Method for detecting moving objects in a blind spot region of a vehicle and blind spot detection device

• US 8,040,227 B2
• Filed: 05/02/2008
• Issued: 10/18/2011
• Est. Priority Date: 05/03/2007
• Status: Active Grant

• Current Assignee: Sony Deutschland GmbH (Sony Group Corp.)
• Original Assignee: Sony Deutschland GmbH (Sony Group Corp.)
• Inventors: Zimmermann, Klaus, Siddiqui, Muhammad, Friedrichs, Fabian
• Primary Examiner(s): Wu; Daniel
• Assistant Examiner(s): TUN, NAY L
• Application Number:
  US12/114,239
• Publication Number:
  US 20080309516A1
• Time in Patent Office: 1,264 Days
• Field of Search: 340903-904, 340/435, 340/436, 340/937, 700300-301, 382/104, 382/107, 382/173
• US Class Current: 340/436
• CPC Class Codes: G06T 7/223 using block-matching
• 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

A method for detecting moving objects in a blind spot of vehicle is provided, comprising: taking a sequence of images of said blind spot region, partitioning each of said images of said sequence into blocks, identifying moving blocks which have moved between consecutive images of said sequence, determining a direction and a distance of said movement of said moving blocks; grouping adjacent moving blocks, for which directions within a predetermined direction interval and distances within a predetermined distance interval have been determined, and determining said moving objects based on said step of grouping. A corresponding blind spot detection device is provided as well.

Claims

1. A method for detecting moving objects in a blind spot region of a vehicle, comprising:

taking a sequence of images of said blind spot region;

partitioning each of said images of said sequence into blocks;

identifying moving blocks which have performed a movement between consecutive images of said sequence;

determining a direction and a distance of said movement of said moving blocks;

splitting said images of said sequence into at least two regions;

segmenting said images by grouping adjacent moving blocks, for which directions within a predetermined direction interval and distances within a predetermined distance interval have been determined, wherein different criteria are used for said segmenting in said at least two regions; and

determining said moving objects based on said step of segmenting.

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

determining at least one false alarm among said moving objects; and

generating an alarm if at least one of said moving objects is not a false alarm.

3. The method according to claim 2, further comprising:

determining an object size of said moving object;

comparing said object size with a predetermined size; and

generating said false alarm if said object size is smaller than said predetermined size.

4. The method according to claim 3, further comprising:

determining a vanishing point within said images, wherein said predetermined size depends on a distance of said moving object from said vanishing point.

5. The method according to claim 2, further comprising:

identifying a vanishing point within said images;

determining an object size of said moving object;

determining a vertical distance of said moving object from said vanishing point; and

generating said false alarm, if said object size is below a predetermined object size and if said vertical distance is above a predetermined distance.

6. The method according to claim 2, further comprising:

evaluating a trajectory of said moving object; and

generating said false alarm based on said step of evaluating.

7. The method according to claim 2, further comprising:

determining a size of said moving object;

calculating a difference between a region surface and a convex hull if said size is greater than a predetermined size; and

redefining said moving objects, if said difference is above a predetermined difference.

8. The method according to claim 1, wherein a different search region and/or a different matchblock size are used as said different criteria.

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

identifying moving objects in said at least two regions using different training data.

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

upscaling or downscaling images differently in said at least two regions.

11. The method according to claim 2, further comprising:

identifying false alarms within different regions using different thresholds.

12. The method according to claim 1, wherein in said step of splitting,

said images of said sequence is split into two regions.

13. A computer program product including a non-transitory computer readible medium encoded thereon with a computer program code, which when executed by a processor causes the processor to perform the method according to claim 1.

14. The method according to any of claims 2 to 5, further comprising:

calculating a histogram and/or edge based statistical measure of at least one of said moving objects;

evaluating said histogram and/or edge based statistical measure; and

generating said false alarm based on said step of evaluating.

15. The method according to claim 14, wherein said statistical measure is a variance, an inter quartier range, a mean, a median, an energy, an edge energy, a uniformity, a smoothness, a third moment and/or an entropy.

16. A blind spot detection device for a vehicle, comprising:

a camera configured to take a sequence of images of said blind spot region;

a splitting region unit configured to split said images of said sequence into at least two regions;

a processor configured

to partition each of said images of said sequence into blocks,

to identify moving blocks which have performed a movement between consecutive images of said sequence,

to determine a direction and a distance of said movement of said moving blocks,

to segment said images by grouping adjacent moving blocks, for which directions within a predetermined direction interval and distances within a predetermined distance interval have been determined, wherein different criteria are used for said segmenting in said at least two regions, and

to determine moving objects; and

an indicator configured to indicate a dangerous situation, if a moving object in said blind spot has been detected.

17. The blind spot detection device according to claim 16, further comprising:

a false alarm rejection unit, connected to said processor and to said indicator, and being configured to identify false alarms among said moving objects, wherein

said indicator is further configured to generate an alarm if at least one of said moving objects is not a false alarm.

18. The blind spot detection device according to claim 16 or 17, further comprising:

an active sensor configured to determine a distance to said moving object.