Portable Ticket Automatic Scanning Technology Based on Lightweight Visual Recognition

In accounting and financial management, bills serve as vouchers for enterprises and government departments. The processing of bill information is fundamental to financial management with importance. We created a portable bill-scanning technique based on lightweight visual recognition. An image acquisition module was used to capture bill images, which are transmitted to a microprocessor. The microprocessor was used to grayscale the images using an improved Canny edge detection operator for bill edge scanning and correct the images using Hough transformation. Then, convolutional neural networks were utilized for financial data extraction, yielding high-quality financial data. The proposed bill-scanning method effectively reduced false alarm rates while maintaining high recall rates, addressing prevalent shortcomings in current bill-scanning technologies.