Hello Giuseppe I never heard about Leptonica, I am going to read about it and see how it will help in my case. I went through your sample code, be honest, it didn't make a lot of sense, I think should read about leptonica first and then it will make sense. Just few question, as you said that you are using this to locate the box contour. So, can you explain what is *seed_sequence value are and what it represent (dimension of the box of medicine or what). I will read more about it and then I will ask you more, but once again thanks for giving me something that may work.
regards, Zia On Aug 2, 3:47 am, "Giuseppe Menga" <[email protected]> wrote: > Dear Zia, > I may give you a different idea for recovering the plate rectangle that > worked for me. > In my case I had to recover the box contour of a medicine in order to detect > the expiration date in it. > I'm using Leptonica for this: > static const char *seed_sequence = "o3.3 + r11 + o10.1 + c15.10 + x4"; > Pix* pix; // the input gray picture > Pix *pixm,*pixM; //the minima and maxima loci > //int w,h; > //pixGetDimensions(pix,&w,&h,NULL); > pixLocalExtrema(pix,0,0,&pixm,&pixM); // to get minima and maxima > Pix *pixMd2 = pixMorphSequence(pixM, seed_sequence, 0); // some cleaning > with opening and closing > float skewAngle,conf; > Pix* pixMdd = pixFindSkewAndDeskew(pixMd2,1,&skewAngle,&conf); // deskew > // find baselines > PTA* pta; > NUMA* numa = pixFindBaselines(pixMdd,&pta,0); // find the horizontal lines > /*numa is an array of ordinates, pta is an array of points (x,y) - the > extreme points of the horizontal lines found for each ordinate. */ > Extraxting with a simple logic the key four points ( or just three) of the > main rectangle I build the clipping box. > Among different techniques this was the one more reliable, as the search for > maxima is fairly insensitive to flares and reflections. > Let me know. > Giuseppe > > ----- Original Message ----- > From: "ZIA" <[email protected]> > To: "tesseract-ocr" <[email protected]> > Sent: Monday, August 02, 2010 2:13 AM > Subject: Re: California License Plate font issues with OCR > > Thanks Andre for finding the font. I will see how can i use that. As > you suggested using coreldraw, i don't have this software, i will try > to see if i can use some other software like MS word. > > I was asking how to extract license plate from image. What I am doing, > i get the image, re-sized, convert to binary image and then run the > sobel edge filter. So now i have an image that shows me the rectangle > part of LP clearly, and I know the ration of height to width is 1 to > 2. I just need to scan the image and look for rectangle and calculate > their ration, to get the correct LP from image, or is there any other > more efficient way. I hope this clear things, if not, I will send you > the image as input image and after edge filter image. > > Zia > > On Jul 31, 9:10 pm, Andres <[email protected]> wrote: > > > Anyway, any of you have any idea, about scanning image and getting the > > >> LP (image was filtered using edge filter, i can see the rectangle box > > >> of LP, just need to figure out, how to scan and how to extract. The > > >> ratio of CA LP is 1 to 2, or 6 to 12 inches (height=6, width=12) > > > >> I'm not sure about being understanding completelly. Could you extend a > > > little ? > > -- > You received this message because you are subscribed to the Google Groups > "tesseract-ocr" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group > athttp://groups.google.com/group/tesseract-ocr?hl=en. -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/tesseract-ocr?hl=en.
