-1

I am newly trying to work with images. I am trying to develop a OCR for hand written images. By referring some books i started my work

Now I have a confusion which will be the best thresholding method. I know Local Adaptive method will be the best one. In that i want to know Canny,Bernsen’s method, or Stroke-Based in which method will be the best. I have algorithm for others but for last one I have nothing. 1.Double-edge detection. 2.Stroke based thersholding

Please help me in this...

karthikeyan
  • 61
  • 1
  • 4

1 Answers1

0

Thresholding is the main problem in image processing. In your project you want better threshold for edge detection. so i suggest method

  1. choose initial threshold example t=100
  2. find background image average pixel value Tback
  3. find foreground image average pixel value Tfore
  4. final threshold=(Tback+Tfore)/2
  5. repeat step 1 to step5 you will get best threshold value. where initial threshold value approximately equal to final threshold.

because when you choose the correct threshold value you will get the best edge detected image.

gnat
  • 6,213
  • 108
  • 53
  • 73
suresh gopal
  • 3,138
  • 6
  • 27
  • 58