Bank Bill Recognition Based on an Image Processing

The paper applies image segmentation and recognition theory, one novel method is proposed for the bill searches in bank system. In general, bank bill of Chinese is color image for encryption techniques, and right-up angle in the bill is key information which is composed of 0~9 number and 26 English crisscross-rows and which is of unique encode for each bill. The rectangular region of right-up angle on the bill is segmented, color image filter algorithms is designed in the light of the computing formula of the intensity of color image. The color background of the bill is filtered in the rectangular region segmented, the key information is only contained. Thus, the color bill image is changed into a grey image with filter effect. Through binarization, image enhancement, image sharp, and image filter, any bill in database can be extracted by Pattern Matching theorem exactly and soon. The technique obtains good result in a practice example.