Optical Character Recognition of Postmark Date Based on Machine Vision
暂无分享,去创建一个
For the purpose of information management on postmark according to the date, the paper put forward a method of postmark date recognition based on machine vision, which could meet the demands of personal postmark collectors. On the basis of the relative theories of machine vision, image processing and pattern recognition, the overall process is introduced in the paper from postmark image acquisition to date recognition. Firstly, threshold method is used to generate binary image from smoothed postmark image. So region of date numbers could be extracted from binary image according to different region features. Then regions of date numbers which are connected or broken could be processed through mathematical morphology of binary image. Individual regions of date numbers are obtained for recognition. Finally, classification and pattern recognition based on support vector machine make date numbers classified and date recognition is implemented correctly
[1] Bin Zhao,et al. Support Vector Machine and its Application in Handwritten Numeral Recognition , 2000, ICPR.
[2] Markus Ulrich,et al. Machine Vision Algorithms and Applications , 2007 .