Pedestrian detection apparatus
View Patent ↗A pedestrian detection apparatus of the present invention extracts a whole-body region and a prospective head region based on a pattern matching between an input image picked up by an infrared camera and whole-body/head model images stored in a model-image storage unit. If the whole-body region has not been recognized, the prospective head region which is at the closest position to a prospective head region recognized from a preceding input image is recognized from the current input image, and an image matching with the head model image of the recognized prospective head region is determined to be a head.
1. The pedestrian detection apparatus for recognizing a pedestrian from an input image picked up by image pick-up means based on a pattern matching between the input image and model images of the pedestrian, comprising:
prospective pedestrian recognizing means for recognizing a whole-body region surrounding a whole-body model image and a prospective part region surrounding a part model image of the whole body from the input image based on a pattern matching using the whole-body model image and the part model image of the pedestrian; and
pedestrian recognizing means for determining whether or not a recognized prospective part region is a part of the pedestrian if a prospective part region exists in the whole-body region recognized by the prospective pedestrian recognizing means,
wherein the part model image is a head model image, and the pedestrian recognizing means determines the prospective part region at the highest position to be a head if a plurality of prospective part regions exist in the whole-body region.
2. The pedestrian detection apparatus according to claim 1 , wherein:
the image pick-up means is one of a far-infrared camera and an infrared camera.
3. The pedestrian detection apparatus according to claim 1 , wherein:
the image pick-up means is provided in at least one of the vehicle and the street.
4. A pedestrian detection apparatus, for recognizing a pedestrian from an input image picked up by image pick-up means based on a pattern matching between the input image and model images of the pedestrian, comprising:
prospective pedestrian recognizing means for recognizing a whole-body region surrounding a whole-body model image and a prospective part region surrounding a part model image of the whole body from the input image based on a pattern matching using the whole-body model image and the part model image of the pedestrian; and
pedestrian recognizing means for determining whether or not a recognized prospective part region is a part of the pedestrian if a prospective part region exists in the whole-body region recognized by the prospective pedestrian recognizing means, wherein
the pedestrian recognizing means compares a preceding prospective part region recognized from a preceding input image with a prospective part region recognized from the current input image if the whole-body region has not been recognized by the prospective pedestrian recognizing means and, determines that the prospective part region recognized from the current input image is the part of the pedestrian if the distance between the both prospective part regions is equal to or smaller than a predetermined distance.
5. The pedestrian detection apparatus according to claim 4 , wherein:
the part model image is the head model image.
6. The pedestrian detection apparatus according to claim 5 , wherein:
the image pick-up means is one of the far-infrared camera and the infrared camera.
7. The pedestrian detection apparatus according to claim 5 , wherein:
the image pick-up means is provided in at least one of the vehicle and the street.
8. The pedestrian detection apparatus according to claim 4 , wherein:
the image pick-up means is one of the far-infrared camera and the infrared camera.
9. The pedestrian detection apparatus according to claim 4 , wherein:
the image pick-up means is provided in at least one of the vehicle and the street.
10. A pedestrian detection apparatus, for recognizing a pedestrian from an input image picked up by image pick-up means based on a pattern matching between the input image and model images of the pedestrian, comprising:
prospective pedestrian recognizing means for recognizing a whole-body region surrounding a whole-body model image and a prospective part region surrounding a part model image of the whole body from the input image based on a pattern matching using the whole-body model image and the part model image of the pedestrian; and
pedestrian recognizing means for determining whether or not a recognized prospective part region is a part of the pedestrian if a prospective part region exists in the whole-body region recognized by the prospective pedestrian recognizing means, wherein:
the pedestrian recognizing means compares a preceding prospective part region recognized from a preceding input image with one or more prospective part regions recognized from the current input image if the whole-body region has not been recognized by the prospective pedestrian recognizing means, and determines the prospective part region being at the closest position to the preceding prospective part region and being within a predetermined distance among the recognized prospective part regions to be the part of the pedestrian.
11. The pedestrian detection apparatus according to claim 10 , wherein:
the part model image is the head model image.
12. The pedestrian detection apparatus according to claim 11 , wherein:
the image pick-up means is one of the far-infrared camera and the infrared camera.
13. The pedestrian detection apparatus according to claim 11 , wherein:
the image pick-up means is provided in at least one of the vehicle and the street.
14. The pedestrian detection apparatus according to claim 10 , wherein:
the image pick-up means is one of the far-infrared camera and the infrared camera.
15. The pedestrian detection apparatus according to claim 10 , wherein:
the image pick-up means is provided in at least one of the vehicle and the street.