CDR Presentation.ppt

Multispectral Camera
Comprehensive Design Review
Team Parente
Simon Belkin, Audrey Finken, Grant George, &
Matthew Walczak
• Scientists desire the ability
to capture measurements
based on reflected energy
• Common camera
technology only captures
light in three primary
wavelengths: Red, Green,
and Blue (RGB) from the
visible spectrum of light
Red = Bowline type rocks
Green = Whittaker/Hillary (Jibsheet class)
Blue = bright drift soil
RGB = 450nm – 650nm
System Design
Deliverables Set at CDR
 Build optical system
 Camera can focus
 Capture images at different distances
 Aberration solution
 Algorithm selected
 Code in progress
 Stepper motor control
 Raspberry Pi control the stepper motor
 Can turn clockwise and counterclockwise
Optical System
Optical System
 Out of focus condition due to the addition of filter
wheel required compensation.
 3.75” (95.3mm) of extension tubing.
 A +33mm convex relay lens.
 Thin lens equation, 1/di = 1/f – 1/do, determines the
distance to the object, do, and to the image, di.
Ray Tracing
Optical Distortions
 Aberrations due to lens occur when the light
reflected off one point of an object does not
converge, (or diverge), to where it is expected to
Filter Wheel Mechanics
Filter Wheel Mechanics
 Problem
 Hall Effect sensor provided in the filter wheel burned out
 Filter wheel doesn’t have a stable starting position
 Current Solution
 Have a static number of rotation done through the stepper
motor to rotate from one filter location to the next
 Using Raspberry Pi Shell Script to turn the Filter wheel
 FDR Solution
 Program the filter wheel location in Python
 Replace Hall Effect Sensor and add to Python program to
identify start location
Stepper Motor Control
gpio -g mode 23 out
gpio -g mode 24 out
echo Turn Counter Clock Wise:
sleep 1
gpio -g write 24 1
for i in $(seq 400); do
echo $i
gpio -g write 23 1
gpio -g write 23 0
Setting gpio pin # 24 to 1 turns Counter Clock Wise and 0 Clock Wise.
Pi to Camera Interface
Pi to Camera Interface
 Problem
 Mightex Camera Driver Installation
 “Fedora 5” to Raspbian “wheezy” supported by the Raspberry Pi
 Current Solution
 Use Windows operating system to test Camera
 Write existing code in python which can be ported to Pi when
 FDR Solution
 Mightex Camera integrated to Raspberry Pi
 GPIO pins if necessary
 Capture image command
 Store image
 Analyze data
Image Processing
Image Processing - Problem
 An optical aberration is a deviation from the
optical predictions
 Optical aberration occurs when monochrome
image captured with and without filter
 It occurs because the optical path and focal length
is modified by putting the filter wheel between the
lens and the camera
 Types of optical aberrations
 Spherical aberration
 Chromatic aberration
Spherical Aberrations
 The aberration arises because the parallel light rays of
incoming light do not converge at the same point after
passing through the lens, resulting in a blurred image
 Cause - filters
Lens without Spherical
Chromatic Aberrations
 The aberration arises because the index of refraction of
the lens material varies with wavelength, therefore the
colors of the light passing through the lens refract by
different amounts and produce a blurred image
 Although we are using a monochromatic camera,
chromatic aberration will still occur and blur the image
Solution – Image Registration
 Image Registration is the process of estimating an optimal
transformation between two images
 SIFT (Scale Invariant Feature Transform) is used to generate a set
of ‘interest’ points which can be used for tracking an image
 Use the corresponding points, to calculate a homography matrix
 A homography matrix is a 3×3 matrix which decides the transformation
between two sets of coordinate systems
 RANSAC(Radom Sample Consensus) algorithm picks a minimum
number of points from a given set and estimates the model.
 It takes points from the set which closely fit the model and calculates the
error in the estimated model.
 This model ensures that only inliers are included and outliers are not
taken into account.
Image Processing - FDR Solutions
 Use image registration to remove blur caused by
 Use SIFT implementation and calculation of
homography matrix in python, and apply a RANSAC
algorithm to find the homography matrix and change
the first image accordingly so that it matches the
orientation of the second image
New Team-Member Roles
 Simon Belkin – Hall effect sensor, integration of sensor
and stepper motor code into Python
 Audrey Finken – Image processing & spherical
aberration mitigation
 Grant George – Pi to camera interface, image storing &
 Matthew Walczak – Ray tracing adjustments &
chromatic aberration mitigation
Final Design Review
 Filter wheel turning and stopping at each filter for
period of time, capturing image, and rotating to
the next filter
 Aberration mitigation for all filters
 Chromatic
 Spherical
 Transmitting captured image data from camera to
Raspberry Pi
 Identifying geological markers present in rocks and
Further Reading

similar documents