The Implementation Regarding The Location And Recognition Of The Words In Scene Images Based On Embedded Platform

With the rapid development of Internet, multimedia technology makes our life various. There is a widespread concern over the images that possess many pieces of information. An increasing number of people intend to study the images which contain the essential information. The images such as scene texts are so popular in our life.The paper will illustrate the scene text, the location of words in scene texts and the identity of words in scene texts.As far as the scene text requirement is concerned, V4L2, OpenCV and the transplant of programs are used.In the module of scene text location, Edge detection method is used firstly to identify the rough location of Chinese domain regarding scene text. According to the results of gray detection, the rough location and the results of gray detection will be integrated. Accordingly, the noise can be deleted and the precise location of text field can be identified; in the module of the identity regarding words in scene texts, text blocks should be preprocessed firstly. In addition, training sets should be trained.In the process of the training, gradient feature extraction and dimensionality reduction of training sets should be done and then the training sets should be uploaded. MQDF should be used to identify text blocks and the information should be presented in the interface. Finally, TTS can be used to read the texts.The system caters to the general scene requirement. It will process general scene texts such as the curbside brands, traffic indicator boards, the banners of propaganda, number plates and so on.(1) To enrich the life of the dim-sighted people, speech recognition system is needed. It can recognize the words of books or newspapers and then tell the people the specific content.(2) The recognition of highway sign is to help the exhausted drivers.(3) It is convenient for travelers to visit splendid scene by the speech recognition system. The system tells travelers something they need by the recognition of sign in big scenic spots.(4) It is beneficial for electronic management, automatic driving and prevention of dangerous conditions.