Hello everyone and welcome back to Computer Vision Lecture Series.

This is Lecture 10 part 2.

We continue talking about the basic concepts of structured light.

Last time we saw that there are for generating depth maps from using structured light concepts

we already have information in the sense of we already know the geometric arrangement

of the camera and the source of the structured light.

So we already know the geometry.

We have already seen the algorithms which can compute correspondences using this information

in this kind of geometry.

So here in this case for example we replace one camera and we use a laser light source

which is called a single strip scanning because a laser can project only single line very

coherent strip of line and therefore it is called single strip camera and scanning means

that the objects are either on the conveyor belt or the source of light is moved along

the objects.

But usually in the industrial settings the objects are moved along the conveyor belt

and the source of light is fixed and there is already a camera CCD camera which is taking

images of these moving objects and accordingly when the objects move their profile or their

height profile or the depth profile in the 3D world is generated and captured and recorded

by the CCD camera using the already discussed correspondence algorithms.

So it is very easy to locate this single strip of line.

Filters are also very cheap and this setup is pretty simple to do and you don't need

high quality cameras either.

So this is a very robust technique for getting very precise or very accurate measurements

of depth.

So essentially what we do is in this optical triangulation strategy for example we have

this curved surface please do not consider this red lines but these red lines are the

projection of this source this strip line source which in our case is a laser right

now and the surface is this curved surface that you can see the curvature in this red

line and what we do is and we already have the camera location with respect to the source

and we move the source along this surface and then we capture the location of each and

every each and every strip of line along this surface when it moves from the camera and

therefore it is called single strip scanning.

Because of our arrangement of our system that is it is in the horizontal direction we have

disparities only in the horizontal direction and therefore we have already always seen

that the stripes of lines are vertical and only moved in this vertical horizontal direction

is the main reason.

Using this kind of strip scanning we can get a very precise and version of a structured

light and we can use this information to localize very with high accuracy the 3D locations of

the surface so we can actually get a nice curvatures of the surfaces that we are scanning.

However this scanning takes time because it needs a lot of images so imagine instead of

going by every pixel you have to go by every line so it is still very complex and depending

on the resolution of the strip line it may take a lot of number of images.

So let us see in a bit more mathematical sense how to find the location of this or how to

localize this 3D point using this structured light setting.

So first what you have is you have your object here and you have laser or a projector whichever

it is and you project this single strip of light on this object and then you capture

the scene using a camera located at a point O and since you already know the location

of the laser projector as well as the camera center and you already have this bright light

shining in your image plane you know this because this light is very distinctive it

has a high contrast and so on and so forth because first in the beginning your camera