Human being detection apparatus, method of detecting human being, and human being detecting program

• US 20070165931A1
• Filed: 12/06/2006
• Published: 07/19/2007
• Est. Priority Date: 12/07/2005
• Status: Active Grant

• Current Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Higaki, Nobuo
• Granted Patent: US 7,840,036 B2
• Time in Patent Office: Days
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/251 involving models
• CPC Class Codes: G06V 20/64 Three-dimensional objects
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

A human being detection apparatus includes a moving object detecting part for detecting a target region from the picture image as a region of the moving object and a target distance as a distance to the moving object, a head region setting part for setting the head region corresponding to the target distance, a sample storing image part for storing a sample image having a predetermined size corresponding to the head region of the human being, and a human being determining part for determining whether or not the moving object is a human being by comparing the head region in the picture image converted to have a predetermined size with the sample image.

Claims

1. A human being detection apparatus for detecting a human being in a plurality of picture images including a moving object shot by a plurality of cameras, comprising:

a moving object detecting part that detects a target region of the moving object in the picture images and a target distance to the moving object from the camera as moving object information on the basis of parallaxes in picture images and a difference between picture images sequentially entered;

a head region setting part that sets a head region as having a predetermined size corresponding to the target distance in the target region detected by the moving object detecting part when it is assumed that the moving object is the human being;

a sample image storing part for storing a sample image having a predetermined size corresponding to the head region of the human being; and

a human being determining part that determines whether the moving object is the human being by converting the head region, set by the head region setting part, in the picture image into an image having the predetermined size and comparing the converted head region with the sample image to determine that the moving object is the human being.

2. The human being detection apparatus as claimed in claim 1, wherein the sample image is an image indicating a shape of the head region, and the human being determining part includes:

a head edge image generating part that generates a head edge image as an edge image corresponding to the head region set by the head region setting part;

a size normalizing part that converts the head edge image generated by the head edge image generating part into a normalized image having the predetermined size;

a complementing part that generates a complemented image in which gaps in the normalized image converted by the size normalizing part are filled; and

a similarity determining part that determines that the moving object is the human being on the basis of a similarity between the complemented image and the sample image.

3. The human being detection apparatus as claimed in claim 1, further comprising an image shifting part that adjusts a reference position of the complemented image to a predetermined reference position of the sample image.

4. The human being detection apparatus as claimed in claim 2, wherein the complementing part fills gaps in the head edge image by enlarging and reducing for a region of the head edge image generated by the head edge image generating part a plurality of times, respectively.

5. The human being detection apparatus as claimed in claim 1, wherein the sample image storing part stores a plurality of sample images of a sample human being in which a direction of the sample human being varies from one of the sample image to another one of the sample images.

6. The human being detection apparatus as claimed in claim 1, wherein the moving object detecting part outputs the moving object information when the target distance is within a predetermined range which is previously determined in accordance with a measured accuracy in determining the human being by the human being determining part.

7. A method of detecting a human being included in a plurality of picture images including a moving object shot by a plurality of cameras, comprising the steps of:

detecting a target region of the moving object in the picture images and a target distance to the moving object from the camera as moving object information on the basis of parallaxes in picture images and a difference between picture images sequentially entered;

setting a head region as having a predetermined size corresponding to the target distance in the target region when it is assumed that the moving object is the human being;

generating a head edge image as an edge image corresponding to the head region;

converting the head edge image into a normalized image having the predetermined size;

generating a complemented image in which gaps in the normalized image converted by the size normalizing part are filled;

calculating a similarity between the complemented image and a sample image of the head region having the predetermined size previously stored in a sample image storage; and

determining that the moving object is the human being on the basis of the similarity.

8. A human being detecting program for causing a computer to operate, to detect a human being from a plurality of picture images shot by a plurality of cameras, to have functions of:

a moving object detecting part that detects a target region of the moving object in the picture images and a target distance to the moving object from the camera as moving object information on the basis of parallaxes in picture images and a difference between picture images sequentially entered;

a head region setting part that sets a head region as having a predetermined size corresponding to the target distance in the target region detected by the moving object detecting part when it is assumed that the moving object is the human being;

a sample image storing part for storing a sample image having a predetermined size corresponding to the head region of the human being; and

a human being determining part that determines whether the moving object is the human being by converting the head region set by the head region setting part in the picture image into an image having the predetermined size and comparing the converted head region with the sample image to determine the moving object is the human being.