Some Studies On The Compression And Anti-counterfeiting For Bill Images

As the amount of bills grows rapidly, processing bills automatically is a very important topic. Then bill images become the major information carries. With the promotion of the image collection technology, data contained by bill images is massive. Huge amount of information put forward high demand to storage capacity and channel bandwidth. Bill image compression becomes a must for the bills management. The research about the image compression has decades of history, but there is less about bill images compression.In the process of bills application, anti counterfeit is also plagued the industry. With the rise of the bar code technology, bar code is widely used in the anti counterfeit fields. But in practical, the identification of bar code becomes difficultly under the influence of complex environment. There have been many literatures about the identification of bar code. Most of them are about the identification of single bar code, but few of them are about the identification of multiple bar codes.In this paper, bill images compression and the identification of QR Code are made a deep study. Reference to the existing achievements, this paper puts forward some improvement methods, mainly including the following aspects:In the research about bill images compression, a method is proposed which is based on Hough transform and mathematical morphology. Through skew correction and redundancy removed, the amount of image data is reduced. Hough transform is used to detect the skew. To serve the shortcomings of traditional Hough transform, the algorithm removes the redundant details by morphological operations. This can achieve the purpose of improving the calculation speed and the detection accuracy. To remove the redundancy, extracting region of interested is used. The algorithm gets the vertex coordinates of the region of interested through scanning. Experiment results show that this algorithm is not only applicable to binary images and gray images, but also color images.In the research about QR Code identification, an identification algorithm for single QR Code is proposed, first. In the aspects of single QR Code identification, related documents are not mention how to take advantage of the position relationship between the position detection graphics to correct the skew. This problem is resolved in our algorithm. At the same time, the algorithm simplifies the extracting operation by utilizing the width, the height and the central points’coordinates of the position detection graphics.Then, an identification algorithm for multiple QR Codes is proposed. The algorithm uses the recognition of the position detection graphics to locate the QR Codes. The steps include that how to judge the pixels meeting the conditions whether belong to the same position detection graphic, how to remove the pixels not belonging to the position detection graphics, and how to confirm the coordinates of the position detection graphics’central pixels. Position detection graphics need to be judged whether belong to the same QR Code, according to the identification of multiple QR Codes. It is realized by utilizing the size and the position relationship between the position detection graphics. Experiment results show that the algorithm is effective.