Method for quickly generating road data containing vectorization information

Abstract

The invention discloses a method for quickly generating road data containing vectorization information. The method comprises the following steps of: constructing a segmented lane topological structure by line segment roads and / or arc segment roads; designing parameters based on the road, such as road target length, road target width, point resolution, central angle and the like; respectively calculating an inner boundary point set, an outer boundary point set and a center line point set in the straight road segment and / or the arc road segment; the inner boundary point set, the outer boundary point set and the center line point set of the linear section road segment and / or the arc section road segment are extracted, and a lane boundary point set containing unaligned points is extracted in a sampling mode to form a data set. According to the method, a large amount of road data can be quickly generated, the road data can comprise lane boundary information and center line information of different spatial alignment degrees, and a large amount of road data can be provided for other related experiments.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent transportation and automatic driving, and in particular to a method for quickly generating road data containing vectorized information. Background Art

[0002] In the study of road topology and autonomous vehicle trajectory planning, road boundaries and lane centerlines play a vital role.

[0003] However, in the prior art, the road boundary lines are usually obtained by manual annotation or with the help of sensors, and then the lane centerline information is extracted from the road boundary lines using algorithms. This type of method for calculating the lane centerline based on the road boundary lines is affected by the density of boundary points on the lane boundary lines and the spatial alignment of boundary points on the two boundaries. Although some data sets can provide lane boundary information and centerline information, the number of lanes in the data set is limited and cannot effectively contain a large amount of data to simulate the spatial misalignment of road boundary points.

[0004] Therefore, there is an urgent need for a method that can quickly generate road data containing a large amount of lane boundary information and centerline information, and can provide lane boundary information and centerline information with different spatial alignments, so as to provide a large amount of data for related research such as lane centerlines. Summary of the invention

[0005] In order to address the deficiencies in the prior art, the present invention provides a method for quickly generating road data containing vectorized information. This method can be used to quickly generate a large amount of road data, and the road data can include lane boundary information, centerline information, etc. with different spatial alignments. The road data generated by this method can be used for other related experiments.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A method for quickly generating road data containing vectorized information comprises the following steps:

[0008] Step 1: construct a segmented lane topology structure from at least one straight road segment and / or at least one arc road segment;

[0009] Step 2: Initialize and set the design parameters in the segmented lane topology structure, including the target road length L road 、Road target width W road , design parameters of the straight segment road, design parameters of the arc segment road;

[0010] The design parameters of the straight road segment include but are not limited to the original point resolution of the inner boundary of the straight road. Original point resolution of the outer boundary of the straight road The original point resolution of the center line of the straight road is

[0011] The design parameters of the arc segment road include but are not limited to the arc segment road center angle θ crv , the radius of the arc section road centerline is R crv , the point resolution on the center line of the arc segment is

[0012] Step 3: Based on the above segmented lane topology and design parameters, calculate the inner boundary point set, outer boundary point set and center line point set in the straight road segment and / or arc road segment respectively;

[0013] Step 4: Based on the inner boundary point set, outer boundary point set and center line point set of the straight segment road segment and / or arc segment road segment obtained in step 3, a lane boundary point set containing misaligned points is extracted by sampling to form a data set.

[0014] Furthermore, for the arc road segment, the number of centerline points and the angular resolution are determined according to the design parameters, and the inner boundary point set of the arc segment is constructed in a loop. Outer boundary point set And the centerline point set

[0015] Furthermore, for the straight road segment, according to its design parameters, a set of boundary points within the straight road segment is constructed in a loop. Outer boundary point set And the centerline point set

[0016] Furthermore, for the straight road segment, the resolution of the points on the inner and outer boundaries of the road is adjusted to improve the misalignment of the points on the inner and outer boundaries of the road.

[0017] Furthermore, for straight road segments, the point resolution on the centerline is greater than the point resolution on the inner and outer boundaries of the road to ensure the accuracy of the centerline.

[0018] Furthermore, for the arc section road center angle θ crv , the radius of the arc section road centerline is R crv The selection range is limited according to the expected speed of the lane.

[0019] Furthermore, in step 4, new road boundary points are generated by uniform sampling, and the number of target boundary points after sampling adjustment ensures the spatial misalignment of points on the inner and outer boundaries.

[0020] Furthermore, Latin hypercube sampling is used for uniform sampling.

[0021] Furthermore, for the set topological structure, new road data is generated by sampling the inner boundary points, outer boundary points, and center line points of the topological structure multiple times.

