Back to results
US 8,527,235 B2Granted
Object position estimation system, object position estimation device, object position estimation method and object position estimation program
AssigneePanasonic Corporation
InventorToru Tanigawa
Life of this patent
Estimating the term…
Abstract
ID likelihoods and position likelihoods of an object are detected by a first observation device, and position likelihoods of the object and tracking states of the object are detected by a second observation device; thus, the object detected by the second observation device and the object ID are associated with each other by an association unit, and based upon information from the second observation device and the association unit, the ID likelihoods of the object detected by the second observation device are determined by a second object ID likelihood determination unit so that the object position is estimated by an object position estimation unit based upon the ID likelihoods and position likelihoods.
Claims8
An object position estimation system, which estimates a position of an object, comprising:
- a first observation unit that acquires pieces of first observation information including positions and identifications of the object at respectively different points of time;
- a first object position likelihood determination unit that, based upon the first observation information, determines first object position likelihoods corresponding to estimated positions of the object at the respectively different points of time;
- a first object identification likelihood determination unit that, based upon the identifications, determines first object identification likelihoods of the object at the respective points of time;
- a second observation unit that attaches second observation identifications to pieces of second observation information including the positions and amounts of feature of the object acquired at respectively different points of time;
- a second object position likelihood determination unit that, based upon the second observation information, determines second object position likelihoods corresponding to estimated positions of the object at the respective points of time;
- an object tracking state determination unit that determines, by making the second observation identifications of two pieces of the second observation information having a same amount of feature of the object, observed at different points of time, mutually associated with each other, tracking state information of the object;
- a second object identification likelihood determination unit that, based upon the tracking state information of the object and the estimated positions of the object, determines second object identification likelihoods of the second observation information;
- an association unit that, based upon the first object identification likelihoods and the first object position likelihoods of the object, calculates a first object association value, and based upon the second object identification likelihoods and the second object position likelihoods, calculates a second object association value;
- and
- an object position estimation unit that, based upon at least one of two groups, (1) the first object identification likelihoods, the first object position likelihoods, and the first object association value of the object, and (2) the second object identification likelihoods, the second object position likelihoods, and the second object association value of the object, estimates the position of the object.
7 more elements
An object position estimation method, which estimates a position of an object, comprising:
- acquiring pieces of first observation information including positions and identifications of the object at respectively different points of time by a first observation unit;
- based upon the first observation information, determining first object position likelihoods corresponding to estimated positions of the object at the respectively different points of time by a first object position likelihood determination unit;
- based upon the identifications, determining first object identification likelihoods of the object at the respective points of time by a first object identification likelihood determination unit;
- attaching second observation identifications to pieces of second observation information including the positions and amounts of feature of the object acquired at respectively different points of time, by a second observation unit;
- based upon the second observation information, determining second object position likelihoods corresponding to estimated positions of the object at the respective points of time by a second object position likelihood determination unit;
- by making the second observation identifications of two pieces of the second observation information having a same amount of feature of the object, observed at different points of time, mutually associated with each other, determining tracking state information of the object by an object tracking state determination unit;
- based upon the tracking state information of the object and the estimated positions of the object, determining second object identification likelihoods of the second observation information by a second object identification likelihood determination unit;
- based upon the first object identification likelihoods and the first object position likelihoods of the object, calculating a first object association value by an association unit;
- based upon the second object identification likelihoods and the second object position likelihoods of the object, calculating a second object association value by the association unit;
- and
- based upon at least one of two groups, (1) the first object identification likelihoods, the first object position likelihoods, and the first object association value of the object, and (2) the second object identification likelihoods, the second object position likelihoods, and the second object association value of the object, estimating the position of the object by an object position estimation unit.
8 more elements
A non-transitory computer readable medium containing computer instructions stored therein for causing a computer processor to perform:
- acquiring first observation information including positions and identifications of an object at respectively different points of time by a first observation unit;
- based upon the first observation information, determining first object position likelihoods corresponding to estimated positions of the object at the respectively different points of time by a first object position likelihood determination unit;
- based upon the identifications, determining first object identification likelihoods of the object at the respective points of time by a first object identification likelihood determination unit;
- attaching second observation identifications to pieces of second observation information including the positions and amounts of feature of the object acquired at respectively different points of time, by the second observation unit;
- based upon the second observation information, determining second object position likelihoods corresponding to estimated positions of the object at the respective points of time by a second object position likelihood determination unit;
- by making second observation identifications of two pieces of the second observation information having a same amount of feature of the object, observed at different points of time, mutually associated with each other, determining tracking state information of the object by an object tracking state determination unit;
- based upon the tracking state information of the object and the estimated positions of the object, determining second object identification likelihoods of the second observation information by a second object identification likelihood determination unit;
- based upon the first object identification likelihoods and the first object position likelihoods of the object, calculating a first object association value by an association unit;
- based upon the second object identification likelihoods and the second object position likelihoods, calculating a second object association value by the association unit;
- and
- based upon at least one of two groups, (1) the first object identification likelihoods, the first object position likelihoods, and the first object association value of the object, and (2) the second object identification likelihoods, the second object position likelihoods, and the second object association value of the object, estimating the position of the object by using an object position estimating unit.
8 more elements
An object position estimation device comprising:
- a first object position likelihood determination unit that, based upon first observation information including positions and identifications of an object at respectively different points of time, from a first observation unit that acquires the first observation information, determines first object position likelihoods corresponding to estimated positions of the object at the respectively different points of time;
- a first object identification likelihood determination unit that, based upon the identifications, determines first object identification likelihoods of the object at the respective points of time;
- a second object position likelihood determination unit that, based upon second observation information including positions and amounts of feature of an object acquired at respectively different points of time, from a second observation unit that attaches second observation identifications to piece of the second observation information, determines second object position likelihoods corresponding to estimated positions of the object at the respectively different points of time;
- an object tracking state determination unit that by making the second observation identifications of two pieces of the second observation information having a same amount of feature of the object, observed at different points of time, mutually associated with each other, determines tracking state information of the object;
- a second object identification likelihood determination unit that, based upon the tracking state information of the object and the estimated positions of the object, determines second object identification likelihoods of the second observation information;
- an association unit that, based upon the first object identification likelihoods and the first object position likelihoods of the object, calculates a first object association value, and based upon the second object identification likelihoods and the second object position likelihoods, calculates a second object association value;
- and
- an object position estimation unit that, based upon at least either one of the groups including (1) the first object identification likelihoods, the first object position likelihoods, and the first object association value of the object, and (2) the second object identification likelihoods, the second object position likelihoods, and the second object association value of the object, estimates the position of the object.
5 more elements
Dependencies are inferred from the claim text.