3D geometric and semantic awareness with deep learning for autonomous guidance
The technology disclosed includes systems and methods for using a deep learning trained classifier to detect obstacles and pathways in an environment in which a robot moves. The system includes logic to receive image information captured by at least one visual spectrum-capable camera and location information captured by at least one depth measuring camera located on a mobile platform. The system includes logic to determine a three-dimensional 3D point cloud of points having 3D information. The system can determine, using an ensemble of trained neural network classifiers, an identity for objects. The system includes logic to determine an occupancy map of the environment. The system includes logic to provide the occupancy map to a process for initiating robot movement to avoid objects in the occupancy map of the environment.
1 . A method for using a deep learning trained classifier to detect obstacles and pathways in an environment in which a robot moves, based upon image information as captured by at least one visual spectrum-capable camera that captures images in a visual spectrum (RGB) range and at least one depth measuring camera, the method comprising:
receiving image information captured by at least one visual spectrum-capable camera and location information captured by at least one depth measuring camera located on a mobile platform;
extracting, by a processor, from the image information, features in the environment;
determining, by a processor, a three-dimensional 3D point cloud of points having 3D information including location information from the depth camera and the at least one visual spectrum-capable camera, the points corresponding to the features in the environment as extracted;
determining, by a processor, using an ensemble of trained neural network classifiers, including first trained neural network classifiers, an identity for objects corresponding to the features as extracted from the images;
determining, by a processor, from the 3D point cloud and the identity for objects as determined using the ensemble of trained neural network classifiers, an occupancy map of the environment; and
providing the occupancy map to a process for initiating robot movement to avoid objects in the occupancy map of the environment.
2 . The method of claim 1 ,
wherein the depth camera is tightly coupled with the at least one visual spectrum-capable camera by (i) an overlapping of fields of view; (ii) a calibration of pixels per unit area of field of view; and (iii) a synchronous capture of images; thereby enabling locations and features of objects to correspond to one another in sets of images captured by the cameras.
3 . The method of claim 2 ,
wherein the calibration of pixels per unit area of field of view of the at least one visual spectrum-capable camera and the depth camera is 1:1.
4 . The method of claim 2 ,
wherein the calibration of pixels per unit area of field of view of the at least one visual spectrum-capable camera and the depth camera is 16:1.
5 . The method of claim 2 ,
wherein the calibration of pixels per unit area of field of view of the at least one visual spectrum-capable camera and the depth camera is 24:1.
6 . The method of claim 1 , further including:
annotating, by a processor, the occupancy map with annotations of object identities at locations corresponding to at least some of the points in the 3D point cloud; and
using the occupancy map as annotated to plan paths to avoid certain ones of objects based upon identity and location.
7 . The method of claim 6 , wherein the occupancy map is one of a 3D map and a 2D grid representation of a 3D map.
8 . The method of claim 1 , further including:
determining, by a processor, using second trained neural network classifiers, an identity for a room based upon objects identified that correspond to the features as extracted from the images;
annotating, by a processor, the occupancy map with annotations of room identities at locations corresponding to at least some of the points in the 3D point cloud; and
using the occupancy map as annotated to plan paths to remain within or to avoid certain ones of rooms based upon identity and location.
9 . The method of claim 1 ,
wherein the field of view of the at least one visual spectrum-capable camera is 1920×1080 and the depth camera is 224×172.
10 . The method of claim 1 ,
wherein the field of view of the depth camera is within one of: (i) a range including −20 and +20 about a principal axis of the depth camera; and (ii) a range of −30 and +30 about a principal axis of the depth camera.
11 . The method of claim 1 ,
wherein alignment for an angle between a principal axis of the depth camera and a principal axis of the robot along a horizontal plane includes values between 0 degrees and +/−30 degrees.
12 . The method of claim 1 ,
wherein alignment for an angle between a principal axis of the depth camera and a principal axis of the robot along a horizontal plane includes values between 0 degrees and +/−45 degrees.
13 . The method of claim 1 ,
wherein alignment for an angle between a principal axis of the depth camera and a principal axis of the robot along a horizontal plane includes values between 0 degrees and +/−90 degrees.
14 . The method of claim 1 ,
wherein alignment for an angle between a principal axis of the depth camera and a principal axis of the robot along a horizontal plane includes values between 0 degrees and +/−120 degrees.
15 . The method of claim 1 , wherein trained neural network classifiers implement convolutional neural networks (CNN).
16 . The method of claim 1 , further including employing trained neural network classifiers implementing recursive neural networks (RNN) for time-based information.
17 . The method of claim 1 , further including employing trained neural network classifiers implementing long short-term memory networks (LSTM) for time-based information.
18 . The method of claim 1 , wherein the ensemble of neural network classifiers includes:
80 levels in total, from an input to an output.
19 . The method of claim 1 , wherein the ensemble of neural network classifiers implements a multi-layer convolutional network.
20 . The method of claim 19 , wherein the multi-layer convolutional network includes:
60 convolutional levels.
21 . The method of claim 1 , wherein the ensemble of neural network classifiers includes:
normal convolutional levels and depth-wise convolutional levels.
22 . A robot system comprising:
a mobile platform having disposed thereon:
at least one visual spectrum-capable camera to capture images in a visual spectrum (RGB) range;
at least one depth measuring camera; and
an interface to a host including one or more processors coupled to a memory storing instructions to implement a method comprising:
receiving image information captured by at least one visual spectrum-capable camera and location information captured by at least one depth measuring camera located on a mobile platform;
extracting, by a processor, from the image information, features in an environment;
determining, by a processor, a three-dimensional 3D point cloud of points having 3D information including location information from the depth camera and the at least one visual spectrum-capable camera, the points corresponding to the features in the environment as extracted;
determining, by a processor, using an ensemble of trained neural network classifiers, including first trained neural network classifiers, an identity for objects corresponding to the features as extracted from the images;
determining, by a processor, from the 3D point cloud and the identity for objects as determined using the ensemble of trained neural network classifiers, an occupancy map of the environment; and
providing the occupancy map to a process for initiating robot movement to avoid objects in the occupancy map of the environment.
23 . A non-transitory computer readable medium comprising stored instructions, which when executed by a processor, cause the processor to implement a method comprising:
receiving image information captured by at least one visual spectrum-capable camera and location information captured by at least one depth measuring camera located on a mobile platform;
extracting, by a processor, from the image information, features in the environment;
determining, by a processor, a three-dimensional 3D point cloud of points having 3D information including location information from the depth camera and the at least one visual spectrum-capable camera, the points corresponding to the features in the environment as extracted;
determining, by a processor, using an ensemble of trained neural network classifiers, including first trained neural network classifiers, an identity for objects corresponding to the features as extracted from the images;
determining, by a processor, from the 3D point cloud and the identity for objects as determined using the ensemble of trained neural network classifiers, an occupancy map of the environment; and
providing the occupancy map to a process for initiating robot movement to avoid objects in the occupancy map of the environment.