection device, and the various modules can be implemented in whole or in part by software, hardware, and a combination thereof. The above modules can be embedded in or independent of the controller in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the controller can call and execute the operations corresponding to the above modules. In other embodiments, the vehicle driving area detection device may include more or fewer modules than those shown in the figure.

[0182] See also Fig.10 In another aspect of the embodiment of the present application, a vehicle device 10 is provided, including a memory 3011 and a processor 3012. The memory 3011 stores a computer program. When the computer program is executed by the processor, the processor 3012 executes the steps of the vehicle driving area detection method provided in any of the above embodiments of the present application. The vehicle device 10 is a device installed on any type of mobile body, such as a vehicle, an electric vehicle, a hybrid electric vehicle, a motorcycle, a bicycle, a personal mobile device, an airplane, a drone, a ship, or a robot, etc.

[0183] The processor 3012 is the control center, which uses various interfaces and lines to connect various parts of the entire computer device, and executes various functions of the computer device and processes data by running or executing software programs and / or modules stored in the memory 3011, and calling data stored in the memory 3011. Optionally, the processor 3012 may include one or more processing cores; preferably, the processor 3012 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user pages and application programs, etc., and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 3012.

[0184] The memory 3011 can be used to store software programs and modules. The processor 3012 executes various functional applications and data processing by running the software programs and modules stored in the memory 3011. The memory 3011 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 3011 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 3011 may also include a memory controller to provide the processor 3012 with access to the memory 3011.

[0185] On the other hand, an embodiment of the present application further provides a storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the vehicle driving area detection method provided in any of the above embodiments of the present application.

[0186] Those skilled in the art can understand that all or part of the processes in the methods provided in the above embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0187] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A vehicle driving area detection method, applied to vehicle equipment, It is characterized in that include: constructing an original map corresponding to a current frame indicating a driving area, wherein the original map includes a plurality of map units; Acquire current point cloud data collected in a current frame, and acquire current driving data of the vehicle equipment, and predict a driving trajectory of the vehicle equipment based on the current driving data; forming a reference baseline based on the driving trajectory line, and calculating a left obstacle projection peak value of each map unit and a right obstacle projection peak value of each map unit based on the current point cloud data and the reference baseline; Based on the left obstacle projection peak value of each map unit, determine the left obstacle point cloud set of each map unit, and based on the right obstacle projection peak value of each map unit, determine the right obstacle point cloud set of each map unit; Based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit, the original map is updated to obtain the left driving area boundary and the right driving area boundary of each map unit, and based on the left driving area boundary and the right driving area boundary of each map unit, a current map corresponding to the current frame is generated.

2. The vehicle driving area detection method according to claim 1, It is characterized in that The constructing of the original map corresponding to the current frame indicating the driving area includes: Acquire a region of interest in front of the vehicle equipment; Based on the division step, the region of interest is divided into a plurality of map units in a longitudinal direction, wherein the longitudinal direction is along a center line of the vehicle equipment and points to a forward direction of the vehicle equipment, and each map unit includes a left initial driving area boundary and a right initial driving area boundary.

3. The vehicle driving area detection method according to claim 1, It is characterized in that The current driving data includes a current driving speed and a current yaw rate. Before predicting a driving trajectory of the vehicle device based on the current driving data, the method further includes: Determining whether current conditions meet map update conditions based on the current driving data; When it is determined based on the current driving data that the current condition satisfies the map update condition, predicting a driving trajectory of the vehicle device based on the current driving data; When the current driving speed is greater than a preset speed and the current yaw rate is less than a first preset yaw rate, determining that the current driving data determines that the original map meets a map update condition; When the current driving speed is less than or equal to the preset speed and the status indicator of the original map indicates a valid state, continue to use the original map; When the current yaw rate is less than a first preset yaw rate and the state indicator of the original map indicates a valid state, continue to use the original map; When the current yaw rate is greater than or equal to the first preset yaw rate, the state mark of updating the original map is invalid.

4. The vehicle driving area detection method according to claim 1, It is characterized in that The current driving data includes a current driving speed and a current yaw rate, and the predicting of the driving trajectory of the vehicle device based on the current driving data includes: Calculating a reference benchmark factor based on the current driving speed and the current yaw rate; Based on the reference benchmark factor, the driving trajectory line is calculated using a trajectory prediction formula; The calculation formula of the reference benchmark factor is as follows: The trajectory prediction formula is as follows: y=0.5*C*x 2 , Among them, yawRate represents the current yaw rate, thre represents the second preset yaw rate, C represents the road curvature, CC represents the reference benchmark factor, vehicleSpeed ​​represents the current driving speed, x represents the distance that the vehicle device has traveled forward in the current frame relative to the previous preset frame of the current frame, and y represents the lateral offset distance of the vehicle device in the current frame relative to the previous preset frame of the current frame.

