1) Describe Projection Schemes For Three-Dimensional Vision. A) It Is Common in Engineering Drawings To Provide Three Views of An

1) Describe Projection Schemes for Three-Dimensional Vision.
A) It is common in engineering drawings to provide three views of an

object to be manufactured—the plan, the side view, and the
elevation.
These views are simple orthographic projections of the object—i.e.,

they are made by taking sets of parallel lines from points on the
object to the flat plane on which it is being projected.
When objects are viewed by eye or from a camera, rays converge to

the lens and so images formed in this way are subject not only to
change of scale but also to perspective distortions.
This type of projection is called perspective projection, although it

includes orthographic projection as the special case of viewing from
a distant point.
This type of projection is called perspective projection, although it

includes orthographic projection as the special case of viewing from
a distant point.
In outdoor scenes, it is very common to see lines that are known to
be parallel apparently converging toward a vanishing point on the
horizon line.
In fact, the horizon line is the projection onto the image plane of the
line at infinity on the ground plane G: it is the set of all possible
vanishing points for parallel lines on G.
In general, the vanishing points of a plane ‘P’ are the projections

onto the image plane corresponding to points at infinity in different
directions on P.
Thus, any plane Q within the field of view may have vanishing
points in the image plane, and these will lie on a vanishing line,
which is the analog of the horizon line for Q.
2) Explain Shape from Texture.
A) Texture can be very helpful to the human eye in permitting depth to
be perceived.
Although textured patterns can be very complex, even the simplest

textural elements can carry depth information.
Ohta et al. (1981) showed how circular patches on a flat surface

viewed more and more obliquely in the distance become first
elliptical and then progressively flatter and flatter.
To disentangle such textured images sufficiently to deduce depths

within the scene, it is first necessary to find the horizon line reliably.
This is achieved by taking all pairs of texture elements and deducing

from their areas where the horizon line would have to be.
To proceed, we make use of the rule:

This information is accumulated in a separate image space and a line
is then fitted to these data: false alarms are eliminated automatically
by this Hough-based procedure.
3) Explain Shape from Shading.

A) The principle underlying this technique is that of modeling the
reflectance of objects in the scene as a function of the angles of
incidence I and emergence e of light from their surfaces.
In fact, a third angle is also involved, and it is called the “phase” g

(Fig. 15.8).
A general model of the situation gives the radiance I in terms of the

irradiance E and the reflectance R:
It is well known that a number of matt surfaces approximate
reasonably well to an ideal Lambertian surface whose reflectance
function depends only on the angle of incidence i—i.e., the angles
of emergence and phase are Immaterial.
E is regarded as a constant and is combined with other constants for

the camera and the optical system
gives a reflectance map in gradient (p, q) space
We now temporarily set the absolute reflectance value R0 equal to

unity.
The reflectance map can be drawn as a set of contours of equal
brightness, starting with a point having R=1 at s= n, and going down
to zero for n perpendicular to s. When s=v, so that the light source is
along the viewing direction shown in Figure
In a more general case, when s =/ v, zero brightness occurs along a
straight line in gradient space (Fig. 15.9(b)). To find the exact shapes
of the contours, we can set R at a constant
value a, which results in:
When objects are at all shiny—such as metal, plastic, liquid, or even

wood surfaces—the specular peak is quite sharp and rather intense:
casual observation may not even indicate the presence of another
peak, since Lambertian reflection is so diffuse (Fig. 15.10).
4) Explain Photometric Stereo.
A) PHOTOMETRIC STEREO:
Photometric stereo is a form of structured lighting that increases the

information available from surface reflectance variations.
Instead of taking a single monocular image of a scene illuminated

from a single several images are source taken, from the same
vantage point, with the scene illuminated in turn by separate light
sources.
The basic idea of photometric stereo is that of cutting down the

number of possible positions in gradient space for a given point on
the surface of an object.
It has already been seen that, for known absolute reflectance R0, a
constant brightness in one image permits the surface orientation to
be limited to a curve of conic cross-section in gradient space.
This would also be true for a second such image, the curve being a
new one if the illuminating source is different.
In general, two such conic curves meet at two points, so there is now
only a single ambiguity in the gradient of the surface at any given
point in the image.
To resolve this ambiguity a third source of illumination can be

employed.
The third image gives another curve in gradient space that should
pass through the appropriate crossing point of the first two curves
(Fig. 15.11)
Advantages:
One is that information on the absolute surface reflectance can be

obtained.
Another is that the assumption of a Lambertian surface can be tested.
Thus, three sources of illumination ensure that the remaining

ambiguity is resolved and permit absolute reflectivity to be
measured.
The calculation is normally carried out by defining a set of nine

matrix components of irradiance, sij being the jth component of light
source vector si. Then, in matrix notation
E= R0Sn
5) Explain Shape from Texture.
A) Same as 2nd answer…

1) Describe Projection Schemes For Three-Dimensional Vision. A) It Is Common in Engineering Drawings To Provide Three Views of An

Uploaded by

Copyright:

Available Formats

1) Describe Projection Schemes For Three-Dimensional Vision. A) It Is Common in Engineering Drawings To Provide Three Views of An

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1) Describe Projection Schemes For Three-Dimensional Vision. A) It Is Common in Engineering Drawings To Provide Three Views of An

Uploaded by

Copyright:

Available Formats

1) Describe Projection Schemes for Three-Dimensional Vision.

A) It is common in engineering drawings to provide three views of an

These views are simple orthographic projections of the object—i.e.,

When objects are viewed by eye or from a camera, rays converge to

This type of projection is called perspective projection, although it

This type of projection is called perspective projection, although it

In general, the vanishing points of a plane ‘P’ are the projections

Although textured patterns can be very complex, even the simplest

Ohta et al. (1981) showed how circular patches on a flat surface

To disentangle such textured images sufficiently to deduce depths

This is achieved by taking all pairs of texture elements and deducing

To proceed, we make use of the rule:

3) Explain Shape from Shading.

In fact, a third angle is also involved, and it is called the “phase” g

A general model of the situation gives the radiance I in terms of the

E is regarded as a constant and is combined with other constants for

gives a reflectance map in gradient (p, q) space

We now temporarily set the absolute reflectance value R0 equal to

When objects are at all shiny—such as metal, plastic, liquid, or even

Photometric stereo is a form of structured lighting that increases the

Instead of taking a single monocular image of a scene illuminated

The basic idea of photometric stereo is that of cutting down the

To resolve this ambiguity a third source of illumination can be

One is that information on the absolute surface reflectance can be

Another is that the assumption of a Lambertian surface can be tested.

Thus, three sources of illumination ensure that the remaining

The calculation is normally carried out by defining a set of nine

You might also like