Autonomous robots and methods of operating the same
Autonomous robots and methods of operating the same are disclosed. An autonomous robot includes a sensor and memory including machine readable instructions. The autonomous robot further includes at least one processor to execute the instructions to generate a velocity costmap associated with an environment in which the robot is located. The processor generates the velocity costmap based on a source image captured by the sensor. The velocity costmap includes velocity information indicative of movement of an obstacle detected in the environment.
1. An autonomous robot comprising:
a sensor;
memory including machine readable instructions; and
at least one processor to execute the instructions to generate a velocity costmap associated with an environment in which the robot is located, the processor to generate the velocity costmap based on a source image captured by the sensor, the velocity costmap including velocity information indicative of movement of an obstacle detected in the environment, the generated velocity costmap corresponding to a combination of a first velocity costmap and a second velocity costmap different than the first velocity costmap, the first velocity costmap based on a first version of the source image, the second velocity costmap based on a second version of the source image, the second version different than the first version.
2. The autonomous robot as defined in claim 1 , wherein the the first version of the source image has a first resolution, the second version of the source image having a second resolution, the second resolution higher than the first resolution.
3. The autonomous robot as defined in claim 2 , further including a motion analyzer to generate a first three-dimensional motion field based on the first version of the source image and a second three-dimensional motion field based on the second version of the source image, the velocity information in the first and second velocity costmaps based on the respective first and second three-dimensional motion fields.
4. The autonomous robot as defined in claim 2 , further including a motion region identifier to identify regions of motion in the first velocity costmap, the second velocity costmap limited to portions of the second version of the source image corresponding to the identified regions of motion of the first velocity costmap.
5. The autonomous robot as defined in claim 4 , wherein the motion region identifier identifies the regions of motion by:
dividing the velocity costmap into multiple blocks, eac block including multiple cells of the velocity costmap; and
grouping adjacent blocks into clusters based on the velocity information and depth information associated with the cells in the adjacent blocks.
6. The autonomous robot as defined in claim 5 , wherein the motion region identifier identifier is to group the adjacent blocks into clusters based on a mean shift analysis of the velocity information and the depth information.
7. The autonomous robot as defined in claim 1 , further including an image resolution adjuster to reduce a resolution of the source image to a coarse image, the first version of the source image corresponding to the coarse image.
8. The autonomous robot as defined in claim 1 , wherein the velocity costmap is two-dimensional.
9. The autonomous robot as defined in claim 1 , wherein the velocity costmap is three-dimensional.
10. The autonomous robot as defined in claim 1 , wherein the sensor is a red-green-blue-depth (RGB-D) camera.
11. A non-transitory computer readable medium comprising instructions that, when executed, cause a robot to at least:
capture a source image of an environment in which the robot is located; and
generate a velocity costmap associated with the environment based on the source image, the velocity costmap including velocity information indicative of movement of an obstacle detected in the environment, the generated velocity costmap corresponding to a combination of a first velocity costmap and a second velocity costmap different than the first velocity costmap, the first velocity costmap based on a first version of the source image, the second velocity costmap based on a second version of the source image, the second version different than the first version.
12. The non-transitory computer readable medium as defined in claim 11 , wherein the first version of the source image has a first resolution, the second version of the source image having a second resolution, the second resolution higher than the first resolution.
13. The non-transitory computer readable medium as defined in claim 12 , wherein the instructions further cause the robot to identify regions of motion in the first velocity costmap, the second velocity costmap limited to portions of the second version of the source image corresponding to the identified regions of motion in the first velocity costmap.
14. The non-transitory computer readable medium as defined in claim 13 , wherein the instructions further cause the robot to:
divide the velocity costmap into multiple blocks, each block including multiple cells of the velocity costmap; and
group adjacent blocks into clusters based on the velocity information and depth information associated with the cells in the adjacent blocks.
15. A method comprising:
capturing a source image with a sensor, the source image including image data indicative of an environment and an obstacle in the environment; and
generating, by executing an instruction with at least one processor, a velocity costmap associated with the environment based on the source image, the velocity costmap including velocity information indicative of movement of the obstacle detected in the environment, the generated velocity costmap corresponding to a combination of a first velocity costmap and a second velocity costmap different than the first velocity costmap, the first velocity costmap based on a first version of the source image, the second velocity costmap based on a second version of the source image, the second version different than the first version.
16. The method as defined in claim 15 , wherein the first version of the source image has a first resolution, the second version of the source image having a second resolution, the second resolution higher than the first resolution.
17. The method as defined in claim 16 , further including identifying regions of motion in the first velocity costmap, the second velocity costmap limited to portions of the second version of the source image corresponding to the identified regions of motion in the first velocity costmap.
18. The method as defined in claim 17 , further including:
dividing the velocity costmap into multiple blocks, each block including multiple cells of the velocity costmap; and
grouping adjacent blocks into clusters based on the velocity information and depth information associated with the cells in the adjacent blocks.
19. An apparatus comprising:
means for capturing a source image of an environment; and
means for generating a velocity costmap associated with the environment based the source image, the velocity costmap including velocity information indicative of movement of an obstacle detected in the environment, the generated velocity costmap corresponding to a combination of a first velocity costmap and a second velocity costmap different than the first velocity costmap, the first velocity costmap based on a first version of the source image, the second velocity costmap based on a second version of the source image, the second version different than the first version.
20. The apparatus as defined in claim 19 , wherein the first version of the source image has a first resolution, the second version of the source image having a second resolution, the second resolution higher than the first resolution.
21. The apparatus as defined in claim 20 , further including means for identifying regions of motion in the first velocity costmap, the second velocity costmap limited to portions of the second version of the source image corresponding to the identified regions of motion in the first velocity costmap.
22. The apparatus as defined in claim 21 , wherein the means for identifying regions of motion in the first velocity costmap includes:
means for dividing the velocity costmap into multiple blocks, each block including multiple cells of the velocity costmap; and
means for grouping adjacent blocks into clusters based on the velocity information and depth information associated with the cells in the adjacent blocks.
23. The apparatus as defined in claim 19 , further including means for reducing a resolution of the source image to a coarse image, the velocity costmap generated based on the coarse image.