Systems and methods for extracting information about objects from scene information
Examples of various method and systems are provided for information extraction from scene information. 2D image information can be generated from 2D images of the scene that are overlapping at least part of one or more object(s). The 2D image information can be combined with 3D information about the scene incorporating at least part of the object(s) to generate projective geometry information. Clustered 3D information associated with the object(s) can be generated by partitioning and grouping 3D data points present in the 3D information. The clustered 3D information can be used to provide, e.g., measurement information, dimensions, geometric information, and/or topological information about the object(s). Segmented 2D information can also be generated from the 2D image information. Validated 2D and 3D information can be produced by cross-referencing between the projective geometry information, clustered 3D information, and/or segmented 2D image information, and used to label the object(s) in the scene.
1. A method of operating an autonomous vehicle in an environment comprising:
a. identifying an environment comprising a scene including one or more objects, wherein the environment has not previously been navigated by an autonomous vehicle;
b. introducing the autonomous vehicle into the environment, wherein the autonomous vehicle is configured with two-dimensional (2D) image capture capability;
c. navigating the autonomous vehicle around the scene and the one or more objects, wherein the navigation is by one or more of the autonomous vehicle, a computer, or a user;
d. acquiring a plurality of 2D images for the scene and the one or more objects via the autonomous vehicle;
e. processing the plurality of 2D images to generate 2D and 3D scene and object information, wherein the processing is by either or both of the autonomous vehicle or the computer and is conducted substantially in real time while the autonomous vehicle is navigating the environment;
f. processing the 2D and 3D scene and object information using either or both of projective geometry or a machine learning library comprising a priori scene and object information, wherein the processing is by either or both of the autonomous vehicle or the computer and is conducted substantially in real time while the autonomous vehicle is navigating the environment;
g. generating autonomous vehicle navigation information from the processed 2D and 3D scene and object information by either or the autonomous vehicle or the computer, wherein the autonomous vehicle navigation information is generated substantially in real time; and
h. incorporating the generated autonomous vehicle navigation information in the navigation of the autonomous vehicle in the environment.
2. The method of claim 1 further comprising:
generating, by the computer, 2D and 3D scene and object information associated with an unknown environment, wherein the generated 2D and 3D scene and object information comprises one or more of:
i. object geo location information;
ii. object geometry information; or
iii. object topology information.
3. The method of claim 1 , wherein the 2D and 3D scene and object information and autonomous vehicle navigation information is used in generating instructions for navigation of the autonomous vehicle in a different environment.
4. The method of claim 1 , wherein the 2D and 3D scene and object information or autonomous vehicle navigation information is used in generating instructions for navigation of the autonomous vehicle in a subsequent autonomous vehicle operating event.
5. The method of claim 1 , wherein the 2D and 3D scene and object information comprise for at least one of the one or more objects one or more of:
a. object detection;
b. object recognition;
c. object identification; and
d. semantic information.
6. The method of claim 1 , wherein environment information comprises information about an orientation of the one or more objects in the scene.
7. The method of claim 1 , further comprising generating a 3D reconstruction of either or both of:
a. the scene including the one or more objects; or
b. at least one of the one or more objects.
8. The method of claim 1 , further comprising incorporating the 2D and 3D scene and object information in a map or survey of the environment.
9. The method of claim 1 , wherein the method is configured to receive user inputs associated with some or all of:
a. image acquisition;
b. object selection;
c. autonomous vehicle navigation in the scene;
d. 2D image processing;
e. 2D and 3D information processing; and
f. scene and object information generation.
10. The method of claim 1 , wherein the autonomous vehicle is an unmanned aerial vehicle.
11. A method of operating an autonomous vehicle in an unknown environment comprising:
a. providing navigation instructions for an autonomous vehicle, wherein the navigation instructions are derived from information generated from operation of the autonomous vehicle in a first unknown environment according to a process comprising:
i. acquiring a plurality of 2D images from the first unknown environment, wherein the first unknown environment comprises a scene and one or more objects, and wherein the plurality of 2D images are acquired via the autonomous vehicle when the autonomous vehicle is operating in the first unknown environment;
ii. processing the plurality of 2D images to generate 2D and 3D scene and object information for the first unknown environment, wherein the processing is by the autonomous vehicle or a computer and is conducted substantially in real time while the autonomous vehicle is operating in the first unknown environment; and
iii. processing the 2D and 3D scene and object information using projective geometry or a machine learning library comprising a priori scene and object information, wherein the processing is by the autonomous vehicle or the computer and is conducted substantially in real time while the autonomous vehicle is moving in the first unknown environment; and
b. operating an autonomous vehicle in a second unknown environment according to the provided autonomous vehicle navigation instructions.
12. The method of claim 11 , wherein the autonomous vehicle operated in the second unknown environment is the same or different from the autonomous vehicle operated in the first unknown environment.
13. The method of claim 11 , wherein the navigation instructions generated from the first unknown environment include information derived from one or more of:
i. object geo location information;
ii. object geometry information; or
iii. object topology information.
14. The method of claim 11 , wherein the navigation instructions generated from the first unknown environment include information derived from one or more of:
a. object detection;
b. object recognition;
c. object identification; and
d. semantic information.
15. The method of claim 11 , wherein the navigation instructions generated from the first unknown environment include information derived from an orientation of at least one one object in the scene.
16. The method of claim 11 , wherein the navigation instructions generated from the first unknown environment include information derived user inputs associated with one or more of:
a. image acquisition;
b. object selection;
c. autonomous vehicle navigation in the scene;
d. 2D image processing;
e. 2D and 3D information processing; and
f. scene and object information generation.
17. The method of claim 11 , further comprising incorporating information derived from operation of the autonomous vehicle in the second unknown environment in the generation of navigation instructions configured for operation of one or more autonomous vehicles in one or more different unknown environments.
18. The method of claim 11 , wherein information associated with the generated navigation instructions is included in a machine learning library for use as a priori information in one or more unknown environments in one or more subsequent autonomous vehicle operations.
19. The method of claim 11 , wherein information associated with the generated navigation instructions is incorporated in a map or survey of the second unknown environment.
20. The method of claim 11 , wherein the autonomous vehicle is an unmanned aerial vehicle.