Process for producing cartographic data by stereo vision

• US 6,175,648 B1
• Filed: 08/11/1998
• Issued: 01/16/2001
• Est. Priority Date: 08/12/1997
• Status: Expired due to Term

• Current Assignee: Cassidian Limited (Airbus SE)
• Original Assignee: Matra Systemes et Information
• Inventors: Sirat, Jacques Ariel, Canu, David, Ayache, Nicolas
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): SHERALI, ISHRAT I
• Application Number:
  US09/132,488
• Time in Patent Office: 889 Days
• Field of Search: 382/154, 382/153, 382/284, 382/106, 382/294, 382/293, 382/276, 345/139, 345/419, 345/424, 345/425, 345/427, 356/12, 356/2, 356/375-376, 356/390, 356/147, 348/42-48, 348/36, 348/51, 348/135, 348/159, 359/462, 359/618, 706/16
• US Class Current: 382/154
• CPC Class Codes: G06T 17/05 Geographic models
• CPC Class Codes: G06T 7/55 from multiple images
• CPC Class Codes: G06V 10/10 Image acquisition document image scanning and transmission H04N1/00; control of digital cameras H04N23/60
• CPC Class Codes: G06V 10/147 Details of sensors, e.g. sensor lenses fingerprint or palmprint sensors G06V40/13; vascular sensors G06V40/145; eye sensors G06V40/19
• CPC Class Codes: G06V 2201/12 Acquisition of 3D measurements of objects

Abstract

The method produces three dimensional cartographic data of a scene from n two-dimensional images of the scene, n being an integer higher than 2. The images are delivered by n respective detectors seeing the scene from different points of view. In a first step, each detector i is calibrated for estimating parameters identifying n affine models F_i (x, y, z) each defining a relation between coordinates x, y, z of any point in the scene and coordinates (p, q)_i of a projection of said any point in an image i among said n images. Each of (n−1) couples of images, (each couple consisting of a same reference image selected among said n images) is matched, each match being made by searching an homolog of each of pixels or zones constituting the reference image along a corresponding epipolar line of the other image of the couple. In each couple of two images, each comprising one reference image, and for each pixel or each zone of the reference image, a curve plotting a degree of similarity as a function of disparity along the epipolar line of the other image of the couple is prepared. All curves are brought into a common reference frame for matching a maximum number of images. All curves are summed and the highest peak is retained. The coordinates x, y, z of each point are derived from the disparity of the retained peak and from the parameter of the n models F_i (x, y, z).

Claims

1. A method for producing three dimensional cartographic data of a scene from n two-dimensional images of the scene, n being an integer higher than 2, said images being delivered by n respective detectors seeing the scene from different points of view, comprising the steps of:

a) calibrating each detector i among said detectors for estimating parameters identifying n models F_i (x, y, z) each defining a relation between coordinates x, y, z of any point in the scene and coordinates (p, q)_i of a projection of said any point in an image i among said n images;

b) matching each of (n−

1) couples of said images, each of said couples consisting of a same reference image selected among said n images, each match being made by searching an homolog of each of a plurality of pixels or zones constituting said reference image along a corresponding epipolar line of the other image of the couple;

c) in each of said (n−

1) couples of two images, each comprising one reference image, and for each pixel or each zone of the reference image, preparing a curve plotting a degree of similarity as a function of disparity along the epipolar line of the other image of the couple;

d) bringing back all said curves into a common reference frame for matching a maximum number of said images;

e) summing said curves and retaining a highest peak among all peaks of the curve resulting from the summation; and

f) deriving the coordinates x, y, z of each point from the disparity of the retained peak and from the parameter of the n models F_i (x, y, z).

2. Method according to claim 1, wherein, during step (a) said curves are brought to a common reference frame by an affine model.

3. Method according to claim 1, comprising, during step (c), a further step of discarding any peak of said curve which has a degree of disparity which is inconsistent with the disparities on all the other curves.

4. Method according to claim 1, wherein a plurality of matchings are carried out with different reference images among all said images.

5. Method according to claim 1, including the step of combining results of matching between couples with a majority vote.

6. Method according to claim 5, wherein said majority vote is carried out while assigning a weighting coefficient which is higher for those peaks which correspond to points of view which are most remote from the reference image.