remove streaks from a image
Mostrar comentarios más antiguos
Hi Guys !!!!!!!!!!
I have some images (grayscale) in which some text is written but with same color (which can easily get converted into black while binarizing the image) lines are drawn from one end to other . lines are continous and three or two pixel wide . some time line also intersect the text......
i have tried using imclose and strel, my lines diappear but my text also gets erased and everything becomes white.... i want to remove the line without affecting the text ...... Plz suggest methods !!!!!!
thanks a lot....
Respuestas (1)
Image Analyst
el 17 de Feb. de 2012
0 votos
rajashekar, where did you upload your image? Can you post it to tinypic.com? People almost always like to have images to look at when giving advice on image processing - weird, huh?
Morphological closing always enlarges an image, unless your background (the paper) is the foreground (white) and your writing is actually the background (black). In that case closing will make the dark lines smaller and broken. I suggest you reverse the sign in your thresholding to make the text true (1 or white) and the background false (0 or black). If you insist on doing it in reverse like you are now, then try imopen() instead of imclose().
5 comentarios
rajashekar
el 18 de Feb. de 2012
rajashekar
el 18 de Feb. de 2012
Image Analyst
el 18 de Feb. de 2012
Are you trying to restore a corrupted image? Which one is the input image? Or did you even post it? What does "I got this image" mean? Does that mean that is the input image that you were supplied with? Or does that mean that is what you got after you tried some operations on the input image? Try to envision the situation from my viewpoint - someone who has no prior background knowledge or context of your situation. Now try to explain as clearly as you can. Tell me what image you are starting with and what is the desired output image. Closing may or may not be the best approach.
rajashekar
el 18 de Feb. de 2012
Image Analyst
el 18 de Feb. de 2012
That's not bad. Looks pretty good to me considering how badly corrupted your input image was. Maybe your expectations are unreasonably high for an image that is corrupted so severely.
Categorías
Más información sobre Image Processing Toolbox en Centro de ayuda y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!