Obstacle detection method and system, electronic equipment and storage medium

Abstract

The invention discloses an obstacle detection method and system, electronic equipment and a storage medium. The obstacle detection method comprises the following steps: acquiring point cloud data corresponding to an obstacle; analyzing the point cloud data and converting the point cloud data into a Cartesian coordinate system to obtain a coordinate value of each point in the point cloud data in the Cartesian coordinate system; constructing a grid of a cylindrical coordinate system according to the coordinate value of each point in the Cartesian coordinate system; mapping the point cloud data into a grid space of a cylindrical coordinate system to obtain a coordinate value of each point in the point cloud data in the grid space of the cylindrical coordinate system; marking ground grids and non-ground grids according to the characteristics of the coordinate values of the adjacent grids; traversing a neighborhood range of a non-ground grid by adopting a region growing algorithm, and marking the same value for the adjacent grids; according to the method, points in non-ground grids with the same values are mapped to a Cartesian coordinate system to obtain obstacle information, so that the robustness and accuracy of ground segmentation are improved, and the risk of ground false detection of a clustering algorithm is reduced.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of detection, and in particular to an obstacle detection method and system, an electronic device and a storage medium. BACKGROUND

[0002] An unmanned vehicle can accurately perceive the surrounding environment through sensors, thereby reducing the occurrence of traffic accidents from the vehicle hardware level. The continuous development of artificial intelligence technology and the progress of computer vision technology have injected new vitality into unmanned vehicles. Currently, vehicle sensors mainly include cameras, laser radars and millimeter wave radars, etc. Laser radars have good distance resolution and angular resolution, have high detection accuracy, strong anti-interference ability, and can be used at night, etc. Laser radars can generate a series of three-dimensional point clouds, and the error accuracy can reach centimeter level. The three-dimensional point clouds reflect the distance, position and reflection intensity of the corresponding target, etc., which can enable the vehicle to quickly detect the drivable area, reproduce the shape, position and size of the obstacle, and further draw a high-precision map of the surrounding environment, thereby meeting the requirements for obstacle detection.

[0003] In the prior art, the ground point cloud segmentation method includes: dividing the collected laser radar data into different polar coordinate channels, dividing each channel into different units based on the radial distance, constructing a polar coordinate grid, selecting a seed point in the polar coordinate grid, and then using an incremental straight line fitting algorithm to segment the ground point cloud; selecting a seed point for plane fitting according to the height value and other coordinate information of the point cloud in the grid unit, and selecting the optimal plane to represent the ground geometry around the vehicle; dividing the point cloud into different regions according to different distance ranges, using different grid unit sizes to divide the point cloud in different regions, and performing plane fitting in each grid to represent the ground shape at the position to segment the ground point cloud. The above ground point cloud segmentation method has the problems of under-segmentation, mis-segmentation and high computational complexity of the ground point cloud, and poor segmentation accuracy for the road surface with significant pitch changes and slopes of the vehicle. SUMMARY

[0004] The technical problem to be solved by the present disclosure is to overcome the defects of the prior art ground point cloud segmentation method, such as under-segmentation, mis-segmentation and high computational complexity, and to provide a video fusion method, system, electronic device and storage medium.

[0005] The present disclosure solves the above technical problems by the following technical solutions:

[0006] The present disclosure discloses an obstacle detection method, which comprises:

[0007] Obtaining point cloud data corresponding to an obstacle;

[0008] parsing and converting the point cloud data into a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system;

[0009] constructing a grid of a cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system;

[0010] mapping the point cloud data into the grid space of the cylindrical coordinate system to obtain coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system;

[0011] labeling ground grids and non-ground grids according to features of adjacent grid coordinate values;

[0012] using a region growing algorithm to traverse a neighborhood range of the non-ground grids and labeling adjacent grids with the same value;

[0013] mapping points in the non-ground grids with the same value to the Cartesian coordinate system to obtain obstacle information.

