I have started a License Plate Recognition project, in which I intend to use IM for image processing.
I am very new to IM and a bit overwhelmed by the image processing richness that IM offers. So far my IM use extends little beyond manually iterating through rotating and croping operations to extract an AOI (Area of Interest) and then applying level, color and other color manip operators (basically to reduce the number of gray pixel and transform those pixels to either black or white while maintaining the alphabetic character features) to the extracted area so that tesseract can reliably translate the characters on the image to ASCII chars. A fundamental question: Is IM suitable for this kind of a project? I am looking for a shell, or better yet Perl or C++, algorithm that can search to for pre-specified quadrilateral object (using rel corner coords, and maybe some color map stats that could represent typical color characteristics of a license plate) that is randomly positionend on an gray-scale image. Also I am looking for an algorithm, that takes an AOI and then (possibly using different color manip approaches) converts dark areas into black and bright areas into white and possibly allows for some etch color conversions. Is it possible to get a list of x-coord, y-coord, color code for each pixel contained in an image using IM? (I know, depending on the image, the list can be VERY long, but if I had a list like that an algorithm could work through that list and manip the image so that it is useful for OCR). And lastly, I am looking for other developers that are interested in colaborating in this project. With kind regards Beat Naef _______________________________________________ Magick-developers mailing list Magick-developers@imagemagick.org http://studio.imagemagick.org/mailman/listinfo/magick-developers
