Anti-counterfeit Identification Research In The Australian Dollars Based On Multi-spectral Images

Anti-counterfeiting technology based on multi-spectral images of the Australiandollars has been the hot spot in the field of paper currency anti-counterfeiting, whichrelates to the national financial security. The research has important theoretical researchvalue and a wide range of application background. Multi-spectral images of the Australiandollars contain the image information of visible light, infrared light and UV light. Throughanalyzing the multi-spectral images of the Australian dollars, we can reliably apart the truefrom the fake, and be able to ensure the safety and reliability of the Australian dollars incirculation by finding out the fake money that traditional anti-counterfeiting technologycan’t identify, as well as the money that are altered and not suitable for circulation.Domestic study of the Australian dollars made of plastics is still in its infancy,especially the Australian plastic money there are many different features compared withother paper currency, which makes that the traditional algorithms are not suitable for theAustralian dollars. The main problems of anti-counterfeiting algorithm based onmulti-spectral images of the Australian dollars will be faced as following: Firstly, themulti-spectral images of the Australian dollars have complex background, bright colorsand rich texture. Secondly, the Australian dollars in circulation have different degree of oldand new, wear, pollution, and so on and so forth. Thirdly, the plastic material andtransparent windows and other new type of anti-counterfeiting technologies are applied tothe Australian dollars. It raises very high demands during our image preprocessing,denomination and direction recognition, as well as authenticity identification.To solve these problems, firstly, we improve the edge detection algorithm throughanalyzing the anti-counterfeiting characteristics of the Australian dollars. Secondly, basedon the IR images of the Australian dollars, we proposed the denomination and directionrecognition algorithm in this thesis. The algorithm is combined with the Australian uniquetransparent window technology. By extracting this area as feature area, we finisheddenomination and direction recognition task of the Australian dollars. Thirdly, by applyingthe multi-area comparison strategy, we put forward the feature area selection algorithm,furthermore, weFor the characteristic of the complex texture in the Australian Dollars, we propose the algorithm based on the new features of the valid gray level information and the valid grayscale distribution information, as well as the texture structure based on the gray levelco-occurrence matrix. In order to improve the classification ability, by using thesupervised training method, we put forward the classifier algorithm for the AustralianDollars based on SVM. This algorithm solves the problems of weak anti-interferenceability, computing workload and low accuracy existing in the traditional algorithm.By comparing with other image algorithms, the experimental results show that theproposed algorithms can better adapt to the anti-counterfeiting characteristics of theAustralian dollars, recognition has significantly increased and the algorithms has the verygood adaptability to noise interference.