Reputation: 1352
I want to split my image into multiple region, where a region is equal to a letter.
My example image is https://myinvoice.be.carrefour.eu/sites/all/themes/custom/myinvoice/images/ticket.jpg
The goal is to detect each letter individually. For example it would detect first the zone of C, then A, R, R and so on.
Any idea of a good algorithm / technique I could use to do that ?
I tought about ranging the pixel array, and when I found a pixel to follow the shape of pixels until I arrive at the beginning point. Save the min/max coordinate and then go to the next one by doing x + maxSHape or something like that but I am afraid it will take a very long time to analyse.
Upvotes: 1
Views: 1994
Reputation: 1032
If the image is that clear. Then that is easy.
1. Thresholding image to binary image (Otsu thresholding will work well)
2. Add some Deliation, Erosion to remove noise
3. Find contour from entire image
4. Add some are filter to select the specific font-characters
You can use opencv and python to implement those task.
Upvotes: 1