Method and apparatus for outputting obstacle information

1. A method for outputting obstacle information by a vehicle comprising:

obtaining three-dimensional point cloud data collected by various vehicle sensors disposed on the vehicle;

determining a set of two-dimensional plane coordinates corresponding to a target obstacle point cloud based on the three-dimensional point cloud data;

determining a candidate direction information set of the target obstacle point cloud based on the set of two-dimensional plane coordinates;

for each piece of candidate direction information in the candidate direction information set, determining a target value of the target obstacle point cloud in a direction indicated by the each piece of candidate direction information, based on the target obstacle point cloud and a circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information;

defining candidate direction information including a minimum target value in the candidate direction information set as direction information corresponding to the target obstacle point cloud;

outputting the direction information corresponding to the target obstacle point cloud; and

tracking an obstacle corresponding to the target obstacle point cloud for route planning based on the direction information corresponding to the target obstacle point cloud, wherein the obstacle corresponds to a physical object within a distance from the vehicle,

wherein the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is determined by: determining a plurality of circumscribing rectangles for the physical object each corresponding to the various vehicle sensors providing the three-dimensional point cloud data, and selecting a circumscribing rectangle of the plurality of circumscribing rectangles with the smallest width-to-length ratio, and wherein tracking the obstacle is further based on the selected circumscribing rectangle, and

wherein the method is performed by at least one processor.

2. The method according to claim 1 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information, based on the target obstacle point cloud and the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information comprises:

performing following target value calculation: determining the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information; calculating an area and a width-to-length ratio of the determined circumscribing rectangle; determining a symmetry point count in the determined circumscribing rectangle of the target obstacle point cloud; and determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and is negatively correlated with the determined symmetry point count; wherein the symmetry point count is a number of symmetry points in the determined circumscribing rectangle with respect to a geometric center line of the determined circumscribing rectangle in a long side direction.

3. The method according to claim 2 , wherein the target obstacle point cloud is set with an associated obstacle type including a motor vehicle or a non-motor vehicle.

4. The method according to claim 3 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count comprises:

defining an average value of first distances of points in the target obstacle point cloud as a first distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the motor vehicle, wherein a first distance of a point in the target obstacle point cloud is a minimum value of distances from the point to four sides of the determined circumscribing rectangle; and

determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio, the determined symmetry point count, and the determined first distance average value, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and the determined first distance average value, and is negatively correlated with the determined symmetry point count.

5. The method according to claim 3 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count comprises:

defining an average value of second distances of points in the target obstacle point cloud as a second distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the non-motor vehicle, wherein a second distance of a point in the target obstacle point cloud is a distance from the point to a central axis line segment, and the central axis line segment is the geometric center line of the determined circumscribing rectangle along the direction indicated by the each piece of candidate direction information; and

determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio, the determined symmetry point count, and the determined second distance average value, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and the determined second distance average value, and is negatively correlated with the determined symmetry point count.

6. The method according to claim 5 , wherein before the outputting the direction information corresponding to the target obstacle point cloud, the method further comprises:

defining the direction information corresponding to the target obstacle point cloud as target direction information;

defining the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the target direction information as a target circumscribing rectangle; and

the outputting the direction information corresponding to the target obstacle point cloud comprises:

outputting the target direction information and at least one of: geometric center coordinates, a length or a width of the target circumscribing rectangle.

7. The method according to claim 6 , wherein the outputting the target direction information and at least one of: geometric center coordinates, the length or the width of the target circumscribing rectangle comprises:

determining a confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle; and

outputting the target direction information and at least one of: the geometric center coordinates, the length or the width of the target circumscribing rectangle, or the determined confidence level.

8. The method according to claim 7 , wherein the determining the confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle comprises:

determining a circumscribing cuboid of the target obstacle point cloud based on the target circumscribing rectangle, in response to determining the associated obstacle type of the target obstacle point cloud being the motor vehicle;

defining an average value of third distances of points in the target obstacle point cloud as a third distance average value, wherein a third distance of a point in the target obstacle point cloud is a minimum value of distances from the point to six sides of the circumscribing cuboid; and

determining the confidence level of the target direction information based on the third distance average value, wherein the determined confidence level is negatively correlated with the third distance average value.

9. The method according to claim 7 , wherein the determining the confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle comprises:

defining an average value of fourth distances of points in the target obstacle point cloud as a fourth distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the non-motor vehicle, wherein a fourth distance of a point in the target obstacle point cloud is a distance from the point to a geometric center line of the target circumscribing rectangle along the direction indicated by the target direction information; and

determining the confidence level of the target direction information based on the fourth distance average value, wherein the determined confidence level is negatively correlated with the fourth distance average value.

10. An apparatus for outputting obstacle information, comprising:

at least one processor; and

a memory storing instructions, the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:

obtaining three-dimensional point cloud data collected by various vehicle sensors disposed on a vehicle;

determining a set of two-dimensional plane coordinates corresponding to a target obstacle point cloud based on the three-dimensional point cloud data;

determining a candidate direction information set of the target obstacle point cloud based on the set of two-dimensional plane coordinates;

for each piece of candidate direction information in the candidate direction information set, determining a target value of the target obstacle point cloud in a direction indicated by the each piece of candidate direction information, based on the target obstacle point cloud and a circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information;

defining candidate direction information including a minimum target value in the candidate direction information set as direction information corresponding to the target obstacle point cloud;

outputting the direction information corresponding to the target obstacle point cloud; and

tracking an obstacle corresponding to the target obstacle point cloud for route planning based on the direction information corresponding to the target obstacle point cloud, wherein the obstacle corresponds to a physical object within a distance from the vehicle, and

wherein the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is determined by: determining a plurality of circumscribing rectangles for the physical object each corresponding to the various vehicle sensors providing the three-dimensional point cloud data, and selecting a circumscribing rectangle of the plurality of circumscribing rectangles with the smallest width-to-length ratio, and wherein tracking the obstacle is further based on the selected circumscribing rectangle.