[0022] Furthermore, based on the obtained lane boundary point set containing misaligned points, a point resolution of Δd is performed between two consecutive lane boundary points. interp The point interpolation is performed to encrypt the lane boundary point density, and then the nearest point on the outer or inner boundary is matched through the nearest neighbor search method to generate a new boundary to improve the point alignment on the inner and outer boundaries.

[0023] Beneficial effects of the present invention:

[0024] 1. A method for quickly generating road data containing vectorized information provided by the present invention can simulate the working conditions on actual roads, and can quickly and massively generate road data based on actual road parameters, thereby providing a sufficient road data basis for related experiments.

[0025] 2. The method of the present invention can simulate the conditions of point alignment and misalignment on lane boundaries, and can be used to verify the influence of point alignment on centerline generation.

[0026] 3. The present invention can generate roads with different topological structures, and can also generate different boundary point densities for each topological structure of the road to obtain different road data.

[0027] 4. The present invention can directly generate the center line of the road lane without the need for additional calculation.

[0028] 5. The road data obtained by the present invention can also be used to construct a road network for autonomous driving tests. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the three-section lane topology structure constructed in this embodiment.

[0030] Figure 2 It is a schematic diagram of some design parameters in the three-section lane topology structure constructed in this embodiment. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0032] A method for quickly generating road data containing lane boundary information comprises the following steps:

[0033] Step 1, construct a segmented lane topology structure; the segmented lane topology structure can be composed of at least one straight road segment and / or at least one arc road segment. For example, the present implementation case shows a three-segment road, which is composed of an arc road and straight roads (node ​​arms) connected at both ends of the arc road. In the specific implementation, the arc road is first constructed, and then two straight roads are constructed from the starting point and the end point of the arc road respectively. Finally, the first straight road segment, the arc road, and the second straight road segment are connected in sequence to generate the final road, such as Figure 1 shown.

[0034] Step 2: For the constructed three-segment road, initialize the design parameters in the segmented lane topology structure, including: the target length of the road is L road , the target width of the road is W road The design parameters of the arc road include: the starting angle of the arc road is θ init , the coordinates of the center of the circle are O(x,y), and the radius of the center line of the arc segment is R crv , the point resolution on the center line of the arc segment is The design parameters of the straight road segment include the original point resolution on the inner boundary and the outer boundary of the straight road segment. and The point resolution on the centerline of the straight road segment is The number of inner and outer boundary points of the straight segment road are and The intersection angle of the two node arms of the road is θ intxn ,like Figure 2 shown.

[0035] Step 3: Based on the above segmented lane topology and design parameters, the inner boundary point set, outer boundary point set and center line point set in the straight line segment road segment and the arc segment road segment are calculated respectively.

[0036] Furthermore, for the arc segment road segment, the steps of obtaining the arc segment road centerline point set according to its design parameters are as follows:

[0037] S1: According to the road node arm crossing angle θ intxn Calculate the center angle θ of the center line of the arc segment road crv :

[0038] θ crv =180-θ intxn (1)

[0039] S2: According to the center angle θ of the center line of the arc segment road crv And the centerline radius R crv Calculate the center line length Lcrv :

[0040]

[0041] S3: According to the center line length L crv and arc segment centerline point resolution Estimating the number of points on the center line

[0042]

[0043] In the formula, F ceil (·) means rounding up the result.

[0044] S4: According to the center angle θ of the center line of the arc segment road crv and the number of points on the center line Calculate the angular resolution Δθ of the circular arc segment crv :

[0045]

[0046] S5: Calculate the number of centerline points of the arc segment road:

[0047]

[0048] S6: construct points on the centerline of the circular arc segment road in a loop. Specifically, when the number of points generated on the outer boundary is less than n, When , construct the points on the arc segment boundary in a clockwise manner and execute S7 and S8; otherwise, jump to step S9.

[0049] S7: Calculate the angle of the i-th boundary point to be generated relative to the x-axis:

[0050]

[0051] S8: Calculate the coordinates of the i-th centerline point to be generated relative to the circle center O(x, y) and update the number of accumulated generated points n:

[0052]

[0053] n=n+1 (11)

[0054] S9: Get the centerline point set of the arc segment road: in, is the centerline point of the i-th arc segment.

