Systems and methods for producing amodal cuboids
This document discloses system, method, and computer program product embodiments for detecting an object. For example, the method includes generating a plurality of cuboids by performing the following operations: defining a plurality of first cuboids each encompassing lidar data points that are plotted on a respective 3D graph of a plurality of 3D graphs; accumulating the lidar data points encompassed by the plurality of first cuboids; computing an extent using the accumulated lidar data points; and defining a second cuboid that has dimensions specified by the extent. The first cuboids and/or the second cuboid may be used to detect the object.
1 . A method for using an object detection in a real world environment, comprising:
generating, by the computing device, a plurality of cuboids by performing the following operations:
defining a plurality of first cuboids each encompassing lidar data points that are plotted on a respective 3D graph of a plurality of 3D graphs;
accumulating the lidar data points encompassed by the plurality of first cuboids;
computing an extent using the accumulated lidar data points;
defining a second cuboid that has dimensions specified by the extent;
selectively using, by the computing device, the plurality of first cuboids or the second cuboid to detect an object in the real world environment;
using the second cuboid to predict a trajectory for the object; and
executing one or more control instructions that cause a robotic system to adjust one or more autonomous movements based on the trajectory predicted for the object.
2 . The method according to claim 1 , further comprising controlling autonomous operations of a robotic system responsive to detection of the object in the real world environment.
3 . The method according to claim 1 , wherein the selectively using the plurality of first cuboids or the second cuboid to detect the object is based on a location of the object in the real world environment.
4 . The method according to claim 1 , wherein generating a plurality of cuboids further comprises:
defining a plurality of third cuboids by adjusting corner coordinates of the plurality of first cuboids; and
selecting coordinates for a center for the second cuboid based on the plurality of third cuboids.
5 . The method according to claim 4 , wherein the coordinates for the center of the second cuboid are selected by aligning a corner of the second cuboid with a closest visible corner of each one of said plurality of third cuboids.
6 . The method according to claim 4 , wherein the selecting coordinates for the center of the second cuboid is based additionally on the plurality of first cuboids.
7 . The method according to claim 6 , wherein the coordinates for the center of the second cuboid are selected by aligning a corner of the second cuboid such that a boundary precision between the second cuboid and each of the plurality of first cuboids is maximized.
8 . The method according to claim 1 , wherein the generating a plurality of cuboids further comprises removing ground points from the lidar data points encompassed by the plurality of first cuboids.
9 . The method according to claim 1 , further comprising:
generating a fourth cuboid using the plurality of first cuboids and the second cuboid; and
selectively using the fourth cuboid to detect the object in the real world environment in addition to or as an alternative to the plurality of first cuboids or the second cuboid.
10 . A system, comprising:
a processor;
a non-transitory computer-readable storage medium comprising programming instructions that are configured to cause the processor to implement a method for using an object detection in a real world environment, wherein the programming instructions comprise instructions to:
generate a plurality of cuboids by performing the following operations:
defining a plurality of first cuboids each encompassing lidar data points that are plotted on a respective 3D graph of a plurality of 3D graphs;
accumulating the lidar data points encompassed by the plurality of first cuboids;
computing an extent using the accumulated lidar data points;
defining a second cuboid that has dimensions specified by the extent;
selectively use the plurality of first cuboids or the second cuboid to detect an object in the real world environment;
use the second cuboid to predict a trajectory for the object; and
executing one or more control instructions that cause a robotic system to adjust one or more autonomous movements based on the trajectory predicted for the object.
11 . The system according to claim 10 , wherein the programming instructions further comprise instructions to control autonomous operations of a robotic system responsive to detection of the object in the real world environment.
12 . The system according to claim 10 , wherein the plurality of first cuboids or the second cuboid is selectively used to detect the object based on a location of the object in the real world environment.
13 . The system according to claim 10 , wherein the programming instructions further comprise instructions to:
define a plurality of third cuboids by adjusting corner coordinates of the plurality of first cuboids; and
select coordinates for a center for the second cuboid based on the plurality of third cuboids.
14 . The system according to claim 13 , wherein the coordinates for a center of the second cuboid are selected by aligning a corner of the second cuboid with a closest visible corner of each one of said plurality of third cuboids.
15 . The system according to claim 13 , wherein the coordinates for the center of the second cuboid are selected based additionally on the plurality of first cuboids.
16 . The system according to claim 15 , wherein the coordinates for the center of the second cuboid are selected by aligning a corner of the second cuboid such that a boundary precision between the second cuboid and each of the plurality of first cuboids is maximized.
17 . The system according to claim 10 , wherein the programming instructions further comprise instructions to remove ground points from the lidar data points encompassed by the plurality of first cuboids.
18 . A non-transitory computer-readable medium that stores instructions that, when executed by at least one computing device, will cause the at least one computing device to perform operations comprising:
generating a plurality of cuboids by performing the following operations:
defining a plurality of first cuboids each encompassing lidar data points that are plotted on a respective 3D graph of a plurality of 3D graphs;
accumulating the lidar data points encompassed by the plurality of first cuboids;
computing an extent using the accumulated lidar data points;
defining a second cuboid that has dimensions specified by the extent;
selectively using, by the computing device, the plurality of first cuboids or the second cuboid to detect the object in the real world environment;
using the second cuboid to predict a trajectory for the object; and
executing one or more control instructions that cause a robotic system to adjust one or more autonomous movements based on the trajectory predicted for the object.