What is automatic recognition? (For beginners)
The term “Auto-ID (Automatic Identification)” is not familiar to the general public, and even among those in related industries, there are many people who know the term but are not familiar with its definition.
However, the technology encapsulated in this term is active everywhere in our daily lives, and the scenes where it is used have already become a familiar sight for everyone.
As proof of this, or perhaps because it is a technology that has become widespread and common, JIS (Japanese Industrial Standards) has definitions and regulations not only for the specifications of the technology, but also for the terminology.
According to “01.01.38 Automatic identification system” in X 0500:2020, “A system for the accurate and unambiguous identification of labels, tags and electronic tags containing data, or of their original characteristics and predetermined characteristics. The data or characteristics described are used to identify the label, tag and electronic tag. The data or characteristics described can be looked up by appropriate sources in the system.
For those of you (including me) who are wondering what the heck is going on, JAISA (Japan Automatic Identification Systems Association), an industry organization for automatic identification, defines it a little more clearly. According to JAISA, it is “the automatic capture and recognition of data from barcodes, magnetic cards, RFID, etc. by devices including hardware and software, without human intervention.
I’m sure there are many people who are familiar with the phrase “barcodes, magnetic cards, RFID data, etc.” or who have a vague idea of what it means.
Yes, supermarket cash registers, credit cards, cash cards, IC cards such as Suica and ICOCA, and electronic money are all based on automatic recognition technology.
In fact, it is rare to find a person who has never been exposed to automatic recognition.
In the following, I’ll explain about automatic recognition from a more bird’s eye view, but before that, I’d like to share some of my latest knowledge.
Nowadays, as image recognition has become widespread, automatic recognition is called AIDC (Automatic Identification & Data Capture) instead of Auto-ID. In fact, the title of the aforementioned JIS X 0500:2020 is also “Automatic identification and data capture technology – Terminology”.
Well, I’ll explain about that in the future.
Types of Automatic Identification Technology
So far I have talked about automatic recognition in a way that is difficult to understand, mainly based on definitions, but now I would like to talk about what exactly automatic recognition technology is.
The current automatic identification technologies can be broadly categorized as follows. These are barcodes, RFID, biometrics, magnetic stripes, OCR, and machine vision.
The following is a general introduction to these six types. (More details in a separate article)
I'm sure you are relatively familiar with this one. When most people hear the word "barcode," they probably think of the JAN code. It is a vertical stripe with thick and thin stripes mixed in. It is printed on various daily necessities and foods, and when you go shopping, the clerk at the cash register reads it and displays the amount of money, right? Or you may have imagined QR codes.
We generally call vertical stripe pattern like JAN code and square speckled pattern like QR code as barcode, but in automatic recognition technology, they are completely different code systems in detail, although they are the same optical reading media.
One-dimensional symbol (linear bar code)
This is the so-called bar code. It is said that there are more than 100 types of barcodes in the world, including the aforementioned JAN code, ITF, NW7, etc. Of these, five types, ITF, CODE39, EAN/UPC, CODE128, and GS1 data bar, have been standardized by ISO as international standards. The JAN code conforms to the EAN/UPC standard (the Japanese JAN is called EAN internationally), so of course it is a global standard.
In Japan, in addition to these five types of barcodes, NW7 (CODABAR) is also a JIS standardized system.
Two dimensional Symbol
Two-dimensional symbols are information carriers (data carriers) with data in both horizontal (X-axis) and vertical (Y-axis) directions, such as QR Code, which we are all familiar with. They can be read by a machine (reader) by optically scanning two dimensions. PDF417, QR Code, DataMatrix, MaxiCode, GS1 Composite Symbol, MicroPDF417, and Aztec Code are the seven types of international standard 2D symbols, and in Japan, PDF417, QR Code, DataMatrix, and Micro QR Code are the four types of JIS standard codes. In Japan, four types of 2D symbols, PDF417, QR Code, DataMatrix, and Micro QR Code, have been established in JIS.
