Method for estimating the motion of a carrier relative to an environment and computing device for navigation system

• US 20120070037A1
• Filed: 09/13/2011
• Published: 03/22/2012
• Est. Priority Date: 09/14/2010
• Status: Active Grant

• Current Assignee: Astrium Sas (Airbus SE)
• Original Assignee: Astrium Sas (Airbus SE)
• Inventors: POLLE, Bernard
• Granted Patent: US 8,548,197 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G01C 21/005 with correlation of navigation data from several sources, e.g. map or contour matching G01C21/30 takes precedence
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30244 Camera pose
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters

Abstract

A method (10) for estimating a carrier'"'"'s movement relative to an environment (20) with respect to which the carrier moves, the carrier having at least one navigation sensor and one vision sensor producing 2D images of the environment installed, The method (10) includes the following steps: i) (12) identifying, in images acquired by the vision sensor, characteristic elements of the images representing characteristic elements of the environment, ii) (14) calculating, based on characteristic elements of a triplet of images acquired at different times, at least one condensed measurement representative of characteristics of the carrier'"'"'s movement during the acquisition of the triplet'"'"'s images, iii) (16) estimating the movement by a navigation filter, based on navigation measurements made by the navigation sensor and on the at least one condensed measurement. A computing device for a navigation system of a carrier is also described.

Claims

1-15. -15. (canceled)

16. Method (10) for estimating a carrier'"'"'s movement relative to an environment (20) with respect to which said carrier is moving, the carrier having at least one navigation sensor and at least one vision sensor producing 2D images of the environment installed, wherein said method comprises the following steps:

(12) identifying, in images acquired by the vision sensor, characteristic elements of the images representing characteristic elements of the environment,

(14) calculating at least one condensed measurement, based on characteristic elements of a triplet of images acquired at different times, the at least one condensed measurement being representative of characteristics of the carrier'"'"'s movement during the acquisition of said triplet'"'"'s images,

(16) estimating the movement by an estimate filter, called the “

navigation filter”

, based on navigation measurements made by the navigation sensor and on the at least one condensed measurement.

17. The method (10) according to claim 16, wherein, during the calculation step (14), a condensed measurement is determined that is representative of a ratio of the distances traveled by the carrier between the acquisition times of two image pairs of said triplet, based on characteristic elements of the triplet'"'"'s images representing characteristic elements of the environment visible on each of said triplet'"'"'s three images.

18. The method (10) according to claim 16, wherein, during the calculation step (14), at least one condensed measurement is determined that is representative of a direction of the carrier'"'"'s displacement between the acquisition times of an image pair of the triplet, based on characteristic elements of this image pair representing characteristic elements of the environment visible on each of this pair'"'"'s two images.

19. The method (10) according to one claim 16, wherein, during the calculation step (14), at least one condensed measurement is determined that is representative of a normalized relative pose between the acquisition times of an image pair of the triplet, based on characteristic elements of this image pair representing characteristic elements of the environment visible on each of this pair'"'"'s two images.

20. The method (10) according to claim 16, wherein the characteristic elements are characteristic points and/or characteristic objects or one or more characteristic surfaces.

21. The method (10) according to claim 16, wherein the triplet'"'"'s images are images selected from a plurality of images acquired by the vision sensor.

22. The method (10) according to claim 21, wherein at least one of the triplet'"'"'s images is an image for which a carrier'"'"'s displacement detection criterion is verified.

23. The method (10) according to claim 22, wherein the carrier'"'"'s displacement detection criterion is verified for a second image when, with a first image selected and with a plurality of characteristic elements of the environment, N in number, visible in said first image:

at least Nb of the N characteristic elements of the environment have disappeared in said second image, or

at least Nb new characteristic elements of the environment have appeared in said second image,

Nb being substantially equal to r_P·

N, where r_P is a predefined percentage.

24. The method (10) according to claim 23, wherein the predefined percentage r_P is equal to or greater than 20%, preferably substantially equal to 25%.

25. The method (10) according to claim 16, comprising a step (18) of searching for aberrant characteristic elements amongst the plurality of characteristic elements of the images, characteristic elements regarded as aberrant not being considered in the calculation step (14).

26. The method (10) according to claim 25, wherein the step (18) searching for aberrant characteristic elements of the images comprises the following sub-steps:

(180) establishing groups of characteristic elements representative of at least five characteristic elements of the environment visible in each image of a pair and/or triplet to be used to calculate a condensed measurement,

(182) calculating, for each group of characteristic elements of the images, an estimate of the condensed measurement to be calculated,

(184) searching for groups of aberrant characteristic elements for which the estimate of said condensed measurement to be calculated is aberrant with respect to all the estimates of said condensed measurement to be calculated.

27. The method (10) according to claim 16, wherein the navigation filter'"'"'s state vector comprises the carrier'"'"'s state at the acquisition times of the triplet'"'"'s images.

28. The method (10) according to claim 27, wherein a condensed measurement representative of a normalized relative pose between the acquisition times of at least one pair of the triplet is supplied to the navigation filter substantially at the time the second image of said pair is acquired.

29. The method (10) according to claim 16, wherein, during the calculation step (14), the error in estimating the at least one condensed measurement is estimated and in that, in the estimation step (16), the carrier'"'"'s movement is estimated based on the estimate of the error made in estimating the at least one condensed measure.

30. Computing device for a navigation system of a carrier having at least one navigation sensor and at least one vision sensor installed, wherein said computing device comprises means configured to estimate the carrier'"'"'s movement according to claim 16.