11. The apparatus according to claim 10 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the target obstacle point cloud and the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information comprises:

performing following target value calculation: determining the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information; calculating an area and a width-to-length ratio of the determined circumscribing rectangle; determining a symmetry point count in the determined circumscribing rectangle of the target obstacle point cloud; and determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and is negatively correlated with the determined symmetry point count; wherein the symmetry point count is a number of symmetry points in the determined circumscribing rectangle with respect to a geometric center line of the determined circumscribing rectangle in a long side direction.

12. The apparatus according to claim 11 , wherein the target obstacle point cloud is set with an associated obstacle type including a motor vehicle or a non-motor vehicle.

13. The apparatus according to claim 12 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count comprises:

defining an average value of first distances of points in the target obstacle point cloud as a first distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the motor vehicle, wherein a first distance of a point in the target obstacle point cloud is a minimum value of distances from the point to four sides of the determined circumscribing rectangle; and

determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio, the determined symmetry point count, and the determined first distance average value, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and the determined first distance average value, and is negatively correlated with the determined symmetry point count.

14. The apparatus according to claim 12 , wherein the determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio and the determined symmetry point count comprises:

defining an average value of second distances of points in the target obstacle point cloud as a second distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the non-motor vehicle, wherein a second distance of a point in the target obstacle point cloud is a distance from the point to a central axis line segment, and the central axis line segment is the geometric center line of the determined circumscribing rectangle along the direction indicated by the each piece of candidate direction information; and

determining the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information based on the calculated area and width-to-length ratio, the determined symmetry point count, and the determined second distance average value, wherein the target value of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is positively correlated with the calculated area and width-to-length ratio, and the determined second distance average value, and is negatively correlated with the determined symmetry point count.

15. The apparatus according to claim 14 , the operations further comprising:

defining the direction information corresponding to the target obstacle point cloud as target direction information;

defining the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the target direction information as a target circumscribing rectangle; and

the outputting the direction information corresponding to the target obstacle point cloud comprises:

outputting the target direction information and at least one of: geometric center coordinates, a length or a width of the target circumscribing rectangle.

16. The apparatus according to claim 15 , wherein the outputting the target direction information and at least one of: geometric center coordinates, the length or the width of the target circumscribing rectangle comprises:

determining a confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle; and

outputting the target direction information and at least one of: the geometric center coordinates, the length or the width of the target circumscribing rectangle, or the determined confidence level.

17. The apparatus according to claim 16 , wherein the determining the confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle comprises:

determining a circumscribing cuboid of the target obstacle point cloud based on the target circumscribing rectangle, in response to determining the associated obstacle type of the target obstacle point cloud being the motor vehicle;

defining an average value of third distances of points in the target obstacle point cloud as a third distance average value, wherein a third distance of a point in the target obstacle point cloud is a minimum value of distances from the point to six sides of the circumscribing cuboid; and

determining the confidence level of the target direction information based on the third distance average value, wherein the determined confidence level is negatively correlated with the third distance average value.

18. The apparatus according to claim 16 , wherein the determining the confidence level of the target direction information based on the target obstacle point cloud and the target circumscribing rectangle comprises:

defining an average value of fourth distances of points in the target obstacle point cloud as a fourth distance average value, in response to determining the associated obstacle type of the target obstacle point cloud being the non-motor vehicle, wherein a fourth distance of a point in the target obstacle point cloud is a distance from the point to a geometric center line of the target circumscribing rectangle along the direction indicated by the target direction information; and

determining the confidence level of the target direction information based on the fourth distance average value, wherein the determined confidence level is negatively correlated with the fourth distance average value.

19. A non-transitory computer storage medium storing a computer program, the computer program when executed by one or more processors, causes the one or more processors to perform operations, the operations comprising:

obtaining three-dimensional point cloud data collected by various vehicle sensors disposed on a vehicle;

determining a set of two-dimensional plane coordinates corresponding to a target obstacle point cloud based on the three-dimensional point cloud data;

determining a candidate direction information set of the target obstacle point cloud based on the set of two-dimensional plane coordinates;

for each piece of candidate direction information in the candidate direction information set, determining a target value of the target obstacle point cloud in a direction indicated by the each piece of candidate direction information based on the target obstacle point cloud and a circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information;

defining candidate direction information including a minimum target value in the candidate direction information set as direction information corresponding to the target obstacle point cloud;

outputting the direction information corresponding to the target obstacle point cloud; and

tracking an obstacle corresponding to the target obstacle point cloud for route planning based on the direction information corresponding to the target obstacle point cloud, wherein the obstacle corresponds to a physical object within a distance from the vehicle, and

wherein the circumscribing rectangle of the target obstacle point cloud in the direction indicated by the each piece of candidate direction information is determined by: determining a plurality of circumscribing rectangles for the physical object each corresponding to the various vehicle sensors providing the three-dimensional point cloud data, and selecting a circumscribing rectangle of the plurality of circumscribing rectangles with the smallest width-to-length ratio, and wherein tracking the obstacle is further based on the selected circumscribing rectangle.