Simple classification scheme for vehicle/pole/pedestrian detection

• US 6,687,577 B2
• Filed: 09/13/2002
• Issued: 02/03/2004
• Est. Priority Date: 12/19/2001
• Status: Active Grant

• Current Assignee: Ford Global Technologies, LLC (Ford Motor Company)
• Original Assignee: Ford Global Technologies, LLC (Ford Motor Company)
• Inventors: Strumolo, Gary Steven
• Primary Examiner(s): Marc-Coleman, Marthe Y.
• Application Number:
  US10/065,067
• Publication Number:
  US 20030114964A1
• Time in Patent Office: 508 Days
• Field of Search: 701/1, 701/301, 348/116, 348/118, 348/148, 348/143, 342/70, 342/118, 342/175
• US Class Current: 701/1
• CPC Class Codes: B60R 21/0134 responsive to imminent contact with an obstacle , e.g. using radar systems
• CPC Class Codes: G08G 1/165 for passive traffic, e.g. including static obstacles, trees
• CPC Class Codes: G08G 1/166 for active traffic, e.g. moving vehicles, pedestrians, bikes

Abstract

A classifying system 10 for a host vehicle 12 includes a sensor 14 and a controller 24 coupled to the host vehicle 12. The sensor 14 detects boundary data of an obstacle 16, 18, 20 and 22. The sensor 14 also generates an obstacle signal, which is received by the controller 24, which then generates a bounding box 26, 28, 30 and 32 for the obstacle 16, 18, 20 and 22. The bounding box 26, 28, 30 and 32 includes a number of vertical pixels 40, corresponding to a maximum height of the obstacle 16, 18, 20 and 22, and a number of horizontal pixels 42, corresponding to a maximum width of the obstacle 16, 18, 20 and 22. The controller 24 includes obstacle classifying logic activating in response to height and width of the obstacle 16, 18, 20 and 22 within parameters of one of a plurality of objects.

Claims

1. A classifying system for a host vehicle comprising:

at least one sensor coupled to the host vehicle and detecting boundary data of an obstacle, said sensor further generating an obstacle signal in response to said boundary data; and

a controller coupled to the host vehicle and receiving said obstacle signal, said controller generating a bounding box for said obstacle in response to said obstacle signal, said bounding box including at least one vertical pixel corresponding to a maximum height of said obstacle and at least one horizontal pixel corresponding to a maximum width of said obstacle, said controller including obstacle classifying logic classifying said obstacle in response to said height and said width and determining a rotation angle of approach of said obstacle.

2. The system of claim 1 further comprising an array of sensors detecting a position and a velocity of said obstacle.

3. The system of claim 1 wherein said obstacle comprises a target vehicle.

4. The system of claim 3 wherein said at least one sensor detects at least one inside corner of said target vehicle, and wherein said controller compares a distance between said inside corner of said target vehicle and a nearest side of said bounding box to a distance between an inside corner of said bounding box and an axis extending from an outer edge of said bounding box to a direction of travel of the host vehicle.

5. The system of claim 3 wherein said obstacle height comprises a body height of said target vehicle.

6. The system of claim 3 wherein said obstacle height comprises a total height of said target vehicle.

7. The system of claim 1 wherein said parameters include at least two different obstacle heights corresponding to different vehicle types.

8. An obstacle classifying method for a host vehicle comprising:

detecting an obstacle through at least one sensor coupled to the host vehicle;

generating an obstacle signal;

receiving said obstacle signal in a controller coupled to the host vehicle;

generating a bounding box around an image of said obstacle in response to said obstacle signal, said bounding box including a number of vertical pixels corresponding to a maximum height of said obstacle and a number of horizontal pixels corresponding to a maximum width of said obstacle;

activating vehicle classifying logic in response to said height and said width of said obstacle within vehicle parameters;

classifying a type of a target vehicle based on said obstacle height;

estimating a rotation angle of said target vehicle in relation to the host vehicle;

activating said vehicle classifying logic in response to said height and said width of said obstacle within other object parameters; and

classifying said obstacle based on at least one of said obstacle height and said obstacle width.

9. The method of claim 8 further comprising:

estimating all rotation angles of said target vehicle; and

eliminating one of an angled away or an angled toward image of said target vehicle trough a measurement of closing velocity.

10. The method of claim 8 further comprising activating an actuator in response to either said step of classifying a type of a target vehicle or said step of classifying said obstacle.

11. An obstacle classifying system for a host vehicle comprising:

at least one sensor coupled to the host vehicle and detecting boundary data of an obstacle and a velocity of said obstacle and a distance from said obstacle to the host vehicle, said sensor further generating an obstacle signal in response to said boundary data; and

a controller coupled to the host vehicle and receiving said obstacle signal, said controller generating a bounding box for said obstacle in response to said obstacle signal, said bounding box including a number of vertical pixels corresponding to a maximum height of said obstacle and a number of horizontal pixels corresponding to a maximum width of said obstacle, said controller including vehicle classifying logic and other object classifying logic, said vehicle classifying logic activating in response to said height and said width of said obstacle within vehicle parameters, said other object classifying logic activating in response to said height and said width of said obstacle within other object parameters, said vehicle classifying logic classifying a type of a target vehicle based on said obstacle height, said vehicle classifying logic further estimating a rotation angle of said target vehicle in relation to the host vehicle, said vehicle classifying logic further determining a direction of travel of said target vehicle based on said obstacle velocity and said rotation angle, said other object classifying logic classifying said obstacle based on said obstacle height and said obstacle width.

12. The system of claim 11 further comprising an array of sensors detecting a position and a velocity of said obstacle.

13. The system of claim 11 wherein said at least one sensor detects at least one inside corner of said target vehicle, and wherein said controller compares a distance between said inside corner of said target vehicle and a nearest side of said bounding box to a distance between an inside corner of said bounding box and an axis extending from an outer edge of said bounding box to a direction of travel of the host vehicle.

14. The system of claim 11 wherein said obstacle height comprises a body height of said target vehicle.

15. The system of claim 11 wherein said obstacle height comprises a total height of said target vehicle.

16. The system of claim 11 wherein said vehicle parameters include at least two different obstacle heights corresponding to different vehicle types.

17. The system of claim 11 wherein said vehicle parameters include said width of said obstacle set at approximately 2 meters.

18. The system of claim 11 wherein said other object parameters include said width of said obstacle less than 1 meter and said height of said obstacle less than 2.5 meters for a pedestrian.

19. The system of claim 11 wherein said other object parameters include said width of said obstacle less than 2 meters and said height of said obstacle greater than a visual field of said at least one sensor for a telephone pole or a tree.

20. The system of claim 11 wherein said other object parameters include said width of said obstacle greater than 2 meters and said height of said obstacle greater than a visual field of said at least one sensor for a wall.

21. The system of claim 11 wherein said classifying logic and other object classifying logic include Hough transform logic.