Research On Key Techniques Of Uyghur-chinese Translation Based On Camera

Camera translation software is a convenient and fast translation tool, it has importantvalue for language learning and cultural exchange. Currently, the study of English–Chinesecamera translation software is relatively mature, and it has been widely used in smart phonesand other mobile carriers. In Xinjiang, Uighur is widely used, but the study of Uighur–Chinese camera translation is of immature. The study of Uighur–Chinese camera translationtechnology has a positive role in promoting Uighur–Chinese cultural exchange and Uighurinformatization development.In this paper, the study was focused on software section of Uighur–Chinese cameratranslation system. The pictures we processed were shoot by Camera. The specific job wecompleted is as follows:Word segmentation. In this paper, the recognition method we used need us to cut theUighur word into characters first. The pictures we shot should be pretreated,such asbinarization, filtering and tilt correction. Then we extract the target word and complete Wordsegmentation. In the study of word segmentation, against the feature of the Uighur wordspicture taking by camera, we proposed a new algorithm based on adaptive threshold. Byexperimental verification, segmentation accuracy rate of this method reached more than96%.Word recognition. Uighur character classification recognition algorithm based ontemplate matching was used to process the characters we get from word segmentation. Weextract the character feature to classify the characters first, and then we calculate both thesimilarity of the original image and the similarity of the projected image between thecharacters to be recognised and templates. Through similarity comprehensive judgment, weget the result of recognition. recognition accuracy rate of this method reached more than92%.After this part, we get codes as the result of recognition.Translation system. Word translation is done by using MATLAB software to callACCESS dictionary database information. Through programming, we establish a connection between ACCESS database and MATLAB software, and by using code sequence for the indexto the query word we achieved translation function. Finally, we used the MATLAB softwareto make the GUI interface. All aspects of the system were connected together. We can use theGUI interface to show the system.