[0055] Furthermore, for the arc segment road segment, the steps of obtaining the inner boundary point set and the outer boundary point set in the arc segment road according to its design parameters are as follows:

[0056] S1: According to the road node arm crossing angle θ intxn Calculate the center angle θ of the center line of the arc segment road crv :

[0057] θ crv =180-θ intxn (12)

[0058] S2: According to the center angle θ of the center line of the arc segment road crv And the centerline radius R crv Calculate the center line length L crv :

[0059]

[0060] S3: According to the center line length L crv and arc segment centerline point resolution Estimating the number of points on the center line

[0061]

[0062] In the formula, F ceil (·) means rounding up the result.

[0063] S4: According to the center angle θ of the center line of the arc segment road crv and the number of points on the center line Calculate the angular resolution Δθ of the circular arc segment crv :

[0064]

[0065] S5: Based on the arc segment road angle resolution Δθ crv , calculate the number of inner and outer boundary points of the arc segment road:

[0066]

[0067] S6: According to the radius R of the center line of the arc segment road crv and the road target width W road Calculate the inner boundary radius R of the arc segment road innr :

[0068] R innr =R crv -W road (17)

[0069] S7: construct points on the inner boundary of the arc segment road in a loop. Specifically, when the number of points n generated on the inner boundary is less than When , construct the points on the arc segment boundary in a clockwise manner and execute S8 and S9; otherwise, jump to S10.

[0070] S8: Calculate the angle of the i-th boundary point to be generated relative to the x-axis:

[0071]

[0072] S9: Calculate the coordinates of the i-th boundary point to be generated relative to the circle center O(x, y) and update the number of accumulated generated points n:

[0073]

[0074] n=n+1 (23)

[0075] Among them, O x , O y The x-coordinate and y-coordinate of the center O are respectively, is the change in x coordinates of the i boundary points, is the change in the y coordinate of the i boundary points, are the x-coordinate and y-coordinate of the i-th boundary point.

[0076] S10: Obtain the inner boundary point set of the arc segment road: in, is the inner boundary point of the i-th arc segment road.

[0077] S11: Record the angle of the last generated point relative to the x-axis

[0078]

[0079] S12: According to the radius R of the center line of the arc segment road crv and the road target width W road Calculate the outer boundary radius R of the arc segment road outr :

[0080] R outr =R crv +W road (25)

[0081] S13: construct points on the outer boundary of the circular arc segment road in a loop. Specifically, when the number of points n generated on the outer boundary is less than When , construct the points on the arc segment boundary in clockwise direction and execute S14 and S15; otherwise, jump to step S16.

[0082] S14: Calculate the angle of the i-th boundary point to be generated relative to the x-axis:

[0083]

[0084] S15: Calculate the coordinates of the i-th boundary point to be generated relative to the circle center O(x, y) and update the number of accumulated generated points n:

[0085]

[0086] n=n+1 (31)

[0087] S16: Obtain the set of outer boundary points of the arc segment road: in, is the outer boundary point of the i-th arc segment road.

[0088] Furthermore, for the straight road segment, according to its design parameters, a straight line segment inner boundary point set, an outer boundary point set and a center line point set are constructed in a loop; the specific steps are as follows:

[0089] S1: Create the first node arm and set its length to L arm_1 The initial points of the inner and outer boundaries and center line of the first node arm are

[0090] S2: Set the direction of the node arm road to:

[0091] θ arm_1 =θ init +90 (32)

[0092] S3: Calculate the unit direction vector of the inner boundary of the node arm:

[0093]

[0094] in, are the changes in the x- and y-coordinates of the first node arm, v innr is the unit direction vector of the inner boundary of the first node arm.

[0095] S4: Construct the points on the inner boundary of the first node arm road in a loop. Specifically, when the cumulative length of the i-th generated point on the inner boundary is Less than L arm_1 , execute step S5; otherwise, jump to step S6.

[0096] S5: Construct the first node arm boundary point And update the cumulative length

[0097]

[0098] S6: Get the boundary point set within the first node arm road: in is the number of boundary points in the first node arm finally generated.

[0099] S7: Calculate the unit direction vector of the outer boundary of the first node arm:

[0100]

[0101] S8: Construct points on the outer boundary of the node arm road in a loop. Specifically, when the cumulative length of the i-th generated point on the outer boundary Less than L arm_1 , execute step S9; otherwise, jump to step S10.

[0102] S9: Construct node arm boundary points And update the cumulative length

[0103]

[0104] S10: Get the outer boundary point set of the first node arm road: in is the number of outer boundary points finally generated.

[0105] S11: Calculate the unit direction vector of the center line of the node arm:

[0106]

[0107] S12: Construct points on the center line of the node arm in a loop. Specifically, when the cumulative length of the i-th generated point on the center line Less than L arm_1 , execute step S13; otherwise, jump to step S14.

