Tracking dynamics using a computerized device
The presently disclosed subject matter there includes a computerized device that comprises a processing circuitry, and one or more sensors including at least one image sensor configured to continuously capture images of an environment; the processing circuitry is configured to process the captured images and: detect an object within the environment; track a trajectory of movement of the object; select a best fitting motion model to the trajectory and determine a state of the object based on the best fitting motion model; and determine at least one intersection point between the object and a target object.
1. A computerized device comprising a processing circuitry, the processing circuitry comprising at least one computer processor; the computerized device further comprises or is otherwise operatively connected to one or more sensors including at least one camera configured to continuously capture images of an environment;
the processing circuitry is configured to process the captured images and:
detect an object moving within the environment;
determine a plurality of location-points, each location-point is determined based on one or more detected appearances of the object in the captured images;
determine object parameters;
determine a trajectory of movement of the object, based on the plurality of location-points and parameters;
select a best fitting motion model to the trajectory from a plurality of available motion models, and determine a state of the object based on the best fitting motion model;
determine at least one intersection point between the object and a target object,
generate an activity world as a virtual environment that corresponds to the environment,
define an activity area having specific boundaries within the activity world, the activity area being assigned with one or more attributes defining one or more constraints on a manner in which an activity of a user should be performed,
detect based on the at least one intersection point an interaction of the user with the object,
following detection of interaction of the user with the object:
determining an area of interaction within the activity world;
determining whether the area of interaction overlaps with the activity area, and if so:
determining whether one or more interaction attributes comply with the one or more attributes assigned to the activity area; and
determining a level of success in performing the activity of the user.
2. The computerized device of claim 1 , wherein the processing circuitry is further configured to:
continuously update the trajectory of the moving object, using additional location-points, to thereby generate an updated trajectory;
compare the updated trajectory with a plurality of motion models to determine a best fitting motion model to the updated trajectory;
identify a change in the state of the object, in case it is determined that a motion model that is different from the selected motion model has a best fit to the updated trajectory.
3. The computerized device of claim 1 , wherein the processing circuitry is configured for determining the trajectory of the object based on the plurality of location-points and parameters, to:
define a sliding window encompassing a sub-group of the plurality of location-points;
shift the position of the sliding window along the plurality of location points; for each window position:
determine, based on a respective sub-group of location-points in the window, a selected motion model having a best fit to a sub-section of the trajectory consisting of the sub-group of location-points;
assign a current state to the object according to the selected motion model determined for the plurality of windows.
4. The computerized device of claim 1 , wherein the processing circuitry is configured to reapply the selected motion model on the trajectory.
5. The computerized device of claim 1 , wherein the processing circuitry is configured for determining the at least one intersection point between the object and the target object, to:
estimate a plurality of future trajectories of the moving object at time of intersection with the target object, thereby obtaining a plurality of candidate intersection points;
calculate a distribution of the plurality of candidate intersection points; and determine a final intersection point if a distance of the object from the target object is less than a predefined distance and a variance of the distribution is less than a predefined value.
6. The computerized device of claim 1 , wherein the processing circuitry is configured, following determination of the at least one intersection point between the object and the target object, to:
determine whether the at least one intersection point, complies with the one or more attributes and determine a level of success in performing the interaction accordingly.
7. The computerized device of claim 1 , wherein processing of the images is performed in real-time while the images are being captured.
8. The computerized device of claim 1 , wherein the processing circuitry is configured to adapt a camera frame rate according to the state of the object.
9. A method of monitoring using activity in an environment using a computerized device, comprising a processing circuitry that includes or is otherwise operatively connected to at least one image sensing device; the method comprising:
capturing images of the environment using the at least one image sensing device;
processing the captured images, comprising:
detecting an object moving within the environment;
determining a plurality of location-points, each location-point is determined based on one or more detected appearances of the object in the captured images;
determining object parameters;
determining a trajectory of movement of the object based on the plurality of location-points and parameters;
selecting a best fitting motion model to the trajectory from a plurality of available motion models, and determining a state of the object based on the best fitting motion model;
determining at least one intersection point between the object and a target object,
generating an activity world as a virtual environment that corresponds to the environment,
defining an activity area having specific boundaries within the activity world, the activity area being assigned with one or more attributes defining one or more constraints on a manner in which an activity of a user should be performed,
detecting based on the at least one intersection point an interaction of the user with the object,
following detection of interaction of the user with the object:
determining an area of interaction within the activity world;
determining whether the area of interaction overlaps with the activity area, and if so:
determining whether one or more interaction attributes comply with the one or more attributes assigned to the activity area; and
determining a level of success in performing the activity of the user.