5. The vehicle driving area detection method according to claim 1, It is characterized in that The forming of a reference baseline based on the driving trajectory line and the calculating of a left obstacle projection peak value of each map unit and a right obstacle projection peak value of each map unit based on the current point cloud data and the reference baseline include: For any map unit, based on the current point cloud data, obtain a left static point cloud and a right static point cloud; Respectively projecting the left stationary point cloud and the right stationary point cloud to the lateral direction to obtain lateral coordinate distribution data of the left obstacle and lateral coordinate distribution data of the right obstacle, wherein the lateral direction is perpendicular to the longitudinal direction and located on the same plane, and the longitudinal direction is along the center line of the vehicle equipment and points to the forward direction of the vehicle equipment; Determine a left projection peak value based on the lateral coordinate distribution data of the left obstacle, and determine a right projection peak value based on the lateral coordinate distribution data of the right obstacle; When the left projection peak value satisfies the left peak value condition, obtaining the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and weighting the left projection peak value and the left obstacle projection peak value corresponding to the previous preset frame to obtain the left obstacle projection peak value of any map unit; When the left projection peak value does not meet the left peak value condition, obtaining the left obstacle projection peak value corresponding to the previous preset frame of the current frame, and obtaining the left obstacle projection peak value of any map unit based on the left obstacle projection peak value corresponding to the previous preset frame; When it is determined that the right projection peak value satisfies the right peak value condition, obtaining the right obstacle projection peak value corresponding to the previous preset frame of the current frame, and weighting the right projection peak value and the right obstacle projection peak value corresponding to the previous preset frame to obtain the right obstacle projection peak value of any map unit; When the right side projection peak value does not meet the right side peak condition, the right side obstacle projection peak value corresponding to the previous preset frame of the current frame is obtained, and the right side obstacle projection peak value of any map unit is obtained based on the right side obstacle projection peak value corresponding to the previous preset frame.

6. The vehicle driving area detection method according to claim 1, It is characterized in that The step of determining the left obstacle point cloud set of each map unit based on the left obstacle projection peak value of each map unit, and determining the right obstacle point cloud set of each map unit based on the right obstacle projection peak value of each map unit includes: Acquire point cloud data corresponding to a previous preset frame of the current frame, and perform compensation correction on the current point cloud data based on the point cloud data corresponding to the previous preset frame to obtain corrected point cloud data; Determine a left obstacle point cloud candidate set for each map unit based on the corrected point cloud data and the left obstacle projection peak value of each map unit, and determine a right obstacle point cloud candidate set based on the corrected point cloud data and the right obstacle projection peak value of each map unit; Based on the left obstacle point cloud candidate set of each map unit, the left obstacle point cloud set of each map unit is determined, and based on the right obstacle point cloud candidate set of each map unit, the right obstacle point cloud set of each map unit is determined.

7. The vehicle driving area detection method according to claim 6, It is characterized in that The determining of the left obstacle point cloud candidate set of each map unit based on the corrected point cloud data and the left obstacle projection peak value of each map unit, and the determining of the right obstacle point cloud candidate set based on the corrected point cloud data and the right obstacle projection peak value of each map unit, comprises: For any map unit, based on the corrected point cloud data and the left obstacle projection peak value of the any map unit, a left stationary point cloud in the any map unit that meets the left screening condition is determined as a left obstacle point cloud candidate set of the any map unit, wherein the left screening condition includes that the pitch position of the point cloud is within a preset pitch interval range, and the lateral coordinates of the point cloud and the left obstacle projection peak value of the any map unit are within a preset left lateral error range; Based on the corrected point cloud data and the right obstacle projection peak value of any map unit, a right stationary point cloud in any map unit that meets a right screening condition is determined as a right obstacle point cloud candidate set of any map unit, wherein the right screening condition includes that the pitch position of the point cloud is within a preset pitch interval, and the lateral coordinate of the point cloud and the right obstacle projection peak value of any map unit are within a preset right lateral error range; The transverse coordinate is a coordinate in the transverse direction, the transverse direction is perpendicular to the longitudinal direction and is located on the same plane, and the longitudinal direction is along the center line of the vehicle equipment and points to the forward direction of the vehicle equipment.

