Measuring the range to an object, in an image, using size categorization

• US 9,058,524 B2
• Filed: 06/28/2013
• Issued: 06/16/2015
• Est. Priority Date: 06/28/2013
• Status: Active Grant

• Current Assignee: Bendix Commercial Vehicle Systems LLC (Estate of Heinz Hermann Thiele)
• Original Assignee: Bendix Commercial Vehicle Systems LLC (Estate of Heinz Hermann Thiele)
• Inventors: Kuehnle, Andreas U.
• Primary Examiner(s): AZARIAN, SEYED H
• Application Number:
  US13/930,910
• Publication Number:
  US 20150003670A1
• Time in Patent Office: 718 Days
• Field of Search: 382/100, 382103-107, 382122-123, 382/154, 382/162, 382/168, 382/173, 382/181, 382/199, 382/206, 382/209, 382/224, 382/232, 382/254, 382/262, 382/274, 382/276, 382286-291, 382/305, 382/312, 375/240.12, 701/2
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 7/73 using feature-based methods
• 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

When determining a range of distance to an object or vehicle in front of or behind the host vehicle, image data of the object is captured and analyzed to determine the relative location of the object relative to the bottom of an image frame containing the object as well as with a feature measurement of the object measured in pixels across a plurality of image frames. The object is classified into one of a plurality of discrete size categories as a function of its relative location and median feature measurement. Once the object is classified, a table lookup is performed to identify the distance range for the object relative to a host vehicle as a function of a monitored median feature measurement of the object and the discrete size category assigned to the object.

Claims

1. A method for estimating distance from a host vehicle to an object, comprising:

capturing image data; and

via a processor;

identifying an object in the image data;

determining a relative location of the object within at least one frame of image data;

determining a median feature measurement, in pixels, of the object;

classifying the object as belonging to at least one of a plurality of size categories as a function of the relative location and median feature measurement of the object;

identifying a distance range to the object based on the classification; and

interpolating a distance from the identified distance range as a function of the median feature measurement of the object.

2. The method according to claim 1, wherein the relative location of the object is a distance to a lower surface of the object measured from a bottom of at least one frame of the image data using a known camera geometry.

3. The method according to claim 1, wherein the identified distance range to the object correlates to a median feature measurement value for the size category to which the object belongs.

4. The method according to claim 1, further comprising monitoring a relative distance to the object as a function of changes in the median feature measurement value of the forward object and the size category to which the object belongs.

5. The method according to claim 1, further comprising iteratively classifying the object.

6. The method according to claim 5, wherein when the iterative classifications of the object result in the object being classified in two adjacent size categories, assigning the forward object to a supercategory that comprises at least the two adjacent categories.

7. The method according to claim 6, further comprising identifying a distance range for the object as a function of the median feature measurement of the object and the supercategory to which the object is assigned.

8. The method according to claim 7, further comprising interpolating a distance from the identified distance range as a function of the median feature measurement of the object.

9. The method according to claim 1, further comprising:

outputting the identified distance range on a graphical display.

10. The method according to claim 1, further comprising:

maintaining the classification of the object as long as the object remains in the image data.

11. The method according to claim 1, further comprising:

identifying the object as a vehicle and performing a table lookup to determine the type of the vehicle.

12. The method according to claim 1, further comprising:

employing a plurality of feature measurements for a plurality of object features when classifying the object.

13. A processor configured to estimate distance from a host vehicle to an object, wherein the processor is configured to:

receive image data;

identify an object in the image data;

determine a relative location of the object within at least one frame of image data;

determine a median feature measurement, in pixels, of the object;

classify the object as belonging to at least one of a plurality of size categories as a function of the relative location and median feature measurement of the forward object;

identify a distance range to the object based on the classification; and

interpolate a distance from the identified distance range as a function of the median feature measurement of the object.

14. The processor according to claim 13, wherein the relative location of the object is a distance to a lower surface of the object measured from a bottom of at least one frame of the image data using a known camera geometry.

15. The processor according to claim 13, wherein the identified distance range to the object correlates to a median feature measurement value for the size category to which the object belongs.

16. The processor according to claim 13, further configured to monitor a relative distance to the object as a function of changes in the median feature measurement value of the object and the size category to which the object belongs.

17. The processor according to claim 13, further configured to iteratively classify the object.

18. The processor according to claim 17, wherein when the iterative classifications of the object result in the object being classified in two adjacent size categories, the processor is further configured to assign the object to a supercategory that comprises at least the two adjacent categories.

19. The processor according to claim 18, further configured to identify a distance range for the object as a function of the median feature measurement of the object and the supercategory to which the object is assigned.

20. The processor according to claim 19, further configured to interpolate a distance from the identified distance range as a function of the median feature measurement of the object.

21. The processor according to claim 13, configured to receive the additional relative location information of the object from a radar unit.

22. The processor according to claim 13, further configured to output the identified distance range on a graphical display.

23. A method for estimating distance from a host vehicle to a forward object, comprising, via a processor:

receiving image data;

identifying an object in the image data;

determining the distance of a lower surface of the object within at least one frame of image data using a known camera geometry;

determining a median feature measurement, in pixels, of the object;

classifying the object as belonging to at least one of a plurality of size categories as a function of the distance and median feature measurement of the object;

identifying a distance range that correlates to a median feature measurement value for the size category to which the object belongs;

interpolating a relative distance within the distance range from the median feature measurement of the object,

monitoring the relative distance to the object as a function of changes in the median feature measurement value of the object and the discrete size category to which the object belongs; and

correlating the relative distance to a following time for the host vehicle behind the object.

24. The method according to claim 23, further comprising:

determining a following time threshold as a function of the relative distance and a relative speed between the host vehicle and the object as determined by at least one of an antilock brake system (ABS) and an adaptive cruise-with-braking (ACB) system; and

providing an alert if the relative distance to the object falls below the following time threshold.