[gmkumar2005]

Hi all
I want to perform edge detection on the video produced by viewfinder.
how to use PyNcvLib for this purpose?
The problem with PyNcvLib is it expects images from a file. I am not able to find a way to convert the pys60 Image object into CNokiaCVImage.
The pycamera has a callback method which passes image as a parameter. I want to use PyNcvLib to perform edge detection inside this call back.
Is there are any other library avaliable which can perform edge detection in realtime

Please help
Kiran

[brunovianna]

hi Kiran

I'm trying to do more or less the same. I managed to do the conversion with the code below. However, after a few frames, the application freezes. There seems to me some memory leak. Let know if you can get around this issue....

import e32, camera, appuifw, key_codes, graphics, PyMelib, random, PyNcvLib

from graphics import Image





def quit():

    global app_lock

    camera.stop_finder()

    camera.release()

    app_lock.signal()



def finder_cb(im):

    global sourceImg, sourceImgInst, imgOps, canvas, targetImg, targetImgInst

    

    finderBmp = im._bitmapapi()

    sourceImg.Bitmap_CreateL(finderBmp)

    

    imgOps.CornerDetect(sourceImgInst, targetImgInst, 20, 25)

    

    altBmp = targetImg.Bitmap()

    altOut = Image.from_cfbsbitmap(altBmp)

        

    canvas.blit(altOut)





sourceImg = PyNcvLib.PyCNokiaCVImage()

sourceImgInst = sourceImg.NewLC()



targetImg = PyNcvLib.PyCNokiaCVImage()

targetImgInst = targetImg.NewLC()

    



imgOps = PyNcvLib.PyCImageOperations()



canvas = appuifw.Canvas()

appuifw.app.body = canvas



appuifw.app.title = u"edge"

appuifw.app.exit_key_handler = quit



canvas = appuifw.Canvas()

appuifw.app.body = canvas



camera.start_finder(finder_cb, size=(240,180))



app_lock = e32.Ao_lock()

app_lock.wait()