A Study Based On Image Recognition Of Old And New Banknotes Clearing System

Image recognition technology is an important means to improve the current banknote sorter sorting efficiency, the main achievement of the denomination of the bill,facing new and old crippled classification requires a high level of real-time and reliability. The requirements of the various commercial banks every day to give a large number of ATM machines with notes, according to the People’s Bank of China, "Currency notes reach seventy percent new condition" turned over to the storage of old banknotes only rely on manual operation is difficult to achieve. Traditional denoising algorithm for image denoising processing are often unable to smooth and protect the edge between balance and sometimes lead to loss of image information due to excessive noise suppression, resulting in a blurred image; or class traditional algorithm will noise suppression is not sufficient the residual excessive noise in the image denoising ineffective. In order to protect the edges of the image, such as singular information while de-noising, the researchers have proposed TV denoising algorithm, the algorithm can play a protective effect on the edges of the image denoising. However, traditional TV algorithm sometimes while retaining edge information remains in smooth regions of the image in more noise, the worst case even to generate false edges.How to accurately identify the RMB banknotes and how to improve the processing speed is the key. This paper focuses on these issues in-depth exploration and research,and to try to solve this problem, banknote sorter, not only can accurately identify the RMB old and new banknotes, and also reached the technical requirements for the speed of the identification. The core technology of this article is that the image pre-processing and feature selection. First, in order to accurately identify the image, the image must be pretreated, such as noise removal, find the center point, and calculating the inclination angle.To compensate for the disadvantages of the traditional TV algorithm while retaining the protective properties of the algorithm for edge, the paper the analysis spatial gradients of the pixels of the image and the pixel gradient information, improved traditional TV algorithm so that it can be carried out to protect the edge of the image on the basis effective denoising processing. First analysis of the gradient information of the image, and then applied to the analysis area gradient information threshold inhibition toovercome the impact of noise on the smooth area, so that you can make up for the traditional TV operators meal even more residual noise in smooth regions there will be the lack of false edges; after denoising pretreated, the algorithm also combines the image pixel gradient information to of iterated function TV algorithm that will be used to solving. Finally the experimental results show that the proposed algorithm can achieve better denoising effect while preserving the image edge, improve image PSNR value and visual effects.