[0108] S13: Constructing the centerline points of the node arms And update the cumulative length

[0109]

[0110] S14: Get the centerline point set of the first node arm road: in The number of centerline points finally generated.

[0111] S15: Create the second node arm and set its length to L arm_2 The initial points of the inner and outer boundaries and center line of the second node arm are

[0112] S16: Set the direction of the node arm road to:

[0113]

[0114] S17: Calculate the unit direction vector of the inner boundary of the node arm:

[0115]

[0116] S18: Construct points on the inner boundary of the node arm road in a loop. Specifically, when the cumulative length of the i-th generated point on the inner boundary Less than L arm_2 , execute step S19; otherwise, jump to step S20.

[0117] S19: Construct node arm boundary points And update the cumulative length

[0118]

[0119] S20: Get the boundary point set inside the second node arm road: in is the number of inner boundary points finally generated.

[0120] S21: Calculate the unit direction vector of the outer boundary of the node arm:

[0121]

[0122] S22: construct points on the outer boundary of the node arm road in a loop. Specifically, when the cumulative length of the i-th generated point on the outer boundary Less than L arm_2 , execute S23; otherwise, jump to S24.

[0123] S23: Construct node arm boundary points And update the cumulative length

[0124]

[0125] S24: Get the outer boundary point set of the second node arm road: in is the number of outer boundary points finally generated.

[0126] S25: Calculate the unit direction vector of the center line of the node arm:

[0127]

[0128] S26: construct points on the centerline of the node arm road in a loop. Specifically, when the cumulative length of the i-th generated point on the centerline Less than L arm_2 , execute S27; otherwise, jump to S28.

[0129] S27: Construct node arm centerline points And update the cumulative length

[0130]

[0131] S28: Get the centerline point set of the second node arm road: in The number of centerline points finally generated.

[0132] Step 4: Based on the inner boundary point set, outer boundary point set and center line point set of the straight road segment and / or arc road segment obtained in step 3, a lane boundary point set containing misaligned points is extracted by sampling to form a data set. The specific steps are as follows:

[0133] S4.1: Construct a road boundary point set based on the inner and outer boundary points of the constructed node arms and arc segments. The road inner boundary point set is:

[0134]

[0135] Among them, F flip (·) Reverse the order of the point set.

[0136] The outer boundary point set is:

[0137]

[0138] The centerline point set is:

[0139]

[0140] The number of inner and outer boundary and centerline point sets are:

[0141]

[0142]

[0143] The serial numbers of the inner and outer boundary point sets are:

[0144]

[0145] S4.2: Generate a set of road boundary points Perform Latin hypercube sampling to generate Random numbers:

[0146]

[0147] The i-th random number The range is:

[0148]

[0149] S4.3: Calculate the sequence number of the inner boundary points that need to be retained:

[0150]

[0151] S4.4: Extract the final road boundary points:

[0152]

[0153] At the same time, the first and last points of the inner boundary need to be retained: and The final set of points is:

[0154]

[0155] S4.5: Generate a set of road outer boundary points Perform Latin hypercube sampling to generate Random numbers:

[0156]

[0157] The i-th random number The range is:

[0158]

[0159] S4.6: Calculate the sequence number of the outer boundary points that need to be retained:

[0160]

[0161] S4.7: Extract the final road outer boundary points:

[0162]

[0163] At the same time, the first and last points of the inner boundary need to be retained: and The final set of points is:

[0164]

[0165] S4.8: Finally, a road dataset is generated, where the road boundary is represented by a point set and Indicates that the road centerline is Indicates that the number of inner and outer boundary points are

[0166] In this embodiment, for a straight road segment, the resolution of points on the inner and outer boundaries of the road is adjusted to improve the misalignment of points on the inner and outer boundaries of the road.

[0167] In this embodiment, for a straight road segment, the point resolution on the centerline is greater than the point resolution on the inner and outer boundaries of the road to ensure the accuracy of the centerline.

[0168] In this embodiment, for the arc segment road center angle θ crv , the radius of the arc section road centerline is R crv The selection range is limited according to the expected speed of the lane.

[0169] In this embodiment, based on the constructed road topology structure, new road data is generated by sampling the inner boundary points, outer boundary points, and center line points of the topology structure multiple times.

[0170] In this embodiment, the target road width W between adjacent road sections is road It can be consistent or inconsistent.