[0014] Preferably, the step of labeling ground grids and non-ground grids according to features of adjacent grid coordinate values comprises:

[0015] obtaining a maximum value and a minimum value of z coordinates of the adjacent grid coordinate values;

[0016] obtaining a height difference of the adjacent grids according to the maximum value and the minimum value of the z coordinates;

[0017] determining an angle of the grid with the ground according to the height difference and a grid size;

[0018] determining whether the angle is greater than a first preset threshold value, if yes, the adjacent grid is a non-ground grid, and if no, the adjacent grid is a ground grid.

[0019] Preferably, after the step of obtaining coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system and before the step of labeling ground grids and non-ground grids according to features of adjacent grid coordinate values, the obstacle detection method comprises:

[0020] obtaining a number of points in a grid, and filtering out the grid if the number of points is less than a second preset threshold value.

[0021] Preferably, the step of determining the angle of the grid with the ground according to the height difference and the grid size comprises:

[0022] determining a value of θ according to a formula, and taking the value of θ as the angle of the grid with the ground;

[0023] Wherein, θ represents the slope of the point cloud in the z-axis direction between adjacent grids; Δ represents the distance between the centers of adjacent grids; Δρ represents the height difference of the z-axis coordinate values of the point cloud in adjacent grids.

[0024] The present disclosure also discloses an obstacle detection system, the obstacle detection method comprising:

[0025] An acquisition module is configured to acquire point cloud data corresponding to an obstacle.

[0026] An analysis and conversion module is configured to analyze and convert the point cloud data to a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system.

[0027] A construction module is configured to construct a grid of a cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system.

[0028] A first mapping module is configured to map the point cloud data to the grid space of the cylindrical coordinate system to obtain coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system.

[0029] A marking module is configured to mark a ground grid and a non-ground grid according to the characteristics of the coordinate values of adjacent grids.

[0030] The marking module is configured to traverse the neighborhood range of the non-ground grid using a region growing algorithm, and mark the near-neighbor grid with the same value.

[0031] A second mapping module is configured to map the points in the non-ground grid with the same value to the Cartesian coordinate system to obtain obstacle information.

[0032] Preferably, the marking module comprises:

[0033] A first acquisition unit is configured to acquire the maximum and minimum values of the z-coordinate of the coordinate values of adjacent grids.

[0034] A second acquisition unit is configured to acquire the height difference of adjacent grids according to the maximum and minimum values of the z-coordinate.

[0035] A determination unit is configured to determine the angle of the grid with the ground according to the height difference and the grid size.

[0036] A judgment unit is configured to determine whether the angle is greater than a first preset threshold value, if yes, the adjacent grid is a non-ground grid, and if not, the adjacent grid is a ground grid.

[0037] Preferably, the obstacle detection system comprises:

[0038] A filtering module is configured to acquire the number of points in a grid, and if the number of points is less than a second preset threshold value, the grid is filtered out.

[0039] Preferably, the determining unit 53 is specifically configured to:

[0040] According to the formula, determine the value of θ as the angle of the grid with the ground;

[0041] wherein θ represents the slope of the point cloud in the z-axis direction between adjacent grids; Δ represents the distance between the centers of adjacent grids; and Δρ represents the height difference of the z-axis coordinate values of the point cloud in adjacent grids.

[0042] The present disclosure also discloses an electronic device comprising a memory, a processor, and a computer program stored in the memory and configured to run on the processor, wherein the processor executes the computer program to implement the obstacle detection method as described above.

[0043] The present disclosure also discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the obstacle detection method as described above.

