|
![本文二维码信息](/download_upload_file.aspx?file_name=/uploadfile/two_dimensional_code_img/zgzzen/2021/6/202106010.png) 二维码(扫一下试试看!) |
Research on Acquisition and Recognition of Image on Encrypted Paper Made of Magnetic Fibers |
Received:January 11, 2021 |
DOI:10.11980/j.issn.0254-508X.2021.06.010 |
Key Words:encrypted paper magnetic fiber image collection image recognition |
Fund Project:陕西省榆林市2020年科技计划项目(CXY-2020-090)。 |
|
Hits: 3202 |
Download times: 1831 |
Abstract:A acquisition and recognition method of the image on encrypted paper made of magnetic fiber was proposed aiming at the problem that the encrypted pattern on encrypted paper made of magnetic fibers cannot be recognized by naked eyes. The encrypted image acquisition device first performed strong magnetic treatment on the magnetic-fiber encrypted paper, then sprayed magnetic powder onto the magnetic-fiber encrypted paper to make the encrypted pattern appear,finally a CCD camera was used to obtain the image on the encrypted paper. A method combining an improved machine vision algorithm and Convolutional Recurrent Neural Networks(CRNN)was proposed based on the original images to construct a recognition model for encrypted paper pattern. The detection and segmentation of the encrypted pattern were realized through the combination of Gaussian filtering, Sobel edge detection operator, improved OTSU, etc., the segmented images were then input into the CRNN network to complete the feature extraction and recognition of the magnetic-fiber encrypted image. Experimental results showed that the accuracy of the proposed model reached 98.37%, indicating it could solve the problem of recognition of encrypted pattern on encrypted paper made of magnetic fiber. |
View Full Text HTML View/Add Comment Download reader |