1. A
SEMINAR
ON
“CONCEPT OF STEREO VISION BASED VIRTUAL
TOUCH SCREEN”
VIVEK R. CHAMORSHIKAR
2. WHAT IS STEREO VISION?
Stereo Vision is a by product of good binocular vision.
BINOCULAR: Involving both eyes at once.
BINOCULAR VISION: Here both eyes aim
simultaneously at the same visual target, vision in which
both eyes work together as a coordinated team equally
and accurately.
STEREO VISION:(stereopsis or stereoscopic vision)
Vision in which two separate images from two eyes are
successfully combined into one image in the brain.
How it works?
3. Why to use Stereo Vision?
Stereo Vision is related to stereopsis.
Stereopsis (stereo means “three-dimensional” or “solid”
and opsis means “sight” or “view”).
Basic Ability of Stereo Vision: The ability to infer
information on the 3-D structure and distance of a scene
from two or more images taken from two different
viewpoints.
Stereo vision is most cost efficient way, instead of using
the costly sensors.
4. Requirements for the system are as-
1. Mouse input should be replaced by touch input.
Create active/inactive spaces for interactions.
2. GUI applications should be designed to enable touch input
events.
Fig 1. Figure showing the efforts faced
between human and machine interaction.
5. 3. Two cameras are needed.
It helps to distinguish interactive parts of captured image .
Accurate and reliable 3-D image is captured.
Accurate dimensions are calculated.
4. Synchronization is needed by two cameras.
The image frames should be captured from two cameras
at the same time and also frame rate of two cameras
should be same.
5. Distance Calibration.
The calibration of distance of blob (object used for input)
should be nearest to the actual distance of screen for good
result.
6. PROBLEMS IN STEREO VISION
Problems to solve in stereo vision are:
1. Correspondence Problem
2. Calibration Problem
3. Synchronization Problem
4. Shadow Problem
5. Sunlight Problem
7. SOLUTION FOR CORRESPONDENCE
PROBLEM
Two algorithms to solve correspondences problem
Correlation-based Algorithm- Checking if one location in
one image looks/seems like another in another image.
Produce a DENSE set of correspondences.
Feature-based Algorithm - Finding features in the image
and seeing if the layout of a subset of features is similar in
the two images.
Produce a SPARSE set of correspondences.
8. APPLICATIONS OF STEREO VISION
1.People Tracking
2.Robotics
3.Random Bin Picking (RBP)
4.Surgeries
5.3-D Underwater Mosaicking
Stereo Vision has many Other Applications:
Driver assistance system
Forensics - Crime Scenes, Traffic Accidents
Mining - Mine face measurement
Civil Engineering - Structure monitoring
Collision Avoidance
Manufacturing- Process Monitoring
9. ADVANATAGES AND DISADVANTAGES OF
STEREO VISION
Advantages of Stereo Vision:
1. Robustness
2. Gives a very dense depth (or range) map.
3. Use to calculate shape of objects.
4. Human motion detection is possible instead of using sensors
for it.
Disadvantages of Stereo Vision:
1. The system must be pre-calibrated.
2. Has to be used in indoor environment
3. Shadow and sunlight present in the experimental area makes
difficult in distance calculation.
10. Tracking of Blob:
Novel algorithm is used for efficient motion detection and
calculating distance of blob.
Combining Blob:
After assigning all the labels to every pixel of the image we
count all the labels other than background labels (i.e. other
than 0) and store its corresponding (x, y) coordinates. The
pixels having same label is considered as a single object and
a box is drawn around it using the maximum and minimum x
and y coordinates.
Height Map:
In computer graphics, heightmap is a image used to store
values, such as surface elevation data, for display in 3D
Computer graphics.
11. STEREO RANGING:
Calculating the distance to objects by making a pair of
observations at different locations.
Range = (Focal length x Camera baseline) / Disparity
C0 - Left Camera
C1 -Right Camera
P -Observed feature point
F -focal length
B -baseline distance
D -distance to observed feature
point
c0, c1 -Pixel center of camera images
v0, p1 -pixel position of observed
feature point
v0, v1 -Pixel displacement of
observed feature point
Disparity (D) = v1-v0
Distance (D) = bf/d
13. CONCLUSION
Stereo vision
Applications
Requirements for the system to use stereo vision.
Advantages and Disadvantages of stereo vision.
The calculation of the distance of the blob from
two cameras.