Method and apparatus for identifying road signs

• US 9,697,430 B2
• Filed: 09/04/2014
• Issued: 07/04/2017
• Est. Priority Date: 10/01/2013
• Status: Active Grant

• Current Assignee: Conti Temic Microelectronic GmbH (Continental AG)
• Original Assignee: Conti Temic Microelectronic GmbH (Continental AG)
• Inventors: Kristensen, Steen
• Primary Examiner(s): Tabatabai, Abolfazl
• Application Number:
  US14/902,246
• Publication Number:
  US 20170017849A1
• Time in Patent Office: 1,034 Days
• Field of Search: 382103, 382104, 340907, 340910, 340917, 340988, 701117, 701118, 701469, 701514
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/30 Noise filtering
• CPC Class Codes: G06V 20/582 of traffic signs
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names

Abstract

The invention relates to a method and a device for recognizing a road sign (Hz, Zz) by means of a camera from a traveling vehicle. The method has the following steps:

• • acquiring at least one image of the surroundings of a vehicle by means of the camera,
  • determining the presence of at least one road sign (Hz, Zz) in the at least one acquired image of the surroundings of the vehicle,
  • determining (also estimating) motion blur in that image segment in which the present road sign (Hz, Zz) is situated,
  • subtracting out motion blur in this image segment, which results in a sharpened image segment, and
  • recognizing the road sign (Hz, Zz) while taking account of the sharpened image segment.

Claims

1. A method of recognizing a road sign (Hz, Zz) by a camera from a traveling vehicle, comprising steps:

acquiring at least one image of surroundings of the vehicle by the camera,

determining that at least one road sign (Hz, Zz) is present in the at least one image of the surroundings of the vehicle,

determining a motion blur in the at least one image in an image segment thereof in which the road sign (Hz, Zz) is situated,

subtracting out at least some of the motion blur in the image segment, which results in a sharpened image segment, and

recognizing the road sign (Hz, Zz) while taking account of the sharpened image segment.

2. The method according to claim 1, wherein the recognizing of the road sign (Hz, Zz) while taking account of the sharpened image segment comprises performing text recognition in the sharpened image segment.

3. The method according to claim 1, wherein the determining of the motion blur comprises analyzing data characterizing a motion of the vehicle.

4. The method according to claim 1, wherein the determining of the motion blur comprises determining and analyzing edges in the image segment in which the road sign (Hz, Zz) is situated.

5. The method according to claim 1, further comprising recognizing a static content (S) of the road sign (Hz, Zz) and partially recognizing the road sign from the acquired image in consideration of the static content, and wherein the determining of the motion blur in the image segment comprises taking into consideration a predetermined road sign template corresponding to the partially recognized road sign, and further comprising recognizing a variable content (V) of the road sign (Hz, Zz) while taking account of the sharpened image segment.

6. The method according to claim 5, wherein the static content (S) comprises a shape and an edge of the road sign which is a main road sign (Hz).

7. The method according to claim 5, wherein the image segment is formed such that the at least one road sign situated in the image segment includes at least one main road sign (Hz) and all of one or more additional signs (Zz) that are associated with the at least one main road sign, wherein an entirety of at least one sign among the at least one main road sign (Hz) is recognized as the static content and a content of the one or more additional signs (Zz) is recognized as the variable content.

8. The method according to claim 1, wherein the at least one road sign includes plural road signs, further comprising forming a separate said image segment respectively for each individual road sign (Hz or Zz) among the plural road signs included in the image, subsequently performing a check as to whether at least two road signs (Hz, Zz) among the plural road signs are assigned to or associated with each other, and, when the check is affirmative, then the motion blur is determined only for a first said image segment and is then adopted for at least a further one of said image segments assigned thereto.

9. The method according to claim 1, wherein, if the recognition of the road sign (Hz, Zz) while taking account of the sharpened image segment is not successful or not reliable, then further comprising creating a synthetic highly sharp image segment is created from a plurality of sharpened image segments extracted from successive images acquired by the camera, each of said sharpened image segments including the road sign (Hz, Zz), wherein the road sign (Hz, Zz) is recognized while taking account of the synthetic highly sharp image segment.

10. A device for road sign recognition comprising:

a camera configured and arranged on a traveling vehicle to acquire image data including at least one image of surroundings of the traveling vehicle, and

an image analysis unit configured and arranged to detect a presence of a road sign (Hz, Zz) from the image data, to determine a motion blur in an image segment in which the road sign (Hz, Zz) is situated, to subtract at least some of the motion blur out of the image segment, which results in a sharpened image segment, to recognize the road sign (Hz, Zz) while taking account of the sharpened image segment, and to generate a corresponding output signal.