Ask Your Question
0

Rotating Dial Number recognition Approach

asked 2016-07-19 17:23:52 -0500

Hello openCV community :)

I am new to image processing & openCV. I am trying to isolate and regonize the numbers in rotating dial of a water meter using openCV. So far i am at the very beginning of the process. Below is the image of a sample water meter i am testing out.

image description

For isolating the number dial i have tried the following:

  • Used ORB feature detector to detect the location of the dial. ORB detects the dial when i use the a similar image of a dial with numbers but it wont work for all numbers

Other alternatives that i am going to try will involve,

  • Estimating the dials location by the Red color ( I am unsure about a method that will let me extract the whole dial using this technique)
  • Look into if I can use something like a neural network to recognize the dial
  • Look into if I can use the feature of white and red on the dial to detect and extract the dial. (Would that be possible?)

What would be the most effective method for this? What kind of pre-processing should i do for the image?

Thank you in advance :)

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
1

answered 2016-07-20 04:53:54 -0500

Some things that I would adapt to your system

  • Don't use keypoint matching for the meter part. I guess that will have a small accuracy over different meters.
  • Like suggested by you, focus on distinct visual properties of the meter.

That being said try a colour base approach that does

  • Translate image to HSV colour space (easier to seperate colors)
  • Retrieve the red regions only.
  • Find the largest blob after some erosion, dilation to clean up noise
  • Find the bounding rectangle of that blob
  • Now double the width from [0, width] to [-width, width], maintaining the height, because that is again a specific property of the meter that can be abused

That will get you pretty far!

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2016-07-19 17:23:52 -0500

Seen: 211 times

Last updated: Jul 20 '16