Device for monitoring surroundings of a vehicle

• US 9,160,986 B2
• Filed: 11/09/2011
• Issued: 10/13/2015
• Est. Priority Date: 11/16/2010
• Status: Active Grant

• Current Assignee: Arriver Software AB (Qualcomm, Inc.)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Nagaoka, Nobuharu, Aimura, Makoto
• Primary Examiner(s): KOSTAK, VICTOR R
• Application Number:
  US13/884,732
• Publication Number:
  US 20130229525A1
• Time in Patent Office: 1,434 Days
• US Class Current: 1/1
• 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
• CPC Class Codes: G08G 1/163 involving continuous checking
• CPC Class Codes: H04N 7/183 for receiving images from a single remote source

Abstract

A device for monitoring surroundings of a vehicle mounted in a vehicle, captures an image of surroundings of the vehicle, extracts from the captured image an image area for a predetermined part of a desired type of object, and sets a predetermined area below the extracted image area. The device extracts candidates for the desired type of object present outside the vehicle based on the predetermined area and determines, for each of the extracted object candidates, whether the object candidate is the desired type of object. When no pattern different from the background is captured in the predetermined area, or when a pattern different from the background is captured only in one of a first and a second areas, then the object in the image area is excluded from the candidates for the desired type of object. Thus, a pedestrian, for example, can be distinguished from an artificial structure.

Claims

1. A vehicle surroundings monitoring device comprising:

an imaging unit which is mounted in a vehicle and captures an image of surroundings of the vehicle; and

an image processing unit having a processor and a memory,

the image processing unit being configured to;

extract possible image areas for a predetermined part of a desired type of object in the captured image;

set a predetermined area below each extracted image area;

when no pattern different from a background is captured in the predetermined area, or when a pattern different from the background is captured only in one of a first and a second areas each of which is a half of the predetermined area divided in a horizontal direction, exclude the extracted area related to said predetermined area and determine candidates for the desired type of object as objects in the remaining extracted image areas;

determine, for each of the candidates, whether the candidate is the desired type of object.

2. The device according to claim 1, wherein the imaging unit is an infrared camera, wherein the image processing unit is configured to extract the possible image areas as high-intensity areas each having an intensity value higher than a predetermined value, and

wherein the image processing unit is configured to determine whether left-right symmetry of an object in the first and the second areas is lower than a predetermined level by examining the left-right symmetry on the basis of intensity values in the first and the second areas, and to exclude the extracted image region related to said first and second areas if the determination is positive.

3. The device according to claim 2, wherein the image processing unit is configured to determine that the left-right symmetry is lower than a predetermined level when a difference between a sum of intensity values of pixels in the first area and a sum of intensity values of pixels in the second area is not smaller than a predetermined value, or when a difference between a variance of the intensity values of the pixels in the first area and a variance of the intensity values of the pixels in the second area is not smaller than a predetermined value.

4. A method of monitoring surroundings of a vehicle, comprising steps of:

capturing an image of surroundings of the vehicle with an imaging unit;

extracting possible image areas for a predetermined part of a desired type of object in the captured image;

setting a predetermined area below each extracted image areas;

when no pattern different from a background is captured in the predetermined area, or when a pattern different from the background is captured only in one of a first and a second areas each of which is a half of the predetermined area divided in a horizontal direction, excluding the extracted area related to said predetermined area and determine candidates for the desired type of objects as objects in the remaining extracted image areas;

determining, for each of the candidates, whether the candidate is the desired type of object.

5. The method according to claim 4, wherein the step of extracting comprises extracting the possible image areas for a predetermined part of a desired type of object as high-intensity areas each having an intensity value higher than a predetermined value, and

wherein the step of excluding comprises:

determining whether left-right symmetry of an object in the first and the second areas is lower than a predetermined level by examining the left-right symmetry on the basis of intensity values in the first and the second areas, and excluding the extracted image region related to said first and second areas if the determination is positive.

6. The method according to claim 5, wherein said excluding an object in the image area comprises:

determining that the left-right symmetry is lower than a predetermined level when a difference between a sum of intensity values of pixels in the first area and a sum of intensity values of pixels in the second area is not smaller than a predetermined value, or when a difference between a variance of the intensity values of the pixels in the first area and a variance of the intensity values of the pixels in the second area is not smaller than a predetermined value.