Some help here guys, please.

When I try to run any script with the line:
opencv.addListener("publish", python.name, "input", OpenCVData().getClass()) 
 
I get these warnings:
 
21693 [opencv] WARN  class org.myrobotlab.framework.Service  - OpenCV.setState(OpenCV) :  NoSuchMethodException - attempting upcasting
21694 [opencv] WARN  class org.myrobotlab.framework.Service  - ouch! need to search through 173 methods
21703 [opencv] INFO  org.myrobotlab.opencv.VideoProcessor  - starting capture
21705 [opencv_videoProcessor] INFO  org.myrobotlab.opencv.VideoProcessor  - using com.googlecode.javacv.OpenCVFrameGrabber
 
and the script stops...

GroG

10 years 8 months ago

Hello Paulo !

Like your Avatar ! ;)

You don't get a video stream ?  Unfortunately, the warning that you listed would not result in your problem. Although it's nice you looked into the log.  

Just so you know, the warning is a reminder to me that the frameworks dynamic upcasting is not as optimized as it could be.  It has nothing to do with the lack of a video stream.

Can you send a No-Worky - http://myrobotlab.org/content/helpful-myrobotlab-tips-and-tricks-0#noWorky

This will allow me to see the whole log file.

GroG,

Tks for the rapid response (as always).

I sent you the "No-Worky".

In Python scripts,

When I use the CvPoint (like bellow), the listener doesn't work, but the script goes until the end.

opencv.addListener("publish", python.name, "input", CvPoint().getClass()) 
 
But when I use OpenCVData, the script stops there.
 
The Python Tracking/OpenCV script works fine and also the OpenCV service without Python.
 
I was testing in an Intel Atlon, but now I'm using my i5 home computer and the results are the same.
 

GroG

10 years 8 months ago

In reply to by pauloSouza

you can't construct an OpenCVData without a "name" .. e.g. OpenCVData("someName")

In the No-Worky log I can see your script - and in it, there is this line..

opencv.addListener("publish", python.name, "input", OpenCVData().getClass()) 

this is supposed to be telling opencv that there is a listener named (python.name) and every time it invokes the "publish" method it should return the resulting data to python.input.....

You don't want "publish" anyway for OpenCVData ...

try this 

opencv = Runtime.createAndStart("opencv","OpenCV")
opencv.addFilter("PyramidDown")
opencv.addFilter("FaceDetect")
opencv.setDisplayFilter("FaceDetect")
 
opencv.addListener("publishOpenCVData", python.name, "input") 
 
 
def input():
    data =  msg_opencv_publishOpenCVData.data[0]
    bb = data.getBoundingBoxArray()
    print 'opencv data ', bb
    for rect in bb:
      print "face found in rectangle ", rect.x, rect.y, rect.width, rect.height
 

it should give the following output

 

Hey GroG,

I am running MRL 1.0.2652 on Windows 7 64-bit OS. When I run the code, my face is detected and bound by a red square, but I'm unable to get the coordinates. 

I have used OpenCV with Python but am only 4 days old on MRL. Please help. I have already sent a 'no-worky'!

Thanks a lot for your help!