How Do QR Codes Work? The Squares That Store Links and Text

They are on restaurant menus, train tickets, parcels, posters and payment terminals. Point your phone camera at one of these pixelated squares and, in a fraction of a second, a link or a message appears. A QR code looks like random noise, yet every square has a job.

Born in a car factory

In the early 1990s, Japanese car factories used traditional barcodes to track parts. But a barcode stores only about 20 characters, and workers had to scan many labels per box. In 1994, engineer Masahiro Hara at Denso Wave, part of the Toyota group, developed a two-dimensional code that could hold far more data and be read very quickly. He called it QR, for Quick Response.

Crucially, Denso decided not to enforce its patent, letting anyone use and create QR codes freely. That openness, along with phone cameras capable of reading them, helped QR codes spread around the world, and their use exploded during the COVID-19 pandemic for contactless menus and check-ins.

The anatomy of a QR code

• Finder patterns: the three large squares in the corners. Their pattern of nested squares is easy for software to spot in any image, and their positions tell the reader which way up the code is. Because there are only three, the empty fourth corner shows the orientation.
• Timing patterns: lines of alternating black and white modules between the finder squares. They let the reader work out the size of the grid and the position of every module, even if the code is slightly distorted.
• Alignment patterns: smaller squares, in larger codes, that help correct for curved or tilted surfaces.
• Data modules: the rest of the grid. Each module is one bit: dark for 1, light for 0. Groups of bits encode numbers, letters, symbols or Japanese characters.
• Quiet zone: the blank border around the code, which separates it from surrounding text and images.

Reading a code

When you point your camera at a QR code, software in the phone looks for the three finder patterns, straightens the image, maps the grid, reads each module as a 0 or 1, and decodes the result. If the data is a web address, the phone offers to open it.

Error correction: why damaged codes still work

QR codes include extra data calculated with Reed–Solomon error correction, the same mathematics used on CDs and in deep-space communication. There are four levels. At the highest, level H, up to 30% of the code can be dirty, torn or covered and it can still be read. That is why designers can place logos in the middle of QR codes.

Be careful what you scan

A QR code is just data; it cannot hack your phone on its own. But it can send you to a fake website. Scammers sometimes stick false codes over real ones on parking meters or posters, a trick called “quishing”. Before entering passwords or payment details, check that the web address is the one you expect.

In short

• QR codes store data as a grid of black and white modules, each a binary bit.
• Three finder squares let cameras locate the code and its orientation instantly.
• Timing and alignment patterns help map the grid, even on curved surfaces.
• Error correction lets codes work even with up to 30% damaged.
• They were invented in 1994 by Masahiro Hara at Denso Wave to track car parts.
• Always check the web address before entering personal or payment details.