Q: How can I improve the results of line removing, and what conditions affect this operation?
A: The Line remover function is an image processing tool that can be used to improve the quality of scanned/faxed papers. To be fully aware of how to optimize the line remover results, it’s important to know what exactly it is doing. The Line remover function is developed to remove noise lines from the input image. The line to be removed cannot be too thick or too short (these limiting values are computed from the DPI value of the image, so the DPI values must be set carefully, and the horizontal and vertical DPI values must be equal), and it must be nearly parallel with the upper/lower sides of the paper (nearly horizontal). If a line is too sheer, it will not be removed. The conditions with the most effect are the continuity of the noise lines, and the DPI value of the image. If the image scanning had poor quality and the continuity of the lines is not warranted, the line remover function can be ineffectual. The incontinuity of the scanned lines can be due to scanning DPI valuewhich is too low, or the original lines where too thin (like pencil drawings). If a poorly scanned paper is available with incontinual lines, morphological binary dilation or closing can be used to make the lines continual.
The noise lines are usually serrated. To remove saw-teeth, line remover operation removes not only the straight line, but the bumps too. If a noiseline crosses text it can be possible that line remover removes some parts of the text along with the line. If it does, reducing the image’s DPI values can be effective (Don’t rescan the image!!! The higher the scanning quality is, the better the line removing results will be, only reduce the set DPI values!).
After applying the line removing operation, the quality of the leaky/saw-tooth text also can be improved with morphological binary dilation or closing.