[0044] The positive progress effect of the present disclosure is that:

[0045] The present disclosure calculates the angle of the grid with the ground through the height difference between adjacent grids, and determines whether it is a non-ground grid. This method discards the idea of fitting the geometric characteristics of the ground, so it is not affected by the slope of the ground and the change in the pitch of the ego vehicle, thereby improving the robustness and accuracy of ground segmentation and reducing the risk of false detection of the clustering algorithm. The present scheme only needs to construct a column coordinate grid with one dimension size, which greatly reduces the computational amount and is simple to calculate after voxelizing the point cloud, and can well meet the real-time requirement. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 Flowchart of the obstacle detection method of Embodiment 1 of the present disclosure;

[0047] Figure 2 Schematic diagram of the laser radar point cloud Cartesian coordinate system of Embodiment 1 of the present disclosure;

[0048] Figure 3 Flowchart of step S105 of Embodiment 1 of the present disclosure;

[0049] Figure 4 Schematic diagram of the laser radar point cloud cylindrical coordinate system of Embodiment 1 of the present disclosure;

[0050] Figure 5 Module schematic diagram of the obstacle detection system of Embodiment 2 of the present disclosure;

[0051] Figure 6 Module schematic diagram of the marking module 5 of Embodiment 2 of the present disclosure;

[0052] Figure 7 FIG. 1 is a structural schematic diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0053] The present disclosure will be further described below by way of examples, but the present disclosure is not limited to the scope of the examples.

[0054] Embodiment 1

[0055] As shown in the drawings, the present embodiment discloses an obstacle detection method, which comprises: Figure 1

[0056] Step S101, acquiring point cloud data corresponding to an obstacle;

[0057] Specifically, the point cloud data corresponding to the obstacle can be acquired by a laser radar.

[0058] Step S102, parsing and converting the point cloud data to a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system;

[0059] Specifically, the point cloud data obtained in step S101 is parsed and converted to the Cartesian coordinate system to obtain coordinate values (x, y, z) of each point in the Cartesian coordinate system.

[0060] x = rcos(ω)sin(α)

[0061] y = rcos(ω)sin(α)

[0062] z = rsin(ω)

[0063] Wherein, x represents the coordinate value of the point on the x-axis in the Cartesian coordinate system; y represents the coordinate value of the point on the y-axis in the Cartesian coordinate system; z represents the coordinate value of the point on the z-axis in the Cartesian coordinate system; ω represents the angle of vertical emission of the point cloud; and α represents the angle of horizontal emission of the point cloud.

[0064] Step S103, constructing a grid of a cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system;

[0065] Specifically, the grid of the cylindrical coordinate system is constructed according to the coordinate values (x, y, z) of each point in the Cartesian coordinate system. The grid size of the cylindrical coordinate system is (Δρ, Δθ, Δz). The construction process of the cylindrical coordinate system is that the angle value of the point cloud in the horizontal plane and the distance value of the point cloud from the origin in the polar radius direction are calculated by the x and y coordinate values of the point cloud, the distance value of the point cloud from the origin in the z-axis direction is obtained by the z coordinate value of the point cloud, and then the grid index of each point cloud is obtained by dividing the above values by the preset grid size, thereby completing the construction of the grid.

[0066] ​Step S104, mapping the point cloud data into the grid space of the cylindrical coordinate system to obtain coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system;

[0067] Step S105, marking the ground grid and the non-ground grid according to the characteristics of the adjacent grid coordinate values;

[0068] Step S106, traversing the neighborhood range of the non-ground grid by using a region growing algorithm, and marking the near neighbor grid as the same value;

[0069] Specifically, the specific steps of traversing the neighborhood range of the non-ground grid by using the region growing algorithm include:

[0070] Step S107, mapping the points in the non-ground grid with the same value to the Cartesian coordinate system to obtain obstacle information.

[0071] Specifically, mapping the points in the non-ground grid with the same value to the Cartesian coordinate system realizes the point cloud clustering effect, and further obtains the obstacle information, which can include the position information and size information of the obstacle. Figure 2 As shown in the figure, the laser radar point cloud Cartesian coordinate system is shown, the laser radar center point is o, three mutually perpendicular number axes are made through the fixed point o, the three axes are all with o as the origin and have the same length unit, the x axis and the y axis are in the horizontal plane, and the z axis is a plumb line.