[0171] In this embodiment, based on the obtained lane boundary point set containing the misaligned points, a point resolution of Δd is performed between two consecutive lane boundary points on the inner boundary (or outer boundary). interp The point difference value is used to encrypt the inner boundary (or outer boundary) lane boundary point density; based on the boundary points on the outer boundary (or inner boundary), the nearest points on the inner boundary (or outer boundary) are matched through the nearest neighbor search method to generate a new boundary to improve the alignment of the points on the inner and outer boundaries. Taking the insertion of points on the inner boundary as an example, the specific steps are as follows:

[0172] S1. Interpolate between two adjacent boundary points on the inner boundary in a loop. Specifically, when the sequence number i of the point on the inner boundary is less than If yes, execute S2; otherwise, jump to S6.

[0173] S2. Extract the inner boundary points with serial numbers i and i+1 and And construct a unit vector from point i to point i+1:

[0174]

[0175] Among them, lgth i,i+1 is the distance between point i and point i+1.

[0176] S3. As the starting point, create interpolation points in a loop. When the cumulative distance of the created mth difference point is Less than lgth i,i+1When , execute S4; otherwise, execute S5.

[0177] S4. Create interpolation points, generate interpolation point coordinates and update the cumulative length of the interpolation points

[0178]

[0179] S5. The set of interpolation points between the i-th and i+1-th points on the inner boundary is:

[0180]

[0181] Where M is the total number of difference points generated.

[0182] S6. After interpolation, the set of inner boundary points is The number of points is:

[0183]

[0184] S7. Starting from the first point on the outer boundary, match the points on the right boundary using the nearest neighbor method in a loop. When , execute S8; otherwise, execute S9.

[0185] S8. Select the i-th point on the outer boundary search On The nearest point is stored in middle.

[0186] S9. Finally, a road dataset is generated, where the road boundary is represented by a point set and express.

[0187] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A method for quickly generating road data containing vectorized information, characterized in that: The steps include: Step 1: construct a segmented lane topology structure from at least one straight road segment and / or at least one arc road segment; Step 2: Initialize and set the design parameters in the segmented lane topology structure, including the target road length L road 、Road target width W road , design parameters of the straight segment road, design parameters of the arc segment road; The design parameters of the straight road segment include but are not limited to the original point resolution of the inner boundary of the straight road. Original point resolution of the outer boundary of the straight road The original point resolution of the center line of the straight road is The design parameters of the arc segment road include but are not limited to the arc segment road center angle θ crv , the radius of the arc section road centerline is R crv , the point resolution on the center line of the arc segment is Step 3: Based on the above segmented lane topology and design parameters, calculate the inner boundary point set, outer boundary point set and center line point set in the straight line segment road segment and / or the arc segment road segment respectively; Step 4: Based on the inner boundary point set, outer boundary point set and center line point set of the straight segment road segment and / or arc segment road segment obtained in step 3, a lane boundary point set containing misaligned points is extracted by sampling to form a data set.

2. A method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For arc road segments, the number of centerline points and angular resolution are determined according to their design parameters, and the inner boundary point set of the arc segment is constructed in a loop. Outer boundary point set And the centerline point set 3. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For a straight road segment, according to its design parameters, a set of boundary points within the straight road segment is constructed in a loop. Outer boundary point set And the centerline point set 4. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For straight road segments, the resolution of points on the inner and outer boundaries of the road is adjusted to improve the misalignment of points on the inner and outer boundaries of the road.

5. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For straight road segments, the point resolution on the centerline is greater than the point resolution on the inner and outer boundaries of the road to ensure the accuracy of the centerline.

6. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For the arc section road center angle θ crv , the radius of the arc section road centerline is R crv The selection range is limited according to the expected speed of the lane.

7. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: In step 4, a uniform sampling method is used to generate new road boundary points. The number of target boundary points after sampling adjustment ensures the spatial misalignment of points on the inner and outer boundaries.

8. A method for quickly generating road data containing vectorized information according to claim 7, characterized in that: Latin hypercube sampling is used for uniform sampling.

9. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: For the set topological structure, new road data is generated by sampling the inner boundary points, outer boundary points and center line points of the topological structure multiple times.

10. The method for quickly generating road data containing vectorized information according to claim 1, characterized in that: Based on the obtained lane boundary point set containing misaligned points, a point resolution of Δd is performed between two consecutive lane boundary points. interp The point interpolation is performed to encrypt the lane boundary point density, and then the nearest point on the outer or inner boundary is matched through the nearest neighbor search method to generate a new boundary to improve the point alignment on the inner and outer boundaries.

Images