@phdthesis{digilib15751,
           month = {January},
           title = {DETEKSI TEPI CITRA KHAT ARAB MENGGUNAKAN
OPERATOR SOBEL DAN CANNY},
          school = {UIN SUNAN KALIJAGA YOGYAKARTA},
          author = {NIM. 10650058 AMIR FUAD ASY?ARI},
            year = {2015},
            note = {Pembimbing :  Bambang Sugiantoro, M.T,CompTIA},
        keywords = {Kata kunci : Arabic Khat, Canny Operator, Computer Vision, Edge Detection,
SNR, Sobel Operator, Timing run, visual.},
             url = {https://digilib.uin-suka.ac.id/id/eprint/15751/},
        abstract = {Khat is one of the branches of Islamic art that comes from the Arabic script.
There are various types of handwriting that can be enjoyed by the beauty of its form.
However, the lack knowledge of lay people about the khat types and forms of khat are
sometimes complicated to make the most people only get to enjoy the beauty of the
handwriting of shape without being able to read or know what the sentence contained
in the khat. One way that can be used to help the layman to recognize the type of khat is
with computer vision. In computer vision there are stages that play an important role in
recognizing the different character. These stages is image segmentation with edge
detection operation. Therefore, this study examines more about the use of edge
detection operators in the process of image segmentation Arabic handwriting. The
operator are used is operator sobel and canny.
The object of this research is the result of handwritten calligraphy image that
has been through the process of digitizing a scanner brands EPSON Stylus TX101 and
then stored in jpg format. The image of khat are 40 images consisting of 8 types of
handwriting and each type of khat are 5 pieces of the image. The steps are being taken
on this study are preprocessing, processing and analysis. And for the testing are done
twice with the testing parameters timing tun, SNR, and visual observation. The first test
is doing preprocessing, and for the second test is not doing preprocessing. It wasbecause
to test the performance of the operator when using preprocessing and without
preprocessing.
The results shows that the timing testing of non preprocessing run Sobel has the
fastest processing time with an average value of 0.4689 seconds / image. Sobel non
preprocessing also have better visual testing the value percentage of 65.25\%.
Meanwhilein the SNR testing, canny preprocessing has a highest average of 8.4901 dB
value. It can be concluded that the Sobel operator is better at timing test run and visual
testing. And canny preprocessing is better on tests SNR.}
}