[0072] In the scheme, the angle of the grid and the ground is calculated through the height difference between the adjacent grids to determine whether it is a non-ground grid. This method discards the idea of fitting the geometric characteristics of the ground, so it is not affected by the ground slope and the change of the vehicle pitch, thereby improving the robustness and accuracy of the ground segmentation and reducing the risk of ground false detection of the clustering algorithm. The scheme only needs to construct a one-dimensional size cylindrical coordinate grid, which greatly reduces the operation amount and is simple to calculate, and can well meet the real-time requirement.

[0073] As shown in the figure, in an implementable manner, step S105 includes: Figure 3

[0074] Step S1051, obtaining the maximum value and the minimum value of the z coordinate of the adjacent grid coordinate values;

[0075] Step S1052, obtaining the height difference of the adjacent grid according to the maximum value and the minimum value of the z coordinate;

[0076] Step S1053, determining the angle of the grid and the ground according to the height difference and the grid size;

[0077] Step S1054, determining whether the angle is greater than a first preset threshold value, if yes, executing step S1055, and if no, executing step S1056;​

[0078] Step S1055, determining that the adjacent grid is a non-ground grid;

[0079] Step S1056, determining that the adjacent grid is a ground grid.

[0080] Specifically, according to the coordinate values (x, y, z) of the point cloud, the detection range, and the grid size (Δρ, Δθ, Δz) of the cylindrical coordinate system, the cylindrical coordinate grid index (ρ, θ, z) of the point cloud is calculated, the maximum and minimum values of the z coordinate of the adjacent grid coordinate value are obtained, the difference Δ between the maximum value of the z coordinate in each grid index (ρ, θ, z) and the minimum value of the z coordinate in the adjacent non-empty grid index is calculated, and the difference is taken as the height difference of the adjacent grid. Wherein, Δρ represents the length of a single grid in the radial direction of the set cylindrical coordinate system grid; Δθ represents the radian of a single grid in the angle direction of the set cylindrical coordinate system grid; Δz represents the length of a single grid in the z-axis direction of the set cylindrical coordinate system grid; ρ represents the index value in the radial direction of the cylindrical coordinate system after each point is converted to the cylindrical coordinate system; θ represents the index value in the angle direction of the cylindrical coordinate system after each point is converted to the cylindrical coordinate system; and z represents the index value in the z-axis direction of the cylindrical coordinate system after each point is converted to the cylindrical coordinate system. As shown in the figure, the laser radar point cloud cylindrical coordinate system, the laser radar center point is o, three mutually perpendicular number axes are made through the fixed point o, the three axes are all taken o as the origin and have the same length unit, the x-axis and the y-axis are in the horizontal plane, and the z-axis is a plumb line. Figure 4

[0081] The scheme, by obtaining the maximum and minimum values of the z coordinate of the adjacent grid coordinate value, obtaining the height difference of the adjacent grid, determining the angle between the grid and the ground, and then determining the non-ground grid through the angle, further determines whether it is an obstacle.

[0082] In an implementable manner, after the step of obtaining the coordinate value of each point in the point cloud data in the grid space of the cylindrical coordinate system, and before the step of marking the ground grid and the non-ground grid according to the characteristics of the adjacent grid coordinate value, the obstacle detection method comprises:

[0083] Obtaining the number of points in the grid, if the number of points is less than a second preset threshold, the grid is filtered out.

[0084] Specifically, after the step of obtaining the coordinate value of each point in the point cloud data in the grid space of the cylindrical coordinate system, the number of points in each cylindrical coordinate grid is counted. If the number of points in the cylindrical coordinate grid is less than a second preset threshold, such as less than 10, the grid is filtered out.

[0085] ​The scheme filters out part of the grid through the number of points in the grid, i.e., preliminarily screens out the non-ground grid, to improve the detection efficiency of the obstacle after the step of obtaining the coordinate value of each point of the point cloud data in the grid space of the cylindrical coordinate system and before the step of marking the ground grid and the non-ground grid according to the characteristics of the adjacent grid coordinate values.

[0086] In an implementable manner, the step S1053 comprises:

[0087] According to the formula, determining the value of θ as the angle of the grid to the ground;

[0088] wherein θ represents the slope of the point cloud in the z-axis direction between adjacent grids; Δ represents the distance between the centers of adjacent grids; and Δρ represents the height difference of the z-axis coordinate value of the point cloud in the adjacent grid.

