Symbol readers are ubiquitous in modern society; they are used in applications ranging from product barcode scanning to passport data reading. Conventional readers employ a camera (sometimes in combination with a light source) to read the symbol. This approach provides unsatisfactory results when the symbol is printed on a curved or shiny surface, or in low contrast situations. Furthermore, for optimal reading results the printed surface has to be located at a specific distance from the lens (the focal distance).
What it does
We discovered a method that enables accurate symbol reading in a wide range of environmental conditions. Testing our prototype showed that the device can read symbols printed on curved, angled, or shiny surfaces. Moreover, it proved to preserve the symbol's geometrical features, thus enabling robust automatic processing.
How it works
Fig. 1 shows the reader: light bounced of a surface is recorded by a “telecentric lens” attached to a CCD camera. The telecentric lens is the heart of the device: it corrects optically for perspective effects, such that object images remain the same size independent of the distance from the lens. The use of this lens ensures that the device can capture undistorted images of symbols printed on angled or curved surfaces. This also means symbols on shiny surfaces can be imaged at an angle to avoid glare. Another advantage of this type of lens is that it has a focal range as opposed to a focal distance. Therefore, the device can properly read symbols within a wide range of distances (3-12 inches in the prototype). Lastly, the sample is directly illuminated and the reflected light directed to the lens via a beam-splitter, optimising contrast.
Fig. 1: Schematic representation of the symbol reader. The light source (12) illuminates the printed symbol (18) through the beam splitter (16). The light reflected from the target (18) is directed by the beam splitter (16) into a telecentric lens (24) attached to a CCD camera (28).