Look a little closer
A scratched or dirty QR Code does not survive simply because the same picture is stored several times. Its data includes mathematically calculated error-correction information, allowing a reader to infer some missing or corrupted pieces from what remains.
A QR Code converts characters into a grid of black and white modules, then adds Reed-Solomon error-correction codewords. When scanned values contain errors, that redundancy helps identify damage and reconstruct the original data blocks.
Creators can choose four correction levels: L, M, Q and H. DENSO WAVE gives approximate codeword recovery rates of 7, 15, 25 and 30 percent respectively. Stronger correction improves resilience but leaves less room for actual content at the same symbol size.
The large squares in three corners are position-detection patterns rather than message data. A camera uses them to determine the code's orientation and scale quickly. The clear border and smaller alignment structures also help locate the grid correctly.
This does not mean covering 30 percent will always work. Damage to essential patterns, blur, weak contrast or a missing quiet zone can cause failure even when the affected area looks small. The percentages refer to codewords, not simply a fraction of visible surface area.
A QR Code's resilience comes from redundancy working together with structural landmarks. For reliable codes, limiting content and preserving adequate size, contrast and clear margins matter as much as choosing a high correction level.
