A Technique For Locating The Overlaid Uyghur Text Lines In Video Images

As the special geographical location and geopolitical environment of Xinjiang, the network has a large number of non-coding images contains a multi-ethnic text. Extracting sensitive words is very important for security of network and management of network information. Videos transmission is an important channel of transmission on the Internet. How to extract the Uyghur texts in videos is of great significance to maintain a good network environment.To solve the problem of detecting the Uyghur text in video sequences, this paper proposes a new method based on combining Harris corner detection and baseline characteristics of Uyghur text: Firstly, make the conversion to input color image into gray scale, then detect the Harris corners, and then determine the baseline of the probable text area according to the distribution of Harris corners, then carry on morphological expanding operations to the corners within the baseline zone. Finally, the locations of the target text regions are obtained. The experimental results of the tests on a relatively large amount of image sequences show that the practicability and the effectiveness of the proposed method are good enough compared to other similar algorithms. in the Xinjiang region, Detection of the Uighur region of text in the video image, plays an important role in video content analysis and content-based image retrieval. We can also use this technology in the Altaic Kazakh and Kirgiz text area detection.