基于高斯模型的污损纸币识别算法研究

Researchon Dirty Banknotes Identification Algorithms Based on Gaussian Model

研究纸币识别问题,提高纸币识别的准确率。针对纸币识别过程中,当待识别的纸币在流通中存在被污染或者磨损,传统的模板匹配的识别算法受纸币污损的影响识别的准确性。为解决上述问题提出一种高斯模型的识别算法,首先对待检测图像进行亮度补偿、边缘检测、倾斜校正等预处理,然后将图像划分为若干个矩形子区域,计算各子区域的灰度平均值作为提取的图像初始特征,通过计算初始特征的先验概率并对后验概率进行修正,对污损区域特征值的校正,最后建立高斯模型完成纸币的识别,克服了传统无法准确识别污损纸币的问题。实验证明,改进方法能够将纸币污损部分校正并将纸币准确识别,取得了满意的效果。

Research notes identification algorithms for improving the accuracy of banknotes recognition.In the process of notes recognition,when the banknotes to be identified were polluted or abrasion in circulation,the traditional matching identification algorithms based on template can not work accurately because of the influence of banknotes fouling.This paper proposed a recognition algorithm based on gaussian model.First,the pretreatments such as detection image edge detection,tilt correction and brightness compensation were carred out.Then,the image was divided into a number of rectangular promoter regions.Each branch area was calculated as the extracted image gray-scale average initial characteristic,and the posterior probability was amended through the initial feature prior probability calculation,so that realize the regional characteristic value of correction fouling.Finally,gaussian model was establish for the identification of complete notes,which overcome the problem that traditional methods cannot accurately identify dirty banknotes.Experiments show that the method can accurately identify fouling and achieve satisfactory results.