METHOD FOR SELF LOCALIZATION USING PARALLEL PROJECTION MODEL

• US 20100245564A1
• Filed: 10/25/2007
• Published: 09/30/2010
• Est. Priority Date: 09/12/2007
• Status: Active Grant

• Current Assignee: Ajou University Industry Cooperation Foundation
• Original Assignee: Ajou University Industry Cooperation Foundation
• Inventors: Cho, We-Duke, Kyong, Yun Tai, Cho, Shung Han, Nam, Yun Young, Hong, Sang Jin
• Granted Patent: US 8,432,442 B2
• Time in Patent Office: Days
• US Class Current: 348/135
• CPC Class Codes: G06T 2207/30244 Camera pose
• CPC Class Codes: G06T 2207/30252 Vehicle exterior; Vicinity of vehicle
• CPC Class Codes: G06T 7/73 using feature-based methods

Abstract

A method of recognizing a self location of an image acquisition device by acquiring an image of two or more reference objects is provided. The method of the present invention comprises setting an actual camera plane, two or more reference object planes, and two or more virtual viewable planes located between the actual camera plane and the reference object planes; projecting the reference objects to a corresponding one of the virtual viewable planes; calculating a distance between a viewing axis and the reference objects and a distance between the viewing axis and images on the actual camera plane, the images corresponding to the reference objects; and sensing the self location of the image acquisition device by using an orientation and a zoom factor of the image acquisition device and coordinates of the reference objects, wherein the zoom factor is a ratio of a length of the reference object plane and a distance between the reference object plane and the virtual viewable plane, and the actual camera plane, the virtual viewable plane, and the reference object plane are perpendicular to the viewing axis.

Claims

1. A method of recognizing a self location of an image acquisition device by acquiring an image of two or more reference objects, the method comprising:

setting an actual camera plane, two or more reference object planes, and two or more virtual viewable planes located between the actual camera plane and the reference object planes;

projecting the reference objects to a corresponding one of the virtual viewable planes;

calculating a distance between a viewing axis and the reference objects and a distance between the viewing axis and images on the actual camera plane, the images corresponding to the reference objects; and

sensing the self location of the image acquisition device by using an orientation and a zoom factor of the image acquisition device and coordinates of the reference objects,

wherein the zoom factor is a ratio of a length of the reference object plane and a distance between the reference object plane and the virtual viewable plane, and

the actual camera plane, the virtual viewable plane, and the reference object plane are perpendicular to the viewing axis.

2. The method of claim 1, further comprising compensating the self location by using a compensation table,

wherein the compensating comprises:

estimating the self location and the orientation by using a zoom factor corresponding to an infinite distance;

calculating a distance between the reference object planes and the virtual viewable plane corresponding thereto, according to the estimated location and orientation; and

recalculating the location and the orientation by using the zoom factors using the distance, selected from the compensation table.

3. A method of recognizing a self location of an image acquisition device by acquiring an image of three or more reference objects, the method comprising:

setting an actual camera plane, three or more reference object planes, and three or more virtual viewable planes located between the actual camera plane and the reference object planes;

projecting the reference objects to a corresponding one of the virtual viewable planes;

calculating a distance between a viewing axis of the image acquisition device and the reference objects and a distance between the viewing axis and images on the actual camera plane, the images corresponding to the reference objects;

calculating an orientation of the image acquisition device by selecting an initial orientation by using two pairs of reference points to minimize an error distance and repeatedly computing the orientation of the image acquisition device by using the initial orientation and the error distance; and

sensing the self location by using the orientation, a zoom factor, and coordinates of the reference objects,

wherein the zoom factor is a ratio of a length of the reference object plane and a distance between the reference object plane and the virtual viewable plane, and

the actual camera plane, the virtual viewable plane, and the reference object plane are perpendicular to the viewing axis.

4. The method of claim 3, wherein the calculating an orientation of the image acquisition device comprises:

selecting an orientation allowing the error distance to be minimized, as the initial orientation by using the two pairs of the reference points;

selecting a smaller one between an error distance corresponding to the initial orientation and an error distance corresponding to an orientation greater than the initial orientation by 90 degrees, as an initial error distance;

calculating an approximate orientation by deducting a minute orientation from the initial orientation, calculating an approximate error distance corresponding to the approximate orientation, and calculating a slope at a location corresponding to the initial orientation by using a ratio of a difference between the initial error distance and the approximate error distance corresponding to the approximate orientation and the minute orientation;

setting a direction of iteration to be negative and setting the approximate error distance to be a present error distance when the slope is positive and setting the direction of iteration to be positive when the slope is not positive;

setting an absolute value of the slope as an initial slope, setting a ratio of the present error distance and a present slope as a next minute orientation, setting a value obtained by adding a value obtained by multiplying the direction of iteration by the minute orientation to a present orientation as a next orientation, and setting an error distance corresponding to the present orientation as a new present error distance;

reiterating the setting operations until the minute orientation becomes a predetermined threshold and repeatedly computing a value obtained by multiplying a ratio of the present error distance and a value obtained by adding to the present error distance to a previous error distance by a present minute orientation as the next orientation; and

determining the previous orientation repeatedly computed to be the orientation of the image acquisition when the minute orientation is the predetermined threshold or less.

5. The method of claim 3, further comprising compensating the location by using a compensation table,

wherein the compensating comprises:

estimating the location and the orientation by using a zoom factor corresponding to an infinite distance;

calculating a distance between the reference object planes and the virtual viewable plane corresponding thereto, according to the estimated location and orientation; and

recalculating the location and orientation by using the zoom factors using the distance, the zoom factors selected from the compensating table.