pvisionTCP3-nerdcam.C File Reference

#include "Beowulf/Beowulf.H"
#include "Component/ModelManager.H"
#include "Devices/BeoChip.H"
#include "Devices/DeviceOpts.H"
#include "Devices/FrameGrabberConfigurator.H"
#include "Image/ColorOps.H"
#include "Image/CutPaste.H"
#include "Image/DrawOps.H"
#include "Image/FilterOps.H"
#include "Image/Image.H"
#include "Image/ImageSet.H"
#include "Image/MathOps.H"
#include "Image/ShapeOps.H"
#include "Neuro/NeuroOpts.H"
#include "Neuro/SaccadeController.H"
#include "Neuro/SaccadeControllerConfigurator.H"
#include "Parallel/pvisionTCP-defs.H"
#include "Simulation/SimEventQueue.H"
#include "Simulation/SimEventQueueConfigurator.H"
#include "Transport/FrameIstream.H"
#include "Psycho/PsychoDisplay.H"
#include "Util/Assert.H"
#include "Util/Timer.H"
#include "Util/sformat.H"
#include "Video/RgbConversion.H"
#include <signal.h>
#include <unistd.h>
Include dependency graph for pvisionTCP3-nerdcam.C:

Go to the source code of this file.

Defines

#define SERVO_PAN   6
 servo to use for pan
#define SERVO_TILT   7
 servo to use for tilt
#define NAVG   20
 Number of frames over which average framerate is computed.
#define PRESCALE   2
 prescale level by which we downsize images before sending them off

Functions

void terminate (int s)
 Signal handler (e.g., for control-C).
void receiveCMAPS (nub::soft_ref< Beowulf > &beo, Image< float > *cmap, int32 *cmapframe)
 function to receive results from slaves
int main (const int argc, char **argv)

Variables

static bool goforever = true
 Will turn false on interrupt signal.

Detailed Description

Grab, process over beowulf, move camera

Definition in file pvisionTCP3-nerdcam.C.


Define Documentation

#define NAVG   20

Number of frames over which average framerate is computed.

Definition at line 88 of file pvisionTCP3-nerdcam.C.

#define PRESCALE   2

prescale level by which we downsize images before sending them off

Definition at line 91 of file pvisionTCP3-nerdcam.C.

#define SERVO_PAN   6

servo to use for pan

This parallel vision processing master is for use with pvisionTCP3. See the pvisionTCP3go script in bin/ for how to launch the slaves. The main difference between the TCP2 and TCP3 versions is that TCP3 uses load-balancing to send images to the slaves, and the master node is not only in charge of grabbing the video and displaying the results, but also of collecting the various conspicuity maps and assembling them into the saliency map. The difference between this and pvisionTCP3-master is that here we also move a pan/tilt camera towards the most salient location in the display.

Definition at line 82 of file pvisionTCP3-nerdcam.C.

#define SERVO_TILT   7

servo to use for tilt

Definition at line 85 of file pvisionTCP3-nerdcam.C.


Function Documentation

void receiveCMAPS ( nub::soft_ref< Beowulf > &  beo,
Image< float > *  cmap,
int32 cmapframe 
)

function to receive results from slaves

Definition at line 392 of file pvisionTCP3-nerdcam.C.

References TCPmessage::getElementFloatIma().

void terminate ( int  s  ) 

Signal handler (e.g., for control-C).

Definition at line 96 of file pvisionTCP3-nerdcam.C.

References goforever.


Variable Documentation

bool goforever = true [static]

Will turn false on interrupt signal.

Definition at line 93 of file pvisionTCP3-nerdcam.C.

Referenced by terminate().

Generated on Sun May 8 08:15:11 2011 for iLab Neuromorphic Vision Toolkit by  doxygen 1.6.3