Presentation: Sensors

Report
EEL5666
Robert Hartwell
12 Mar 2012
 1: Sensors
• Sensors tested
 2: Testing
Conditions
• Sensor response to conditions
 3: Platform
Integration
• Sensor integration on platform
 4:Additional Testing
• Additional conditions or aspects that require
additional testing.

Cell phone
• IP Webcam Server Application

Color Recognition - Matlab
• Colors correspond to menu items

Menu Card
• Color in the box next to the drink you want
% clear all
% clc
%
% % while(1)
% order = 0;
%
% format shortG
% prompt={'Vodka','Rum','OJ','Coke'};
% name='Drink Choices';
% lines=1;
% def={'1', '1', '1', '1'};
% options.Resize='on';
% answer=inputdlg(prompt,name,lines,def,options);
% vodka=str2num(char(answer(1)));
% rum=str2num(char(answer(2)));
% oj=str2num(char(answer(3)));
% coke=str2num(char(answer(4)));
%
% order = 1 % vodka+rum+oj+coke;
%
50
100
150
% % Locatate the IP Camera
url = 'http://192.168.1.4:8080/shot.jpg';
% ss = imread(url);
% fh = image(ss);
200
% This shows a continuous stream from the IP camera
% while(1)%while(order ~= 0)
% ss = imread(url);
% set(fh,'CData',ss);
% drawnow;
% end
250
%%
%% Find Green Object
% This script reads in an image file and then attempts to find a green
% object in the image. It is designed to find one green ball and highlight
% that ball on the original image
%% Read in Image
% First we read the specified image from the file and bring it into MATLAB
% as a variable. We also display the image to ensure it is correct.
% greenBall1 = imread('greenBall1.jpg');
greenBall1 = imread(url);
greenBall1 = uint8(greenBall1);
imagesc(greenBall1);
%% Extract each color
% Next we using indexing to extract three 2D matrices from the 3D image
% data corresponding to the red, green, and blue components of the image.
r = greenBall1(:,:,1);
g = greenBall1(:,:,2);
b = greenBall1(:,:,3);
%% Calculate Green
% Then we perform an arithmetic operation on the matrices as a whole to try
% to create one matrix that represents an intensity of green.
justGreen = 2*g - r - b;
colorsPlot(r,g,b,justGreen);
%%
close
%% Threshold the image
% Now we can set a threshold to separate the parts of the image that we
% consider to be green from the rest.
bw = justGreen > 95;
imagesc(bw);
colormap(gray);
%% Remove small groups
% We can use special functions provided by the Image Processing toolbox to
% quickly perform common image processing tasks. Here we are using
% BWAREAOPEN to remove groups of pixels less than 30.
ball1 = bwareaopen(bw,30);
imagesc(ball1);
%% Dilate
% We will use IMDILATE to dilate and combine nearby pixels
se1 = strel('square',5);
ball2 = imdilate(ball1,se1);
imagesc(ball2)
%% Find center
% Now we are using REGIONPROPS to extract the centroid of the group of
% pixels representing the ball.
s = regionprops(ball2, {'centroid','area'});
if isempty(s)
error('No ball found!');
else
[maxArea, id] = max([s.Area]);
hold on, plot(s(id).Centroid(1),s(id).Centroid(2),'wp','MarkerSize',20,'MarkerFaceColor','r'), hold off
disp(['Center location is (',num2str(s(id).Centroid(1),4),', ',num2str(s(id).Centroid(2),4),')'])
end
%% Show marker on original image
% Finally we will plot the center on the original image to clearly evaluate
% how well we have found the center.
imagesc(greenBall1);
hold on, plot(s(id).Centroid(1),s(id).Centroid(2),'wp','MarkerSize',20,'MarkerFaceColor','r'), hold off
300
50
100
150
200
250
300
350
400
450
EVO 3D
Ranging IR IR Array
Switch
General
Purpose
Color
Recognition
IR Distance
Sensor
IR Return
Physical
Location
Contact
within Array Detection
Radio
Comm
How it
Performs
Function
Creates
Server for IP
Cam
Emulation
IR Return
Strength
Alters
Voltage
Signal
IR Return
Allows
toggles
Voltage
High/Low in Signal
Array
RF receive
and transmit
What it
Controls
Order
Confirmation
Collision
Avoidance
Line
Following
Comm with
laptop
Docking
Radio
Frequency
Conditions tested in Matlab
• Hosting IP webcam server
annihilates the phone battery
Metric
Resolution
Quality
Equalization
Threshold
Range
(480x320)
(640x480)
1-100
NA
Color
Variable
Effect on
Large
Processing
time
Minimal
Slight
Negligible
Effect on
accuracy
Minimal
Large
Large
Large
Ideal
(480x320)
50
NA
TBD
Conditions tested with X-CTU
• Surprisingly small FoV
• FoV skewed left
• Ambient lighting had little effect
Metric
Lights Off
Lights On Blinds Open
Outside
Baseline
250-290
260-290
260-290
TBD
Δ 500
Distance
~17 in
~17 in
~17 in
TBD
Break 1200 ~ 12 in
distance
~12 in
~ 12 in
TBD
FoV @ 12
in
~ 3 in
~ 3 in
TBD
~3 in
Conditions tested with X-CTU
• Ambient lighting had little effect
• Very small “sweet spot”
Metric
Lights Off
Lights On
Blinds Open
Outside
Min Range
1/8 in
1/8 in
1/8 in
TBD
1 in
1 in
TBD
Max Range 1 in
Conditions tested with Voltmeter
and Matlab
Metric
XBee
Switch
Closes
Switch
NA
Yes
Tx/Rx
with
Matlab
Yes
NA
• Xbee fully function with Matlab
and AVR / Board
• Still working debounce routine for
switches
XBee
IR Ranger
IR Array
Bump Switch

similar documents