Transcript 3-D Spatial Mapping Device

THE PRINT-SCAN Machine
3-D Spatial Mapping Device
Nia Cook
Stephen Tan
Anil Rohatgi
Senior Design
Final Report Presentation
ECE4006
Spring2005
Introduction
• Project Goal
Use of 3-D imaging techniques to measure
the detailed physical structure of the
interior of a confined space and map it into
a virtual 3-D environment
• Prototype
The PRINT-SCAN Machine
Project Specifications
• 10cm*10cm*10cm cubic volume
• Ability to capture physical detail (preferably at
the micron level)
• Ability to measure the size and shape of objects
• Ability to measure relative positions of multiple
objects within the volume
• Cannot employ imaging techniques using x-rays
Project Constraints
•
•
•
•
•
Objects are stationary within volume
Objects have low reflectivity
Objects are not in contact with neighbor
A four month time limit
Design cannot exceed $500 budget
Component List
•
•
•
•
•
•
•
•
Sharp GP2D12 distance measuring sensor
IR Mirrors
HP Inkjet Printers
HP 5-49A Ink Cartridges
HP 5-29A Ink Cartridges
D1984 Data Capture with WINDAQ software
Constructed ten centimeter volume
Driver Circuit (L298N and SN74LS04N)
Theoretical Design
Project Technical Details
• Box construction
– 10 cm cube with open top
– Tracks on inside to stabilize mirrors
– Flaps on box for data threshold segmentation
• Driver Circuit
– SN74LS04N inverter toggles the L298N H-bridge so
that printer moves back and forth
– Function generator provides 100 mHz square wave
as input
– Power supply inputs 7 – 8 V for reasonable printer
head speed
Project Technical Details
• Laser Sensors
– Read distance as a function of voltage
– Records voltages in Excel
– 10 cm to 80 cm range
• Mirrors
– Reflective for 850 nm laser sensor
– Angled at 45 degrees to reflect the laser
beam to the object
– Incremented upwards to capture object height
Prototype Design
Data Reconstruction
Distance 1- Power Regression
3
2.5
Voltage (V)
• Sensor
characterization
• Power regression
line:
8.0082x^(-0.837)
• Correlation
percent: 99.64%
• Inverse regression
applied to data
-0.873
y = 8.0082x
R2 = 0.9964
2
Series1
1.5
Power (Series1)
1
0.5
0
0
5
10
Distance (in)
15
20
Data Reconstruction
• Data Imported from Microsoft Excel to Matlab for
processing
• Data needs to be segmented into vertical divisions
Data Reconstruction
• Matrix structure and corresponding
coordinate values
Data Reconstruction
• Three reconstruction techniques
Spline fit
Point cloud
Mesh Grid
Data Reconstruction
• Video demonstration
result:
– Attempted to scan
two rubber wheels
staggered inside the
volume
– Managed to
reconstruct shape
and location,
however, recovering
the spacing between
the objects did not
function.
– Errors were in the
data, not in the data
processing
Lessons Learned
•
•
•
•
Scheduling
IR Sensor Interaction
Power Drive
Calculations
Conclusion
• Although the device did not perform as well in real life as expected,
there was adequate data to support proof of concept.
• With better equipment, and more funding, the design could be
extended to achieve the optimal goals of the project.
Initial Specifications:
Achieved Specifications:
10cm3 confined volume (top open)
The volume is approximately 10 cm3
Capture physical detail, at micron level
The actual resolution of the sensors do not
give the physical details of the object at
the micron level
Measure size and shape of objects
Our design outlines the shape of an object
Measure relative positions of multiple
objects within the volume
During our product demonstration, we
employed two objects, we were able to
calculate their positions relative to each
other, and however mapping the spacing
between the objects was a problem.
Occlusion was not a factor.
Cannot employ imaging techniques using xrays
Our device design, does not employ any
x-ray imaging techniques
Questions?
??