As Philippe says above, you will need to find the algorithm, and then pass it to your FPGA.
I have never heard of an open source OCR engine for specialized equipment, and I have been using OCR since 1997, when there were some European machines (CGK) with OCR in the hardware. Around the same time, they moved OCR back to software.
Here you can find something.
You also need to find an algorithm that matches the types of images you want to read. Some questions you need to ask:
- Are you reading a fixed size font?
- Are you reading a fixed pitch or proportional font?
- Do you process b / w or color images?
- What resolutions or DPI images do you work with?
- Do you need to remove background noise or color or fulfill a threshold value?
- Do you need images for correction?
- How many characters do you need? 0-9 or the whole alphabet?
- How fast should it be?
- How accurate should it be?
- Is the text in one place or in random order? If by chance, how do you find text areas?
If your needs are quite simple, you can get away from some if the idea is in a CodeProject link. Writing a good OCR commercial engine took many years of work, and they are still fine-tuned to increase accuracy and speed.
source share