8. The vehicle driving area detection method according to claim 6, It is characterized in that The determining of the left obstacle point cloud candidate set of each map unit based on the corrected point cloud data and the left obstacle projection peak value of each map unit, and the determining of the right obstacle point cloud candidate set based on the corrected point cloud data and the right obstacle projection peak value of each map unit, comprises: For any map unit, fitting the point cloud corresponding to the left obstacle projection peak of the any map unit to obtain a left obstacle fitting curve; determining the point cloud whose distance to the left obstacle fitting curve is within a preset left distance range as the left obstacle point cloud candidate set of the any map unit; The point cloud corresponding to the right obstacle projection peak of any map unit is fitted to obtain a right obstacle fitting curve; and the point cloud whose distance to the right obstacle fitting curve is within a preset right distance range is determined as a right obstacle point cloud candidate set of any map unit.

9. The vehicle driving area detection method according to claim 6, It is characterized in that The step of determining the left obstacle point cloud set of each map unit based on the left obstacle point cloud candidate set of each map unit, and determining the right obstacle point cloud set of each map unit based on the right obstacle point cloud candidate set of each map unit, comprises: For any map unit, if the number of point clouds in the left obstacle point cloud candidate set of the any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the left obstacle point cloud candidate set of the any map unit to obtain the left obstacle point cloud set of the any map unit; When the number of point clouds in the right obstacle point cloud candidate set of any map unit meets the preset number of point clouds, point clouds meeting the deletion condition are deleted from the right obstacle point cloud candidate set of any map unit to obtain the right obstacle point cloud set of any map unit, wherein the deletion condition includes that the distance between point clouds is less than the preset interval distance.

10. The vehicle driving area detection method according to claim 1, It is characterized in that The updating of the original map based on the left obstacle point cloud set of each map unit and the right obstacle point cloud set of each map unit to obtain the left driving area boundary and the right driving area boundary of each map unit, and generating the current map corresponding to the current frame based on the left driving area boundary and the right driving area boundary of each map unit, including: Sort the point clouds in the left obstacle point cloud set of each map unit according to the longitudinal coordinate to obtain the left sorted point cloud set of each map unit, and determine the position of the point cloud with the smallest transverse coordinate from the left sorted point cloud set of each map unit as the left driving area boundary of each map unit, wherein the longitudinal coordinate is the coordinate in the longitudinal direction, the transverse coordinate is the coordinate in the transverse direction, the transverse direction is perpendicular to the longitudinal direction and located on the same plane, the positive direction of the transverse direction is toward the left side of the vehicle device, and the longitudinal direction is along the center line of the vehicle device and points to the forward direction of the vehicle device; The point clouds in the right obstacle point cloud set of each map unit are sorted according to the longitudinal coordinates to obtain the right sorted point cloud set of each map unit, and the position of the point cloud with the largest transverse coordinate is determined from the right sorted point cloud set of each map unit as the right driving area boundary of each map unit; The current map is generated based on the left driving area boundary of each map unit and the right driving area boundary of each map unit.

11. The vehicle driving area detection method according to claim 1, It is characterized in that The method further comprises: For any map unit, if there is a gap unit at the driving area boundary of any map unit and the gap unit meets the repair condition, the gap unit is interpolated based on the driving area boundary of the adjacent map unit on the same side as the gap unit to obtain a repaired gap unit.

12. The vehicle driving area detection method according to claim 1, It is characterized in that The method further comprises: A map unit to be cleaned that meets a drawing error condition is obtained, and the map unit to be cleaned is cleaned, wherein the drawing error condition includes at least one of the following: the map unit to be cleaned does not have driving area boundaries on both sides, the left driving area boundary or the right driving area boundary is smaller than a preset size value, and the number of point clouds in the map unit to be cleaned is smaller than a preset total number of point clouds.

13. The vehicle driving area detection method according to claim 1, It is characterized in that The method further comprises: Traversing the current point cloud data, if the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is smaller than the left driving area boundary of the map unit where the point cloud is located or larger than the right driving area boundary of the map unit where the point cloud is located, marking the point cloud as a point cloud within the non-driving area; If the point cloud is within the area of ​​interest of the current map, and the lateral coordinate of the point cloud is greater than the left driving area boundary of the map unit where the point cloud is located, and smaller than the right driving area boundary of the map unit where the point cloud is located, the point cloud is marked as a point cloud within the drivable area; If the point cloud is within the region of interest of the current map, and the lateral coordinates of the point cloud and the left driving area boundary or the right driving area boundary of the map unit where the point cloud is located are within a preset boundary error range, the point cloud is marked as a boundary point cloud; If the point cloud is not within the region of interest of the current map, the point cloud is marked as a point cloud in an unknown state.

14. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, the vehicle driving area detection method according to any one of claims 1 to 13 is implemented.

15. A vehicle device, It is characterized in that The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the vehicle driving area detection method according to any one of claims 1 to 13.

Images