-1

I have no idea how to do image recognition, but I have to extract the red-outlined portion from below image and read the graph, supposing the following:

1 - Off Duty
2 - Sleeper Berth
3 - Driving
4 - On Duty

My program should read the graph and, in this example, show a result of 333333333111111222222222444444

A driver's log with a graph indicating when they've been working or off duty

doppelgreener
  • 4,809
  • 10
  • 46
  • 63
Ronak Shah
  • 35
  • 7
  • 2
    So far what did you do ? Did you try any software library ? Did you put your on hands on any code that doesn't work for you ? Check out those links : http://stackoverflow.com/questions/152028/are-there-any-ok-image-recognition-libraries-for-net and http://stackoverflow.com/questions/9727579/c-sharp-image-recognition-library. [AForge](http://code.google.com/p/aforge/) seems to be a good start – Stephan Sep 17 '12 at 07:48
  • 1
    You drastically need to break this task down. It's certainly not suitable to encompass everything you need to achieve here into one SO question. You've got to interface with some sort of image scanning API, you then have the issue of OCR/coding a way of reading that particular part of the image every time accurately. And then you have to decide on which platform and how you're going to represent the data from the scan. –  Sep 17 '12 at 07:48
  • @Stephan Thanks for reply. yes I have used Idea of using Aforge Shapechecker function but there are many fields with rectangle Shape.How Can I extract that part only which i want. – Ronak Shah Sep 17 '12 at 08:41

1 Answers1

1

Assuming there are no alignment issues, you could divide the task as follows

  1. cut the red box rectangle from scan
  2. XOR the resulting image with the content of the same region of a blank document (this will attenuate the grid lines)
  3. scan in the X direction
  4. look for maximum in the Y direction
  5. map the found Y values to ranges you require
zeFrenchy
  • 6,541
  • 1
  • 27
  • 36