Intelligent book positioning for library using RFID and book spine matching

In library, the management of books is very complication and timing costing. The location of books could be altered by librarian, students, teachers and any one around the library. Therefore, allocating a book is not an easy task in big library. Indoor positioning is an important technology to help storage management and customer services providing. RFID provides a good wireless platform to facilitate indoor positioning. However, duo to the small width of each book spine, adopting positioning based RFID alone is not enough to locate books in a library. In this work, we combine image matching with L-GEM based RBFNN to enhance the accuracy and robustness of the book locating system. We apply this new method in a library environment to position the certain books. Experimental results show that the proposed method is highly accurate and robust to white noise of RFID signals.