Implementing OCR on FPGA

Character recognition systems have contributed tremendously to the automation of data entry process. Many systems are present in software for character recognition. Digital signal processing has traditionally been done using enhanced microprocessors. While the high volume of generic product provides a low cost solution, its performance fall seriously short for many applications. Until recently, the only alternatives were to develop custom hardware (typically board level or ASIC), buy expensive fixed function processors (e.g. an FFT chip), or use an array of microprocessors. FPGAs offer an opportunity to accelerate your digital signal processing application up to 1000 times over a traditional DSP microprocessor. The main goal here is designing an optical character recognition (OCR) system in FPGA and to develop algorithms which are suitable for hardware implementation of OCR. The application proposed here is automatic calibration of measuring instrument based on OCR system in FPGA. This solution benefits from significant reducing of work quantity for calibration as well as from better replication and objectivity of measuring. The system entirely eliminates influence of the operator on the uncertainty calibration. As system is using FPGA there are many issues to be overcome like speed of operation, programmability and power consumption.