1 | initial version |
getRotationMatrix2D, warpAffine
http://docs.opencv.org/doc/tutorials/imgproc/imgtrans/warp_affine/warp_affine.html
2 | No.2 Revision |
getRotationMatrix2D, warpAffine
http://docs.opencv.org/doc/tutorials/imgproc/imgtrans/warp_affine/warp_affine.html
P.S.:
I now understood that you first need the angle. I think the easiest solution would be to just use two different angles that have the distance of abut half a character. You then use both angles and applies your OCR first on both images. (Or first on the full image and then on the first part of the second). In your case, you could maybe rotate about 15 degrees. (Or you just select some random angles and later merge all detected characters)