I use tesseract ocr to recognize the text on the image. I have a problem with images that have curly lines. I tried various method such as threshold, gaussian filter, extract by color.. but I cannot remove it. I want to delete the lines without loosing the numbers
This is the image using the threshold method
I’m using opencv for image processing and tesseract 4.0 for recoginize the text
Any hint or some sense of direction will be much appreciated. Thanks in advance for your help.
I tried multiple approaches. The following one is the closest I could get to.
Simple Algorithm:
This is just a starter. You can get far better result if you incorporate adaptive threshold techniques and morphological operations.
(I have the code available in case you need it)