Systems and methods for accelerated obstacle detection for a robotic device
A device may receive, from a robotic device, time-of-flight sensor data that includes a point cloud derived from depth images of an environment with a floor and one or more obstacles, and may shift the depth images by a number of pixels to generate shifted depth images. The device may subtract the shifted depth images from the depth images to generate final images, and may calculate Euclidean distances associated with the final images. The device may generate masks for the final images associated with Euclidean distances that are less than a search radius, and may calculate a covariance matrix based on the masks. The device may calculate eigenvectors for the covariance matrix, and may generate an occupancy grid for the environment based on the eigenvectors for the covariance matrix.
1 . A method, comprising:
receiving, by a device and from a robotic device, time-of-flight sensor data that includes a point cloud of depth images of an environment with a floor and one or more obstacles;
shifting, by the device, the depth images by a number of pixels to generate shifted depth images;
subtracting, by the device, the shifted depth images from the depth images to generate final images;
calculating, by the device, Euclidean distances associated with the final images;
generating, by the device, masks for the final images associated with Euclidean distances that are less than a search radius;
calculating, by the device, a covariance matrix based on the masks;
calculating, by the device, eigenvectors for the covariance matrix; and
generating, by the device, an occupancy grid for the environment based on the eigenvectors for the covariance matrix.
2 . The method of claim 1 , wherein generating the occupancy grid for the environment comprises:
segmenting points associated with the floor from the point cloud based on the eigenvectors; and
clustering remaining points of the point cloud to generate clusters.
3 . The method of claim 2 , wherein generating the occupancy grid for the environment comprises:
constructing convex hulls based on the clusters;
projecting the convex hulls onto the floor to identify edges of the one or more obstacles; and
generating the occupancy grid for the environment based on the edges of the one or more obstacles.
4 . The method of claim 2 , wherein the clusters identify the one or more obstacles in the remaining points of the point cloud.
5 . The method of claim 1 , wherein calculating the eigenvectors for the covariance matrix comprises:
calculating the eigenvectors for the covariance matrix in near real-time.
6 . The method of claim 1 , wherein the occupancy grid enables navigation decisions for the robotic device in the environment.
7 . The method of claim 1 , wherein the occupancy grid identifies the one or more obstacles in relation to a current location of the robotic device.
8 . A device, comprising:
one or more processors configured to:
receive, from a robotic device, time-of-flight sensor data that includes a point cloud of depth images of an environment with a floor and one or more obstacles;
shift the depth images by a number of pixels to generate shifted depth images;
subtract the shifted depth images from the depth images to generate final images;
calculate Euclidean distances associated with the final images;
generate masks for the final images associated with Euclidean distances that are less than a search radius;
calculate a covariance matrix based on the masks;
calculate eigenvectors for the covariance matrix; and
generate an occupancy grid for the environment based on the eigenvectors for the covariance matrix,
wherein the occupancy grid enables navigation decisions for the robotic device in the environment.
9 . The device of claim 8 , wherein the eigenvectors identify edges of the one or more obstacles in the point cloud.
10 . The device of claim 8 , wherein the one or more processors, to calculate the eigenvectors for the covariance matrix, are configured to:
calculate the eigenvectors for the covariance matrix without utilizing a k-dimensional tree graph.
11 . The device of claim 8 , wherein the one or more processors are further configured to:
perform one or more actions based on the occupancy grid.
12 . The device of claim 11 , wherein the one or more processors, to perform the one or more actions, are configured to:
provide the occupancy grid to the robotic device; or
calculate, based on the occupancy grid, a path for the robotic device and instruct the robotic device to utilize the path.
13 . The device of claim 11 , wherein the one or more processors, to perform the one or more actions, are configured to:
identify, from the one or more obstacles and based on the occupancy grid, an obstacle to move and instruct the robotic device to move the obstacle; or
calculate, based on the occupancy grid, movement instructions and provide the movement instructions to the robotic device.
14 . The device of claim 11 , wherein the one or more processors, to perform the one or more actions, are configured to:
identify, from the one or more obstacles, an obstacle based on the occupancy grid; and
provide an identity of the obstacle to the robotic device.
15 . A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors of a device, cause the device to:
receive, from a robotic device, time-of-flight sensor data that includes a point cloud derived from depth images of an environment with a floor and one or more obstacles;
shift the depth images by a number of pixels to generate shifted depth images;
subtract the shifted depth images from the depth images to generate final images;
calculate Euclidean distances associated with the final images;
generate masks for the final images associated with Euclidean distances that are less than a search radius;
calculate a covariance matrix based on the masks;
calculate eigenvectors for the covariance matrix;
generate an occupancy grid for the environment based on the eigenvectors for the covariance matrix; and
perform one or more actions based on the occupancy grid.
16 . The non-transitory computer-readable medium of claim 15 , wherein the one or more instructions, that cause the device to generate the occupancy grid for the environment, cause the device to:
segment points associated with the floor from the point cloud based on the eigenvectors; and
cluster remaining points of the point cloud to generate clusters.
17 . The non-transitory computer-readable medium of claim 16 , wherein the one or more instructions, that cause the device to generate the occupancy grid for the environment, cause the device to:
construct convex hulls based on the clusters;
project the convex hulls onto the floor to identify edges of the one or more obstacles; and
generate the occupancy grid for the environment based on the edges of the one or more obstacles.
18 . The non-transitory computer-readable medium of claim 16 , wherein the clusters identify the one or more obstacles in the remaining points of the point cloud.
19 . The non-transitory computer-readable medium of claim 15 , wherein the occupancy grid identifies the one or more obstacles in relation to a current location of the robotic device.
20 . The non-transitory computer-readable medium of claim 15 , wherein the eigenvectors identify edges of the one or more obstacles in the point cloud.