Aiming at problems such as the image obtained by contact image sensor（CIS） is not as clear as the scanned images

uniformly illumination and defocused imaging and other issues in the processing of series recognition on paper currency.So how to extract useful information from the low-quality images is the key step for the automatic character recognition.A solution is presented to use the homomorphic filtering to remove effect of uneven illumination and then the improved Bernsen algorithm is used for image binarization.The experimental results show the effectiveness of this algorithm

the algorithm can effectively overcome the impact of the uneven illumination and it has good binarization effect.