[0089] The scheme determines the angle of the grid to the ground through the formula to ensure the accuracy of obstacle detection.

[0090] Embodiment 2

[0091] As shown in Figure 5 The embodiment discloses an obstacle detection system, and the obstacle detection method comprises the following steps:

[0092] An acquisition module 1 is configured to acquire point cloud data corresponding to an obstacle;

[0093] Specifically, the point cloud data corresponding to the obstacle can be acquired by a laser radar.

[0094] An analysis and conversion module 2 is configured to analyze and convert the point cloud data to a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system.

[0095] Specifically, the point cloud data obtained in step S101 is analyzed and converted to the Cartesian coordinate system to obtain coordinate values (x, y, z) of each point in the Cartesian coordinate system.

[0096] x = rcos(ω)sin(α)

[0097] y = rcos(ω)sin(α)

[0098] z = rsin(ω)

[0099] wherein x represents the coordinate value of the point in the x-axis of the Cartesian coordinate system; y represents the coordinate value of the point in the y-axis of the Cartesian coordinate system; z represents the coordinate value of the point in the z-axis of the Cartesian coordinate system; ω represents the angle of vertical emission of the point cloud; and α represents the angle of horizontal emission of the point cloud.

[0100] A constructing module 3 is configured to construct a grid of the cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system;

[0101] Specifically, the grid of the cylindrical coordinate system is constructed according to the coordinate values (x, y, z) of each point in the Cartesian coordinate system. The grid size (Δρ, Δθ, Δz) of the cylindrical coordinate system. The cylindrical coordinate construction process is that the angle value of each point in the horizontal plane and the distance value of each point from the origin in the polar radius direction are calculated by the x, y coordinate values of the point cloud, the z coordinate of the point cloud is used to obtain the distance value of the point cloud from the origin in the z-axis direction, and then the grid index of each point cloud is obtained by dividing the preset grid size, and the construction of the grid is completed.

[0102] A first mapping module 4 is configured to map the point cloud data into the grid space of the cylindrical coordinate system to obtain the coordinate value of each point in the point cloud data in the grid space of the cylindrical coordinate system;

[0103] A first marking module 5 is configured to mark the ground grid and the non-ground grid according to the features of the adjacent grid coordinate values;

[0104] A second marking module 6 is configured to traverse the neighborhood range of the non-ground grid by using a region growing algorithm, and mark the near neighbor grid with the same value.

[0105] Specifically, the specific steps of traversing the neighborhood range of the non-ground grid by using the region growing algorithm include:

[0106] A second mapping module 7 is configured to map the points in the non-ground grid with the same value into the Cartesian coordinate system to obtain obstacle information.

[0107] Specifically, the points in the non-ground grid with the same value are mapped into the Cartesian coordinate system, which realizes the point cloud clustering effect and further obtains the obstacle information, wherein the obstacle information can include the position information and the size information of the obstacle. Figure 2 As shown in the figure, the laser radar point cloud Cartesian coordinate system is shown, the laser radar center point is o, three mutually perpendicular number axes are made through the fixed point o, the three axes are all with o as the origin and have the same length unit, the x-axis and the y-axis are in the horizontal plane, and the z-axis is a plumb line.

[0108] In the scheme, the angle of the grid and the ground is calculated by the height difference between the adjacent grids, whether it is a non-ground grid is judged, the method discards the fitting ground geometric feature idea, so it is not affected by the ground slope and the change of the self-vehicle pitch, the robustness and accuracy of the ground segmentation are improved, and the risk of ground false detection of the clustering algorithm is reduced. The scheme only needs to construct a one-dimensional size cylindrical coordinate grid, the operation amount is greatly reduced after the point cloud is voxelized, and the calculation is simple, and the real-time requirement can be well met.

[0109] As shown in the figure, Figure 6As shown, in an implementable mode, the marking module 5 comprises:

[0110] The first obtaining unit 51 is configured to obtain a maximum value and a minimum value of z coordinates of adjacent grid coordinate values.

