I am currently working on a BSc thesis concerning 2D barcodes. We decided to work with Zxing due to its huge community and its performances.
I am having issues when trying to recognize a QR Code among many others (think about a piece of paper with a grid with 10x10 QR Codes). Sometimes it works and it recognizes, but most of the times I get one of these errors:
com.google.zxing.ChecksumException com.google.zxing.FormatException
What I believe is that it finds a possible QR Code (perhaps out of focus, or cut out) and tries to decode it, if it fails, the whole image gets thrown away.
I would like to find an efficient way to tell the program to keep looking throughout the whole image for a valid QR Code. For me it is important to recognize even just 1 code in the whole set.
Things I did:
- Kept a proper Quiet zone between the codes (I chose 3 instead of 4, since I need to have, in an area, as many QR codes as possible)
- Used hint 'TRY.HARDER'
The image is from a webcam and it is 640x480px. The quality is fairly decent.
An idea of the pattern:
Even if most of them are corrupted or badly exposed but even just 1 barcode is readable, I would like to be able to find it. I don't mind to lose a bit of speed (at the moment the decoding speed, if the barcode is found, is around 1-2ms).
Any suggestion?