Lane mask generation for autonomous machine
In various examples, object fence corresponding to objects detected by an ego-vehicle may be used to determine overlap of the object fences with lanes on a driving surface. A lane mask may be generated corresponding to the lanes on the driving surface, and the object fences may be compared to the lanes of the lane mask to determine the overlap. Where an object fence is located in more than one lane, a boundary scoring approach may be used to determine a ratio of overlap of the boundary fence, and thus the object, with each of the lanes. The overlap with one or more lanes for each object may be used to determine lane assignments for the objects, and the lane assignments may be used by the ego-vehicle to determine a path or trajectory along the driving surface.
1. A processor comprising:
processing circuitry to perform one or more operations of a machine in an environment using a mask, the mask generated for a lane, at least in part, by:
obtaining image data generated using an image sensor, the image data representative of an image depicting a field-of-view of the image sensor;
determining that first points of the image are associated with a first polyline corresponding to the lane and second points of the image are associated with a second polyline corresponding to the lane;
connecting the first points to generate the first polyline and the second points to generate the second polyline;
based at least on the connecting the first points and the second points, connecting, in a crisscross pattern, the first points of the first polyline to the second points of the second polyline to generate a plurality of triangles; and
generating a polygon corresponding to the lane.
2. The processor of claim 1 , wherein the mask is generated using a triangulation algorithm.
3. The processor of claim 2 , wherein the triangulation algorithm includes monotone polygon triangulation.
4. The processor of claim 1 , wherein the determining that the first points of the image are associated with the first polyline and the second points of the image are associated with the second polyline uses one or more deep neural networks (DNNs) processing the image data.
5. The processor of claim 1 , wherein the one or more operations include at least one of: assigning an object to a lane using the mask, determining a path or trajectory through at least a portion of the environment, performing obstacle avoidance, or updating a world model.
6. The processor of claim 1 , the performing the one or more operations is further executed using an object fence corresponding to an object detected in the environment.
7. The processor of claim 1 , wherein the one or more operations include assigning one or more objects to the lane represented by the mask.
8. A method comprising:
obtaining image data generated using an image sensor, the image data representative of an image;
determining, based at least on the image data, one or more first points of the image that are associated with a first polyline and one or more second points of the image that are associated with a second polyline;
connecting the one or more first points to the one or more second points to generate one or more triangles associated with the image;
determining one or more pixels of the image that are included in the one or more triangles;
generating a polygon that at least encloses the one or more pixels; and
performing one or more operations using a machine and based at least on the polygon.
9. A system comprising:
one or more processors to:
obtain image data generated using an image sensor, the image data representative of an image depicting a field-of-view of the image sensor;
determine at least a first location within the image of a first polyline corresponding to a lane and a second location within the image of a second polyline corresponding to the lane;
connect one or more first points of the first polyline to one or more second points of the second polyline to generate one or more triangles associated with the image;
determine one or more pixels of the image that are included in the one or more triangles;
generate a polygon that at least encloses the one or more pixels; and
perform one or more operations by a machine using the polygon.
10. The system of claim 9 , wherein the one or more processors are further to generate at least a second portion of the polygon using a lane extension algorithm to extend the lane beyond a detected portion of the lane.
11. The system of claim 10 , wherein the lane extension algorithm includes at least one of a curve fitting algorithm or a lane extrapolation algorithm.
12. The system of claim 9 , wherein the one or more processors are further to generate a second polygon corresponding to a virtual lane.
13. The system of claim 9 , wherein the performance of the one or more operations comprises:
comparing an object representation corresponding to an object to the polygon; and
determining, based at least on the comparing, to assign the object to the lane.
14. The system of claim 13 , wherein the one or more processors are further to generate an object fence using a bounding shape corresponding to the object.
15. The system of claim 9 , wherein the determination of the first location within the image of the first polyline and the second location within the image of the second polyline uses one or more deep neural networks (DNNs) processing the image data.
16. The system of claim 9 , wherein the system is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
17. The system of claim 9 , wherein the connection of the one or more first points of the first polyline to the one or more second points of the second polyline comprises:
connecting each of the one or more first points of the first polyline to at least a point from the one or more second points to generate one or more first triangles of the one or more triangles; and
connecting each of the one or more second points of the second polyline to at least a point from the one or more first points of the second polyline to generate one or more second triangles of the one or more triangles.
18. The method of claim 8 , wherein the generating the polygon comprises:
determining the one or more pixels of the image that are included within the plurality of triangles; and
generating the polygon to enclose the one or more pixels.
19. The method of claim 8 , wherein:
the one or more first points include a first plurality of points;
the one or more second points include a second plurality of points;
the method further comprises connecting the first plurality of points to generate the first polyline and the second plurality of points to generate the second polyline; and
the connecting the plurality of first points and the plurality of second points to generate the one or more triangles occurs after the connecting the plurality of first points to generate the first polyline and the plurality of second points to generate the second polyline.
20. The method of claim 8 , wherein the connecting the one or more first points to the one or more second points comprises:
connecting each of the one or more first points to at least a point from the one or more second points; and
connecting each of the one or more second points to at least a point from the one or more first points.