Obstacle detection apparatus and method

• US 20010019356A1
• Filed: 02/09/2001
• Published: 09/06/2001
• Est. Priority Date: 02/29/2000
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Hattori, Hiroshi, Onoguchi, Kazunori, Takeda, Nobuyuki
• Granted Patent: US 6,801,244 B2
• Time in Patent Office: Days
• US Class Current: 348/119
• CPC Class Codes: H04N 7/181 for receiving images from a plurality of remote sources

Abstract

A plurality of cameras respectively input an image from different camera positions. The plurality of cameras is on a moving object. An image memory stores a plurality of images input by the plurality of cameras. An image transformation unit transforms one image input by a first camera using each of a plurality of transformation parameters each representing a geometrical relationship among a predetermined plane, the first camera, and a second camera, and generates a plurality of transformed images from a view position of the second camera. A matching processing unit compares each of the plurality of transformed images with another image input by the second camera for each area consisting of pixels, and calculates a coincidence degree of each area between each transformed image and another image. An obstacle detection unit detects an obstacle area consisting of areas each having a coincidence degree below a threshold from another image.

Claims

1. An obstacle detection apparatus, comprising:

a plurality of image input units configured to input images from at least two different camera positions, said plurality of image input units being on a moving object;

an image memory configured to store a plurality of images input by said plurality of image input units;

an image transformation unit configured to transform one image input by a first image input unit using each of a plurality of transformation parameters each representing a geometrical relationship among a predetermined plane, the first image input unit and a second image input unit, and to generate a plurality of transformed images from a camera position of the second image input unit, each of the plurality of transformed images corresponding to each of the plurality of transformation parameters;

a matching processing unit configured to compare each of the plurality of transformed images with another image input by the second image input unit for each area consisting of pixels, and to calculate a coincidence degree of each area between each transformed image and another image; and

an obstacle detection unit configured to detect an obstacle area consisting of areas each having a coincidence degree below a threshold from another image.

2. The obstacle detection apparatus according to

claim 1

,

wherein the first image input unit and the second image input unit are respectively a left camera and a right camera on a vehicle in order to take a left image and a right image of front area along advance direction of the vehicle.

3. The obstacle detection apparatus according to

claim 2

,

wherein an optical axis of the left camera and an optical axis of the right camera are parallel along the advance direction.

4. The obstacle detection apparatus according to

claim 2

,

wherein the plurality of transformation parameters respectively correspond to different postures of the vehicle on a road plane, and each transformation parameter is previously calculated based on the geometrical relationship among the road plane, the left camera, and the right camera for corresponding postures of the vehicle.

5. The obstacle detection apparatus according to

claim 2

,

wherein said image transformation unit transforms one of the left and right images into a plurality of images using each transformation parameter from a view position of the other of the left and right cameras.

6. The obstacle detection apparatus according to

claim 1

,

wherein said image transformation unit transforms the one image using a standard transformation parameter, and generates one transformed image from the camera position of the second image input unit.

7. The obstacle detection apparatus according to

claim 1

,

wherein said matching processing unit divides each transformed image into a plurality of areas, compares each area with a search area including a correspondence area on another image by shifting the correspondence area in the search area, and calculates the coincidence degree between the area and the correspondence area for each shift position.

8. The obstacle detection apparatus according to

claim 7

,

wherein said matching processing unit selects a maximum coincidence degree from coincidence degrees of the plurality of transformed images for each area of the same position, and creates a coincidence degree image consisting of maximum coincidence degrees of each area.

9. The obstacle detection apparatus according to

claim 7

,

wherein, if a coincidence degree of one shift position for the area of one transformed image is above a threshold, said matching processing unit selects the coincidence degree as the maximum coincidence degree for the area and completes the matching processing for the same area of all transformed images.

10. The obstacle detection apparatus according to

claim 7

,

wherein at least one of a size of the search area, a shift pitch of pixels in the search area, and a number of the transformation parameters, is changed according to a position of the area on the transformed image.

11. An obstacle detection method, comprising the steps of:

inputting images from at least two different camera positions through first and second cameras on a moving object;

storing a plurality of input images;

transforming one image input by the first camera using each of a plurality of transformation parameters each representing a geometrical relationship among a predetermined plane, the first camera, and the second camera;

generating a plurality of transformed images from the camera position of the second camera;

comparing each of the plurality of transformed images with another image input by the second camera for each area consisting of pixels;

calculating a coincidence degree of each area between each transformed image and another image; and

detecting an obstacle area consisting of areas each having a coincidence degree below a threshold from another image.

12. The obstacle detection method according to

claim 11

,

wherein the first camera and the second camera are respectively a left camera and a right camera on a vehicle in order to take a left image and a right image of an area in front of the vehicle.

13. The obstacle detection method according to

claim 12

,

wherein optical an axis of the left camera and an optical axis of the right camera are parallel along the advance direction.

14. The obstacle detection method according to

claim 12

,

wherein the plurality of transformation parameters respectively correspond to different postures of the vehicle on a road plane, and each transformation parameter is previously calculated based on the geometrical relationship among the road plane, the left camera, and the right camera for corresponding postures of the vehicle.

15. The obstacle detection method according to

claim 12

,

further comprising the step of:

transforming one of the left and right images into a plurality of images using each transformation parameter from a view position of the other of the left and right cameras.

16. The obstacle detection method according to

claim 11

,

further comprising the steps of:

transforming the one image using a standard transformation parameter; and

generating one transformed image from a view position of the second camera.

17. The obstacle detection method according to

claim 11

,

further comprising the steps of:

dividing each transformed image into a plurality of areas;

comparing each area with a search area including a correspondence area on another image by shifting the correspondence area in the search area; and

calculating the coincidence degree between the area and the correspondence area for each shift position.

18. The obstacle detection method according to

claim 17

,

further comprising the steps of:

selecting a maximum coincidence degree from coincidence degrees of the plurality of transformed images for each area of the same position; and

creating a coincidence degree image consisting of maximum coincidence degrees of each area.

19. The obstacle detection method according to

claim 17

,

further comprising the steps of:

if a coincidence degree of one shift position for the area of one transformed image is above a threshold,

selecting the coincidence degree as the maximum coincidence degree for the area; and

completing the matching processing for the same area of all transformed images.

20. The obstacle detection method according to

claim 17

,

wherein at least one of a size of the search area, a shift pitch of pixels in the search area, and a number of the transformation parameters is changed according to a position of the area on the transformed image.

21. A computer-readable memory containing computer-readable instructions, comprising:

an instruction unit to input images from at least two different camera position through first and second cameras on a moving object;

an instruction unit to store a plurality of input images;

an instruction unit to transform one image input by the first camera using each of a plurality of transformation parameters each representing a geometrical relationship among a predetermined plane, the first camera, and the second camera;

an instruction unit to generate a plurality of transformed images from the camera position of the second camera;

an instruction unit to compare each of the plurality of transformed images with another image input by the second camera for each area consisting of pixels;

an instruction unit to calculate a coincidence degree of each area between each transformed image and another image; and

an instruction unit to detect an obstacle area consisting of areas each having a coincidence degree below a threshold from another image.