Design And Implementation Of FPGA-based Printed Digital Recognition System

Character recognition is an important branch of pattern recognition technology.Most of today's character recognition technology is implemented based on computer software,while character recognition technology based on hardware platforms is relatively rare.With the rapid development of the integrated circuit industry,digital image processing technology has made great progress on hardware platforms such as digital signal processing(DSP)and field programmable gate array(FPGA).This article aims to take advantage of the fast parallel processing speed of FPGA in the field of image processing,design and implement a printed digital recognition system completely built on the FPGA platform.In order to complete the expected digital recognition function,after summarizing three different digital recognition algorithms,this article decided to choose a digital recognition algorithm based on feature statistics and feature matching based on the feasibility of the three algorithms on the FPGA hardware platform.This system uses a hardware description language to describe each module of the digital recognition system,and the three modules of the digital recognition system's image acquisition,image storage control,and image signal processing are designed in detail.During the construction of the system,Altera's Cyclone IV series chip is used as the core device of the character recognition system,the CMOS camera is used to collect image data in real time,and the 32 MB synchronous dynamic random-access memory(Synchronous Dynamic RandomAccess Memory)chip is used for storage.For video image data,the processed image is displayed on the LCD screen,and the result of digital recognition is displayed on the sevensegment digital tube.The algorithms used in the image preprocessing of this system specifically include grayscale conversion,median filtering,image binarization,and projection segmentation.When the system is designing each module,the image processing module design uses the pipeline design idea,and the image storage control module uses the ping-pong operation design idea.This paper uses Mentor's Modesim simulation tool to simulate the digital recognition function of the system,and download the project file to the FPGA development board for boardlevel debugging.The verification results show that the FPGA-based digital recognition system has reached the design expectations and fulfilled the functional requirements of printed digital recognition.Compared with the traditional digital recognition system based on computer software,the functional requirements of body digital recognition show that this system has good recognition accuracy,significantly improved processing speed,and reasonable system resource consumption.