Bar-Code Searching and Identifying System Based on Shape and Texture

• Main Authors: Po-Chun Lin, 林柏君
• Other Authors: Shih-Yen Huang
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/65968128519121790205

碩士 === 國立勤益科技大學 === 電子工程系 === 99 === As the application of barcode is extremely wide, which pushes forward the relevant study of barcode locating; where, the traditional study is focused on the texture and shape. The texture based is including: Gabor Transform [1], Directional gradient[2] and Hoff Transform[3]. Gabor Transform is to adopt convolution to take the energy of texture, which primarily takes the repetition property in barcode area that the attribute of vectors were taken .The directional gradient and Hoff Transform based on directional gradient in different angle. The deficiency of this method is that when angle gradient divided in more detail, the computing cost is more either. Reversely, if angle divides roughly, the target with the angle out of the algorithm shall be missed.　　　 Besides, the shape based method is to retrieve the shape characteristics as the base of positioning the barcode. Briefly, the algorithm based on texture attribute get higher accuracy but higher computing cost, by contrast, based on shape attribute the computing cost is less but with lower accuracy. [4].　　　 Since the barcode locating is a primary process to decode barcode automatically. This is proposed a barcode locating method that adopts shape attribute in initializing the suspected target so as to reduce the computing cost, then follows the texture attribute in confirming and then the barcode can be located. Finally, this thesis defined a scan line which went across the located barcode from the left margin to the right margin, and then decoded it according to the pixel value of this scan line. This thesis applied this proposed barcode locating method to 200 templates can locate 99.5% correctly. Obviously, the proposed barcode locating method can locate barcode with high accuracy.