Analysis And Research Of Counterfeit Money Recognition Feature Based On Multi-spectral

In recent years, with the sustained, rapid and healthy development of China’s economy, the wealth of society is accumulated quickly, but the phenomenon of trafficking of counterfeit money also spreads widely. Counterfeit money flooding can interfere with the normal order of the circulation of money, undermine the principle of social credit,。It not only affects the stability of the economic life of the masses of the people, but also directly impedes the economic development of a country. China’s public security organs always see combating and preventing counterfeit crime as an important task. Do everything possible to "chase the source of digging dens, cracking down on gangs, cracking major cases", actively in the fight against and prevention work, and have achieved some success.Usually, when the document retrieval gets counterfeit money, he bases on the counterfeit variety of physical, chemical characteristics of the counterfeit money to classify. This new counterfeit and counterfeit notes in the library have been compared, classified, and thus to determine sources of counterfeit dens, so public security organs can master the distribution and circulation of counterfeit money. This provides a reliable basis for combating counterfeit crime. With the continuous expansion of the library of counterfeit money, the document retrieval needs to spend a lot of time and energy to match similar counterfeit money in the counterfeit money library.In order to solve the above problems, Kunming Shubo Lande Science And Technology Co., Ltd. and Yunnan Economic Investigation Corps joint research and development the string and system of sources of counterfeit money. In this system, the new counterfeit money is irradiated by multi-spectral. Then analyze the counterfeit money special image in the different spectral characteristics, and these characteristics are digitized, the counterfeit money is classified.The paper studies the image characteristics of counterfeit money in the infrared, ultraviolet, and white light. Counterfeiting and dens need to consider the cost of counterfeiting and fraud paper, ink is no uniform standard, so the consequences is that the same dens of counterfeit money have a similar image features; but there will be some difference image features in the counterfeit money from different dens. Daily under white light, that distinction is not apparent, but under infrared, ultraviolet spectra, some features are clearly revealed.Image feature of counterfeit money is analysed under different spectra. According to the type of image feature, the counterfeit image feature is digitized by the type of processing. In the infrared spectrum, counterfeit money only shows the information of gray value. Therefore, the method of the averaging gray value is used to analyze some special areas of counterfeit money. In the ultraviolet spectrum, the most prominent feature of counterfeit money is its fluorescence characteristics. Fluorescent region is segmented by Fuzzy C-Means clustering algorithm. And then count the number of spots in the divided regions. The number indicates to distinguish the source of the counterfeit money. Under the white light, due to the different amount of ink, counterfeit money from different dens in the same regions shows different colors. Find the cluster centers of such tone with HIS color model and fuzzy clustering. Classify counterfeit money basing on the value of the cluster center.Finally, this paper describes to the image matching principle and the similarity principle applied by the system. Then classify new counterfeit money, and discuss the classification results.At present, the system has been put into use, reduced the contrast range for document retrieval, and achieves target of saving time and improving the efficiency.