[0111] The second obtaining unit 52 is configured to obtain a height difference of adjacent grids according to the maximum value and the minimum value of the z coordinates.

[0112] The determining unit 53 is configured to determine an angle of a grid with the ground according to the height difference and a grid size.

[0113] The judging unit 54 is configured to judge whether the angle is greater than a first preset threshold value, if yes, determining that the adjacent grid is a non-ground grid, and if not, determining that the adjacent grid is a ground grid.

[0114] In an implementable mode, the obstacle detection system comprises:

[0115] The filtering module 8 is configured to obtain a point number in a grid, and filter out the grid if the point number is less than a second preset threshold value.

[0116] In an implementable mode, the determining unit 53 is specifically configured to:

[0117] determine a value of θ according to a formula, and take the value of θ as the angle of the grid with the ground.

[0118] wherein θ represents a slope of point cloud in a z axis direction between adjacent grids; Δ represents a distance between centers of adjacent grids; and Δρ represents a height difference of z axis coordinate values of point cloud in adjacent grids.

[0119] Embodiment 3

[0120] Figure 7 A structural schematic diagram of an electronic device provided in Embodiment 5 of the present disclosure is shown. The electronic device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the obstacle detection method provided in Embodiment 1 when executing the program. Figure 7 The electronic device 50 shown is merely an example, and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0121] As Figure 7 shown, the electronic device 50 can be in the form of a general computing device, for example, it can be a server device. The components of the electronic device 50 can include but are not limited to the above-mentioned at least one processor 51, the above-mentioned at least one memory 52, and a bus 53 connecting different system components including the memory 52 and the processor 51.

[0122] Bus 53 includes a data bus, an address bus, and a control bus.

[0123] Memory 52 can include volatile memory, such as random access memory (RAM) 521 and / or cache memory 522, and / or non-volatile memory, such as read only memory (ROM) 523.

[0124] Memory 52 can also include a program / utility 525 having a set of program modules 524, including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which can include implementation of a network environment, alone or in combination.

[0125] Processor 51 can execute instructions for various functions and data processing, such as the obstacle detection method provided by Embodiment 1 of the present disclosure, by running computer programs stored in memory 52.

[0126] Electronic device 50 can also communicate with one or more external devices 54, such as a keyboard or a pointing device, through input / output (I / O) interface 55. Further, model generation device 50 can communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet, through network adapter 56. As depicted, network adapter 56 communicates with the other modules of model generation device 50 through bus 53. It should be appreciated that although not shown, other hardware and / or software modules could be used in conjunction with model generation device 50, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0127] It should be noted that although several units / modules or sub-units / modules of an electronic device are mentioned in the foregoing detailed description, such a division is merely exemplary and not mandatory. Indeed, according to an embodiment of the present disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into units / modules embodied by multiple units / modules.

[0128] Embodiment 4

[0129] The present embodiment provides a computer readable storage medium, having stored thereon a computer program, the program being executed by a processor to implement a video fusion method.

[0130] More specifically, the readable storage medium can include, but is not limited to, a portable disc, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0131] In possible implementation manners, the present disclosure can also be implemented in the form of a program product, which includes program codes for causing terminal equipment to execute a video fusion method when the program product is run on the terminal equipment.

[0132] The program codes for executing the present disclosure can be written in any combination of one or more programming languages, and can be executed completely on a user equipment, partially on the user equipment, as a stand-alone software package, partially on the user equipment and partially on a remote equipment, or completely on a remote equipment.

[0133] Although the specific implementation manners of the present disclosure are described above, those skilled in the art should understand that this is only an illustration, and the protection scope of the present disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these implementation manners without departing from the principles and essence of the present disclosure, and these changes and modifications all fall within the protection scope of the present disclosure.

