| Summary: | 碩士 === 臺北市立教育大學 === 數學資訊教育學系數學資訊教育教學碩士學位班 === 100 === With the advances in digital technology, digital image processing is getting more mature and technology which brings more convenience for human life is being used more broadly in our daily lives, such as face recognition, recognition of vehicle license plates, motion detection and so on.
In this study, hope through the “webcam” aimed at the Mandarin textbook of elementary school low grades, by image processing techniques to detect the input video skin color range. And by capturing the images in the textbooks and converting images to a binary image, the amount of processed data is reduced. By contour extraction and the design of the morphological mask, Text outside of the noise is filtered out. Then by means of the histogram, the phonetic symbols next to the text are filtered and using Tesseract-OCR done character recognition . Finally, the content of the textbook is output through MS SAPI.
Based on the experiment, the researcher found that clear images can be gotten by image capturing with textbook and camera at a distance of 35 to 40 cm and with the correct cutting rate of 89.9% and recognition rate of 85.6% about the Chinese text.
|
|---|