10. The method of claim 9 , further comprising:
continuously updating the trajectory of the object using additional location-points, to thereby generate an updated trajectory;
comparing the updated trajectory with a plurality of motion models to determine a best fitting motion model to the updated trajectory;
identifying a change in the state of the object, in case it is determined that a motion model that is different from the selected motion model has a best fit to the updated trajectory.
11. The method of claim 9 further comprising, for determining the trajectory of the moving object based on the plurality of location-points and parameters:
defining a sliding window encompassing a sub-group of the plurality of location-points;
shifting the position of the sliding window along the plurality of location points;
for each window position:
determining, based on a respective sub-group of location-points in the window, a selected motion model having a best fit to a sub-section of the trajectory consisting of the sub-group of location-points; and
assigning a current state to the moving object according to the selected motion model determined for the plurality of windows.
12. The method of claim 9 further comprising, for determining the at least one intersection point between the object and the target object:
estimating a plurality of future trajectories of the moving object at a time of intersection with the target object, thereby obtaining a plurality of candidate intersection points;
calculating a distribution of the plurality of candidate intersection points; and
determining a final intersection point if a distance of the object from the target object is less than a predefined distance and a variance of the distribution is less than a predefined value.
13. The method of claim 9 further comprising following determination of the at least one intersection point between the object and the target object:
determining whether the at least one intersection point, complies with the one or more attributes, and determining a level of success in performing the interaction accordingly.
14. The method of claim 9 wherein processing of the images is performed in real-time while the images are being captured.
15. The method of claim 9 further comprising adapting a camera frame rate according to the state of the object.
16. A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by at least one computer, cause the at least one computer to perform:
obtaining captured images of an environment;
processing the captured images, comprising:
detecting an object moving within the environment;
determining a plurality of location-points, each location-point is determined based on one or more detected appearances of the object in the captured images;
determining object parameters;
determining a trajectory of movement of the object based on the plurality of location-points and parameters;
selecting a best fitting motion model to the trajectory from a plurality of available motion models, and determining a state of the object based on the best fitting motion model;
determining at least one intersection point between the object and a target object,
generating an activity world as a virtual environment that corresponds to the environment,
defining an activity area having specific boundaries within the activity world, the activity area being assigned with one or more attributes defining one or more constraints on a manner in which an activity of a user should be performed,
detecting based on the at least one intersection point an interaction of the user with the object,
following detection of interaction of the user with the object:
determining an area of interaction within the activity world;
determining whether the area of interaction overlaps with the activity area, and if so:
determining whether one or more interaction attributes comply with the one or more attributes assigned to the activity area; and
determining a level of success in performing the activity of the user.
17. A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by at least one computer, cause the at least one computer to perform:
obtaining captured images of an environment;
processing the captured images, comprising:
detecting an object within the environment;
determining a plurality of location-points, each location-point is determined based on one or more detected appearances of the object in the captured images;
determining object parameters;
determining a trajectory of movement of the object based on the plurality of location-points and parameters;
selecting a best fitting motion model to the trajectory from a plurality of available motion models, and determining a state of the object based on the best fitting motion model; and
determining at least one intersection point between the object and a target object, said determining comprising:
estimating a plurality of future trajectories of the moving object at a time of intersection with the target object, thereby obtaining a plurality of candidate intersection points;
calculating a distribution of the plurality of candidate intersection points; and
determining a final intersection point if a distance of the object from the target object is less than a predefined distance and a variance of the distribution is less than a predefined value.
18. A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by at least one computer, cause the at least one computer to perform:
obtaining captured images of an environment;
processing the captured images, comprising:
detecting an object within the environment;
determining a plurality of location-points, each location-point is determined based on one or more detected appearances of the object in the captured images;
determining object parameters;
determining a trajectory of movement of the object based on the plurality of location-points and parameters;
defining a sliding window encompassing a sub-group of the plurality of location-points;
shifting the position of the sliding window along the plurality of location-points;
for each window position:
determining, based on a respective sub-group of location-points in the window, a selected motion model having a best fit to a sub-section of the trajectory consisting of the sub-group of location-points;
assigning a current state to the object according to the selected motion model determined for the plurality of windows,
and
determining at least one intersection point between the object and a target object.