Claims

1. An obstacle detection method characterized by, The obstacle detection method comprises: Obtain point cloud data corresponding to the obstacle; Parse and convert the point cloud data to a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system; Construct a grid of a cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system; Map the point cloud data to the grid space of the cylindrical coordinate system to obtain coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system; Label ground grids and non-ground grids according to features of adjacent grid coordinate values; Use a region growing algorithm to traverse a neighborhood range of the non-ground grids, and label adjacent grids with the same value; Map points in the non-ground grids with the same value to the Cartesian coordinate system to obtain obstacle information.

2. The obstacle detection method of claim 1, wherein, The step of labeling the ground grids and the non-ground grids according to the features of the adjacent grid coordinate values comprises: Obtain a maximum value and a minimum value of z coordinates of the adjacent grid coordinate values; Obtain a height difference of the adjacent grids according to the maximum value and the minimum value of the z coordinates; Determine an angle of the grid with the ground according to the height difference and a grid size; Determine whether the angle is greater than a first preset threshold value, and if yes, determine that the adjacent grid is a non-ground grid, and if not, determine that the adjacent grid is a ground grid.

3. The obstacle detection method according to claim 1 or 2, characterized by, After the step of obtaining the coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system, and before the step of labeling the ground grids and the non-ground grids according to the features of the adjacent grid coordinate values, the obstacle detection method comprises: Obtain a number of points in the grid, and filter out the grid if the number of points is less than a second preset threshold value.

4. The obstacle detection method of claim 2, wherein, The step of determining the angle of the grid with the ground according to the height difference and the grid size comprises: According to the formula, The value of θ is determined and is taken as the angle of the grid to the ground; Wherein, θ represents a slope of point cloud in a z-axis direction between adjacent grids; Δ represents a distance between centers of adjacent grids; Δρ represents a height difference of z-axis coordinate values of point cloud in adjacent grids.

5. An obstacle detection system characterized by, The obstacle detection method comprises: An obtaining module is configured to obtain point cloud data corresponding to an obstacle; An analysis and conversion module is configured to parse and convert the point cloud data to a Cartesian coordinate system to obtain coordinate values of each point in the point cloud data in the Cartesian coordinate system; A construction module is configured to construct a grid of a cylindrical coordinate system according to the coordinate values of each point in the Cartesian coordinate system; A first mapping module is configured to map the point cloud data to the grid space of the cylindrical coordinate system to obtain coordinate values of each point in the point cloud data in the grid space of the cylindrical coordinate system; A labeling module is configured to label ground grids and non-ground grids according to features of adjacent grid coordinate values; The labeling module is configured to use a region growing algorithm to traverse a neighborhood range of the non-ground grids, and label adjacent grids with the same value; A second mapping module is configured to map points in the non-ground grids with the same value to the Cartesian coordinate system to obtain obstacle information.

6. The obstacle detection system of claim 5, wherein, The labeling module comprises: A first obtaining unit is configured to obtain a maximum value and a minimum value of z coordinates of adjacent grid coordinate values; A second obtaining unit is configured to obtain a height difference of adjacent grids according to the maximum value and the minimum value of the z coordinates; A determination unit is configured to determine an angle of the grid with the ground according to the height difference and a grid size; The judging unit is configured to judge whether the angle is greater than a first preset threshold value, and if yes, determine that the adjacent grid is a non-ground grid, and if not, determine that the adjacent grid is a ground grid.

7. The obstacle detection system of claim 5 or 6, wherein The obstacle detection system comprises: The filtering module is configured to acquire a point number in a grid, and filter out the grid if the point number is less than a second preset threshold value.

8. The obstacle detection system of claim 6, wherein, The determining unit is specifically configured to determine the value of θ according to a formula, determines the value of θ and takes the value of θ as the angle between the grid and the ground; Wherein, θ represents a slope of point cloud in a z-axis direction between adjacent grids; Δ represents a distance between centers of adjacent grids; and Δρ represents a height difference of z-axis coordinate values of point cloud in adjacent grids.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory for running on the processor, characterized in that, The processor executes the computer program to implement the obstacle detection method according to any one of claims 1 to 4.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the obstacle detection method according to any one of claims 1 to 4.

Images