Affiliations: School of Forensic Science and Technology, Criminal Investigation Police University of China, Shenyang, Liaoning 110854, China | E-mail: lizhen_wjjyjsx@126.com
Correspondence:
[*]
Corresponding author: School of Forensic Science and Technology, Criminal Investigation Police University of China, Shenyang, Liaoning 110854, China. E-mail: lizhen_wjjyjsx@126.com.
Abstract: With the rapid development of network technology and information technology, the amount of information contained in images has increased significantly. How to effectively extract text information from complex images has become the focus of current research in this field. Firstly, the Canny algorithm in the edge detection algorithm is improved and applied to the edge detection of complex images. Then the K-means algorithm is optimized to achieve better clustering effect of pixels. Finally, the text information in the image is extracted from the two. The results show that under the influence of salt and pepper noise from 0% to 90%, the quality factor obtained by the improved Canny algorithm is at least 0.4, and the detection accuracy is higher; The minimum peak signal-to-noise ratio of the algorithm is 38, and the maximum mean square error is 30, which are both better than the LOG algorithm and the traditional Canny algorithm, and have better noise reduction effect and image fidelity. It is used together in the extraction process of image text information, and the text recognition accuracy rate of the combined algorithm reaches a maximum of 93%, and is stable at more than 90%, indicating that this method has a high text recognition accuracy rate and provides text extraction for complex images. A reference path is available.
Keywords: Edge detection, canny algorithm, text extraction, K-means algorithm
