0

I have a collection of signal images with signal as a black line, grid of red lines in the background, and black letters and numbers that are of no importance to me.

I need to extract the black curvy signal line as a time series, and completely ignore the letters and numbers. Something like the graph below, but printed and scanned, therefore not so crisp.

Ideally, the output should be a CSV of the black line.

...26.931, 24.740, 25.806, 24.364, 24.477, 23.901, 23.175, 23.227...

Is there an off the shelf method? what is the simplest/easiest way?enter image description here

Rosa
  • 155
  • 10
  • if it is image then crop image to keep only pixels inside black rectangle (without rectangle) and then only plot has black pixels which you can find by checking all pixels in image using PIL/pillow or converting image to numpy array. But plot has not only points with internesting data but also black lines between them which may not be needed but it hard to say which pixels to skip. – furas Sep 07 '19 at 01:13
  • 1
    Thanks, but I can't crop every image manually, since I have a few thousands images. – Rosa Sep 07 '19 at 02:53

0 Answers0