QR Code, of course, is a matrix type symbol. QR Code is, of course, a matrix-type symbol. The amount of data that can be stored as information is 10 to 100 times larger than that of a barcode, and it also has an error correction function that allows it to be read again even if the symbol is damaged (and the error correction level can be set in several levels! ).
2. RFID( Radio Frequency Identification )
An RF tag is an automatic recognition system that uses radio waves to read and rewrite information stored in an IC chip from an information carrier (data carrier) equipped with an IC chip and antenna.
There are three types of classification based on the frequency band of the radio waves used, from the lowest to the highest: LF (Low Frequency: medium wave band), HF (High Frequency: short wave band), and UHF (Ultra High Frequency: very short wave band).
The five frequency bands standardized by ISO/IEC are "135 KHz or lower," "13.56 MHz," "433 MHz," "860 MHz to 960 MHz," and "2.45 GHz." In Japan, the LF band (135 KHz or lower), HF band (13.56 MHz), UHF band (920 MHz band within 860 MHz to 960 In Japan, however, the use of the LF band (135 KHz or lower), HF band (13.56 MHz), UHF band (920 MHz within 860 MHz to 960 MHz), and UHF band (2.45 GHz) is permitted.
Due to the differences in the wavelengths of the radio waves, the communication distance, the degree of diffraction (wrap-around) to shielding objects, and the degree of attenuation caused by moisture and water droplets (moisture absorbs radio waves) differ.
However, it is important to note that the UHF band, whose use has recently been accelerating, has different bandwidths and channel widths that can be used within the 860MHz to 960MHz band according to the Radio Law of each country.
Another classification of RFID systems is whether the information carrier (data carrier) has a battery or not. RF tags with a battery are called active tags, while those without a battery are called passive tags. Active tags have a built-in battery, so they transmit their own radio waves, allowing them to communicate over longer distances in the same frequency band, but they must be managed so that they do not run out of battery power, and they must be maintained. Passive tags, on the other hand, use radio waves (electromagnetic waves) emitted from the reader as an electromotive force for communication, so the communication distance is shorter, but the tag itself can be operated semi-permanently without maintenance.
As the word implies, identification by biological recognition. Biometric identification is based on the characteristics of an individual organism, such as fingerprints, palmprints, retina, iris, veins, and voice. It is often used to authenticate individuals using their own biometric data, and we (Asterisk) are not currently working on it, so this is the extent of our introduction.
4. Magnetic Stripe
Magnetic stripes are an automatic recognition technology that reads and writes information stored in the black stripes familiar to us from credit cards and cash cards. Magnetic stripes, such as railroad commuter passes and boarding passes with the entire back side coated with black or brown magnetic material, and telephone cards with a silver back side, which we rarely see nowadays, are also practical applications of this technology.
Since information is read and written by magnetic changes, it is important to note that stored data is easily lost due to the degree of coercive magnetism. There are probably many people who have had the misfortune of being unable to use their cards because they were placed together with or near something that emits magnetic force.
4. OCR ( Optical Character Recognition )
Optical character recognition technology. It reads and recognizes printed or handwritten characters and numbers directly by optical means. Some of you may "know" or "use" the ones installed in scanners and multifunction devices.
In actual business applications, it is used to read the machine-readable zone (MRZ) of passports, for example, or to check the expiration date of perishable goods or stockpiles. However, the current state of optical character recognition technology has yet to achieve a high level of literacy due to the large individual differences in the recognition of handwritten characters and numbers.
It is also difficult to reduce the misreading rate of printed characters to "0", and in many cases, OCR fonts, which were developed to eliminate misreading, are used for optical character recognition.
6. Machine Vision
The processing of images taken by a camera, etc., and of captured images and video data by a computer. In general, it is often called computer vision.
In recent years, with the evolution of AI and neural networks, as well as the deepening of machine learning and deep learning methods, the accuracy of this technology has increased remarkably, and it is being put to practical use in a variety of situations.