types of Automatic Identification
The traditional way of entering data into a computer is through the keyboard. However, this is not always the best nor the most efﬁcient solution. In many cases automatic identiﬁcation may be an alternative. Various technologies for automatic identiﬁcation exist, and they cover needs for different areas of application. Below a brief overview of the different technologies and their applications is given.
In systems for speech recognition, spoken input from a predeﬁned library of words are recognized. Such systems should be speaker-independent and may be used for instance for reservations or ordering of goods by telephone. Another kind of such systems are those used to recognize the speaker, rather than the words, for identiﬁcation.
This kind of identiﬁcation is used for instance in connection with toll roads for identiﬁcation of cars. Special equipment on the car emits the information. The identiﬁcation is efﬁcient, but special equipment is needed both to send and to read the information. The information is also inaccessible to humans.
By the use of a TV-camera objects may be identiﬁed by their shape or size. This approach may for instance be used in automatons for recirculation of bottles. The type of bottle must be recognized, as the amount reimbursed for a bottle depends on it’s type.OCR – Optical Character Recognition
Information contained in magnetic stripes are widely used on credit cards etc. Quite a large amount of information can be stored on the magnetic stripe, but specially designed readers are required and the information can not be read by humans.
The bar code consists of several dark and light lines representing a binary code for an eleven-digit number, ten of which identify the particular product. The bar code is read optically, when the product moves over a glass window, by a focused laser beam of weak intensity which is swept across the glass window in a specially designed scanning pattern. The
reﬂected light is measured and analysed by a computer. Due to early standardization, bar codes are today widely used and constitute about 60 % of the total market for automatic identiﬁcation.
The bar code represents a unique number that identiﬁes the product, and a price look-up (PLU) is necessary to retrieve information about price etc. The binary pattern representing the barcode takes up much space considering the small amount of information it actually contains. Also, the barcodes are not readable to humans. Hence, they are only useful when the information can be printed elsewhere in a human readable form or when human readability is not required. Laser-scanning of barcodes is therefore only in a few cases an alternative to optical character recognition.
Printing in magnetic ink is mainly used within bank applications. The characters are written in ink that contains ﬁnely ground magnetic material and they are written in stylized fonts which are speciﬁcally designed for the application. Before the characters are read,
the ink is exposed to a magnetic ﬁeld. This process accentuates each character and helps
simplify the detection. The characters are read by interpreting the waveform obtained
when scanning the characters horizontally. Each character is designed to have its own speciﬁc waveform. Although designed for machine reading, the characters are still readable to humans. However, the reading is dependent on the characters being printed with magnetic ink.
Optical Mark Reading.
This technology is used to register location of marks. It may be used to read forms where the information is given by marking predeﬁned alternatives. Such forms will also be readable to humans and this approach may be efﬁcient when the input is constrained and may
be predeﬁned and there is a ﬁxed number of alternatives.
Optical Character Recognition.
Optical character recognition is needed when the information should be readable both to humans and to a machine and alternative inputs can not be predeﬁned. In comparison with the other techniques for automatic identiﬁcation, optical character recognition is unique in that it does not require control of the process that produces the information