how to remove object that contain more than X% black pixels

3 visualizzazioni (ultimi 30 giorni)
i have some roi and i want to remove roi that has more then 90% black pixels how can i do that???

Risposte (1)

Image Analyst
Image Analyst il 26 Dic 2014
Modificato: Image Analyst il 26 Dic 2014
Is your ROI in the form of a binary image? Are there multiple regions in that binary image? Just call regionprops and find blobs where the (FilledArea - Area) / FilledArea > 0.1.
labeledImage = bwlabel(binaryImage);
measurements = regionprops(labeledImage, 'FilledArea', 'Area');
allAreas = [measurements.Area];
allFilledAreas = [measurements.FilledArea];
darkRatios = (allFilledAreas - allAreas) / allFilledAreas;
darkIndexes = find(darkRatios > 0.1);
newLabeledImage = ismember(labeledImage, darkIndexes);
% Relabel
labeledImage = bwlabel(newLabeledImage > 0);
% Measure again with new blobs.
measurements = regionprops(labeledImage, 'FilledArea', 'Area');
  3 Commenti
Lukasz Jarod
Lukasz Jarod il 27 Dic 2014
Modificato: Lukasz Jarod il 27 Dic 2014
well i dont understand what you doing in this code;/
http://i58.tinypic.com/amze38.jpg for example i have this image then for each labeled componetnt i do template matching but first i remove some of the object. and now i want to remove somehow boject on the left(as you can see object is almost 90% filld with black pixels and there is no letters that are 90% filled with black) maybe you could tell me step by step what hapens in your code , or it will not work for me now? or maybe there is another way to do it?
Image Analyst
Image Analyst il 27 Dic 2014
It's only 9 lines of code, that do what you asked for.
If you really wanted OCR instead, then you should have said so initially and posted your image.

Accedi per commentare.

Categorie

Scopri di più su Convert Image Type in Help Center e File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by