Simultaneous location and mapping (SLAM) using dual event cameras
AssigneeSamsung Electronics Co., Ltd.
InventorsSebastien Derhy, Lior Zamir, Nathan Henri Levy
Life of this patent
Estimating the term…
Abstract
A method for simultaneous localization and mapping (SLAM) employs dual event-based cameras. Event streams from the cameras are processed by an image processing system to stereoscopically detect surface points in an environment, dynamically compute pose of a camera as it moves, and concurrently update a map of the environment. A gradient descent based optimization may be utilized to update the pose for each event or for each small batch of events.
Claims20
A method for simultaneous localization and mapping (SLAM), comprising:
- receiving, from first and second image sensors, first and second event streams, respectively, each comprising a plurality of event packets indicating asynchronous events corresponding to points of surfaces in an environment, wherein the first and second image sensors are arranged with overlapping fields of view, wherein a number of asynchronous events is controlled in event frame units and the event frame is formed by the movement of the first and second image sensors;
- computing depths of the points represented by the first event stream based on relative pixel locations of common points represented by the second event stream;
- and
- dynamically computing a pose of at least the first image sensor with respect to a reference element in the environment, and updating a map of the environment, based at least on the points represented by the first event stream and the computed depths thereof.
A simultaneous localization and mapping (SLAM) system comprising:
- a first image sensor that provides a first event stream of event packets indicating asynchronous events corresponding to points of surfaces in an environment;
- a second image sensor arranged to have an overlapping field of view with that of the first image sensor and providing a second event stream of event packets indicating asynchronous events corresponding to points of surfaces in the environment;
- an image processing system comprising at least one processor that executes instructions read from a memory to:
- compute depths from the first and second image sensors stereoscopically based on points of matching features represented by the first and second event streams;
- and
- dynamically compute a pose of at least the first image sensor with respect to a reference element in the environment, and update a map of the environment, based at least on the points of matching features represented by the first event stream and the computed depths thereof,
- wherein a number of asynchronous events is controlled in event frame units and the event frame is formed by the movement of the first and second image sensors.
Dependencies are inferred from the claim text.