Object tracking device and object tracking method
Provided are an object tracking device and an object tracking method that allow multiple objects to be tracked with high accuracy. An object tracking device ( 20 ) includes an input interface ( 21 ), a processor ( 23 ), and an output interface ( 24 ). The input interface ( 21 ) is configured to acquire sensor data. The processor ( 23 ) is configured to detect multiple detection targets from the sensor data and perform tracking using a Kalman filter for each of the multiple detection targets. The output interface ( 24 ) is configured to output detection results of the detection targets. The processor ( 23 ) allows overlapping of detection results during the process of tracking the multiple detection targets.
1 . An object tracking device comprising:
an input interface configured to acquire sensor data;
a processor configured to detect multiple detection targets from the sensor data and perform tracking using a Kalman filter for each of the multiple detection targets; and
an output interface configured to output detection results of the detection targets,
wherein the processor is further configured to allow overlapping of detection results during a process of tracking the multiple detection targets, without associating multiple observation values with multiple Kalman filters to each other in a one-to-one manner,
wherein when the detected multiple detection targets may be regarded as an identical object, the processor is configured to make the detection target having a smallest estimated range, among estimated ranges based on probability density distributions of positions of the multiple detection targets, be representative of the object.
2 . The object tracking device according to claim 1 , wherein the processor is further configured to perform a control in which the Kalman filter is put into an initial state for the detection target, a tracking preparation state, and a tracking state.
3 . The object tracking device according to claim 2 , wherein the processor is further configured to put the Kalman filter into the tracking state upon successively detecting an identical detection target.
4 . The object tracking device according to claim 1 , wherein the processor is further configured to stop tracking performed by the Kalman filter when an identical detection target is not detected in a prescribed number of successive detection operations.
5 . The object tracking device according to claim 1 , wherein the processor is configured to track positions and velocities of mass points respectively representing the multiple detection targets using a virtual space, the virtual space being a two-dimensional space having a prescribed fixed value in a z-axis direction in a coordinate system consisting of three real space axes that are an x axis, a y axis, and a z axis.
6 . An object tracking method comprising:
acquiring sensor data;
detecting multiple detection targets from the sensor data tracking using a Kalman filter for each of the multiple detection targets; and
outputting detection results of the detection targets,
wherein a processor that executes the tracking allows overlapping of detection results during a process of tracking the multiple detection targets, without associating multiple observation values with multiple Kalman filters to each other in a one-to-one manner,
wherein when the detected multiple detection targets may be regarded as an identical object, the processor is configured to make the detection target having a smallest estimated range, among estimated ranges based on probability density distributions of positions of the multiple detection targets, be representative of the object.
7 . An object tracking device comprising:
an input interface configured to acquire multiple sensor data obtained using different sensing methods;
a processor configured to detect multiple detection targets from the multiple sensor data and perform data processing for tracking by using a Kalman filter for each of the multiple detection targets,
wherein the processor is configured to allow overlapping and linking of detection results of the multiple sensor data to one of the multiple detection targets, without associating multiple observation values with multiple Kalman filters to each other in a one-to-one manner,
wherein when the detected multiple detection targets may be regarded as an identical object, the processor is configured to make the detection target having a smallest estimated range, among estimated ranges based on probability density distributions of positions of the multiple detection targets, be representative of the object.
8 . An object tracking method comprising:
acquiring multiple sensor data obtained using different sensing methods; and
detecting multiple detection targets from the multiple sensor data and performing data processing for tracking by using a Kalman filter for each of the multiple detection targets,
wherein a processor that executes the data processing allows overlapping and linking of the detection results of the multiple sensor data to one of the multiple detection targets, without associating multiple observation values with multiple Kalman filters to each other in a one-to-one manner,
wherein when the detected multiple detection targets may be regarded as an identical object, the processor is configured to make the detection target having a smallest estimated range, among estimated ranges based on probability density distributions of positions of the multiple detection targets, be representative of the object.