METHOD AND SYSTEM FOR TRACKING OBJECTS

• US 20130031045A1
• Filed: 07/25/2011
• Published: 01/31/2013
• Est. Priority Date: 07/25/2011
• Status: Active Grant

• Current Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Toyota Motor Engineering & Manufacturing North America Incorporated (Toyota Motor Corporation)
• Inventors: James, Michael Robert, Samples, Michael Edward
• Granted Patent: US 8,799,201 B2
• Time in Patent Office: Days
• US Class Current: 706/52
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 7/215 Motion-based segmentation
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

An object tracking system and method operable to minimize processing time for tracking objects is provided. The system includes a pair of filters operable to make associations between dynamic objects newly detected and previously detected. One of the pair of filters makes an association when the predicted location of the previously detected objects is within a predetermined distance of the newly detected object. The other of the pair of filters makes an association based upon the probability that a newly detected dynamic object is a previously detected dynamic object. The remaining unassociated dynamic objects are then localized so as to form discrete matrices for optimization filters.

Claims

1. An object detection system operable to minimize processing time for tracking objects, the object detection system comprising:

a memory storage system used for storing detected objects from the previous iteration so as to define a first object group of dynamic objects at the beginning of the predetermined interval;

a sensor operable to detect objects in a predetermined geographic area at a predetermined interval so as to define a second object group detected at the end of the predetermined interval;

a processor in communication with the sensor, the processor operable to determine which of the objects are dynamic and to predict the location for each of the dynamic objects in the first object group at the end of the predetermined interval; and

a pair of filters operable to associate the dynamic objects from the first group with the dynamic objects from the second group so as to track the dynamic objects, the processer processing the predicted location of dynamic objects from the first group and dynamic objects from the second object groups through the pair of filters, wherein one of the pair of filters associates objects from the second object group with a corresponding dynamic object from the first group when the predicted location is within a predetermined distance of each other, and the other of the pair of filters associates each of the dynamic objects from the first object group with one of the corresponding objects in the second object group when the probability that the dynamic object from the second group is the dynamic object from the first group is greater than a predetermined probability.

2. The system as set forth in claim 1, further including an optimization filter, the processor further executing the optimization filter so as to associate the unassociated dynamic objects from the first group with the unassociated dynamic objects from the second group.

3. The system as set forth in claim 1, further including a processing filter operable to sort the unassociated dynamic objects from second object groups into a plurality of discrete categories, and wherein the processor uses association to associate the unassociated objects from the second group with dynamic objects from the first group.

4. The system as set forth in claim 3, wherein the processor actuates the optimization filter so as to process each of plurality of discrete categories independently so as associate the unassociated objects from the second group with dynamic objects from the first group.

5. The system as set forth in claim 1, further wherein one of the plurality of discrete categories includes unassociated objects located within a predetermined distance of the predicted location of dynamic objects from the first object group.

6. The system as set forth in claim 1, further including a segmentation module, the segmentation model associating data from objects detected in the first object group with objects detected in the second object group when the predicted location of objects in the first object groups are within a predetermined distance of detected objects in the second object group.

7. The system as set forth in claim 1, further including a counting module, the counting module operable to discard an unassociated object from the first object group after a predetermined number of predetermined intervals.

8. The system as set forth in claim 7, wherein the predetermined number of predetermined intervals is related to the number of times the dynamic object was associated.

9. The system as set forth in claim 1, further including a mapping module operable to associate each object with a grid coordinate.

10. The system as set forth in claim 1, further including an association matrix, the association matrix having an artificial dynamic object for each of the unassociated dynamic objects from the first and second object groups.

11. Method of tracking objects comprising the steps of:

storing data from previous iterations so as to define a first object group;

gathering data related to objects at predetermined intervals of time so as to detect a second object group;

processing the data so as to classify the object as being either dynamic or static;

predicting the location of dynamic objects in the first object group; and

using a association to associate objects from the first object group with objects from the second object group, wherein the remaining unassociated objects are associated with each other using an optimization filter.

12. The method as set forth in claim 11, wherein an object from the second object group is associated with an object from the first object group when the object from the second object group is within a predetermined distance of the predicted location of an object from the first object group.

13. The method as set forth in claim 11, wherein an object from the second object group is associated with an object from the first object group when the probability that the object from the second object group is an object from the first object group, is above a predetermined probability.

14. The method as set forth in claim 11, further including the step of localizing the remaining unassociated objects so as to form discrete matrices, and processing the discrete matrices using an optimization filter.

15. The method as set forth in claim 14, wherein localization is performed using association.

16. The method as set forth in claim 15, wherein the association is based upon distance.

17. The method as set forth in claim 15, wherein the association is based upon probability.