Applying Digital Image Processing Technologies To Books' Counting Of Library

In this thesis, we attempt to apply the digit image processing technology to the job counting books of a library. Our method is using the digital camera to capture the image of a row of book on the shelf, then extract the information about numbers of books from the image by the image processing algorithms. Moreover, we want to recognize which books the image includes.the image process is divided into four phases, pre-processing, image segmentation, Character segmentation ,Character recognition. In the pre-processing phrase, we convert the RGB image to gray image and enhance the contrast of image to facilitate the next processing. In the image segmentation phrase, we first use the edge detection algorithm to handle the processed image, then use line-scan to gain the edge of books which we can use for image segmentation. By doing this ,we get a lot of small image that only include one book. At this point , we get the number information of books too.In order to recognize book, we must obtain the enough features to distinguish books. One way we can adopt is character recognition. First we need segment the character of spine of book, the sole character image is benefit to our recognition. When we finish character recognition, we get the enough features too. we induct eigenvector and propose the definition and extraction methods of feature of letter and number, and both tree classification and min-distance classification are investigated to the recognition process.