This paper presents the detailed implementation of computer vision, image processing, and barcode decoding techniques in libraries for more efficient tracking of borrowed books. The main idea of the system relies on implementing image processing, machine learning, and a database that contributes to the system's ability to identify the barcodes of the books and log in their titles, editions, authors, borrower IDs, dates, and times.
The experimental results show that the proposed system is accurate, fast, reliable, and capable of recording borrowed books, where the accuracy of the system when capturing the barcode and performing the process of self-borrowing and/or self-returning for one book, or two books or three books reached 100%. The accuracy of capturing the barcode and performing the process of self-borrowing and/or self-returning a book in upright or horizontal positions was 100%. The accuracy of capturing the barcode and performing the process of self-borrowing and/or self-returning books in highlight, or medium light was 100%. The system was unable to capture the barcode and carry out the process of self-borrowing and/or self-returning books in low light without flashing the mobile phone at a rate of 100%. The accuracy of capturing the barcode and performing the process of self-borrowing /self-returning books in low light using the mobile flash was 100%. The accuracy of capturing colored barcodes in the dark by relying on the mobile flash and performing a self-borrowing and/or self-returning process was 100%.