National Instruments NI myRIO, Essentials Manual

Page: 120 / 163

114 CHAPTER 16. AUTO-PAN CAMERA
plane (refer to Extract Luminance Plane
3
but select hue instead of luminance,
of course), dual-threshold the hue image with user-defined low and high values
to select the pixels that match the defined hue (see
Dual-Threshold Grayscale
to Binary
4
), remove noise pixels and other small objects (see Remove Small
Objects
5
), find the energy center (centroid) of the image (see Find Energy Center
(Centroid)
6
), and report the number of detected objects (see Report Area of
Particles
7
this step serves only to count the particles, therefore choose any single
measurement such as “Center of Mass X”).
Select the thresholding
Range
control and the
Centroid
and
Number
of Particles
indicators when you import this script into the Vision Assis-
tant Express VI. The hue thresholds will likely need adjustment occasionally as
lighting conditions change.
Camera Position Controller
The webcam is mounted to a servomotor as shown in Figure 16.3 on the facing
page; the servomotor pans the camera to any desired angle over a 180
◦
range.
The camera position controller can be completely implemented in the
(myRIO) update ports state. In this state the tracked object’s measured X-
position from the vision script is converted into a value between 0 and 100, and
the PID Advanced VI compares this measured position to the desired position to
generate a command signal to the servo via the PWM Express VI; the command
signal manipulates the servomotor to minimize the difference (error signal)
between the measured position and the desired position.
T
IP
: Use the “Quick Drop” shortcut (Ctrl + space bar) to look up and
place a LabVIEW element by name. See Appendix C on page 139 for more
LabVIEW tips.
Measured X -position
The vision script produces two outputs, the tracked object’s position and the
number of objects detected. Enclose the PID and PWM VIs and associated code
in a Boolean case structure so that these run
only
when a single tracked object
is detected. This way the camera will remain stationary when the tracked object
disappears and will also not be distracted by competing features.
3
http://youtu.be/1mmzsd46_Gs
(1:02)
4
http://youtu.be/0TFh4zB6e3Y
(1:56)
5
http://youtu.be/Xfo9wK_4B7Q
(1:09)
6
http://youtu.be/XkYfR141x0k
(0:57)
7
http://youtu.be/N8T3HDdlN-E
(1:44)