Several projects involved in recent work all require online products to more or less perceive a large number of offline items / goods. Warehouse management requires the system to sense incoming goods, vehicle management requires the system to sense incoming and outgoing vehicles, automatic cash collection requires the system to sense goods to be paid, and document management requires the system to sense paper documents.
Perceived items are very important for products in the fields of logistics, warehousing, cashiering, and the Internet of Things. Offline items are separated from online systems. The system must be able to sense / collect the item information to continue the subsequent process.
Since you need to perceive an item, the item must first have a logo. The most original label is to write important information about the item with a paper label and paste it on the surface of the item. We will call this label label 1.0 for the time being.
The problems with this traditional label are also obvious: one is that the label is easily damaged, leading to the loss of information; the other is that it is difficult to perceive the item by means of informationization. Of course, there is no way out. We will talk about OCR compatibility later Kind of traditional paper label.
Label 2.0: barcode / two-dimensional code
The use of bar codes / two-dimensional codes has become popular, and we use them almost every day. The bar code and the two-dimensional code both use black and white colors to represent binary 0 or 1, so that the graphics can be converted into binary information of 0101, plus appropriate decoding rules to represent the corresponding information.
A barcode can be understood as a line of a QR code, so it can carry less information than a QR code. Bar codes can generally only store tens of bytes, and two-dimensional codes can store hundreds of bytes of information. Moreover, the two-dimensional code is more robust than the barcode, and can still be identified even if a small part is damaged.
The [asset name] [asset number] [specification model] [department] information in the fixed asset card above generates a barcode or two-dimensional code, and enters the era of labeling 2.0. Friendly labeling bar code / two-dimensional code labels and traditional The labels are used together, and both the human-readable text information and the machine-readable two-dimensional code information are written on the pasted labels.
Scanning codes with tools such as code scanning guns and mobile phones can obtain corresponding information, and it is also very convenient to directly write information into the database of the system, and then the functions of outbound, inbound, and inventory of fixed assets can be realized.
Tag 2.0 has greatly improved its interaction with information systems compared to 1.0, but requires a code scanning gun or a handheld device as a collection device, and can only identify a single item at a time (imagine that the cashier needs to place items in order when the supermarket checks out Scanning scenes), the efficiency is not particularly high.
Tag 3.0: RFID
I think RFID is the closest way to users' insensitivity. Although I have known RFID for a long time, I haven't actually touched it before. Until I went to Decathlon and found that the automatic cash register system used RFID technology, the experience was very good. When you check out, you only need to put the products into a checkout basket one by one, and the name and amount of the product will be displayed on the display. Then you can scan the code to check out.
RFID, short for Radio Frequency Identification. The RFID system requires a Tag tag and a reader reader. The tag is responsible for storing data (analog two-dimensional codebook code), and the reader (analog code gun) is responsible for transmitting radio frequency signals to form an electromagnetic field. After the tag enters the electromagnetic field formed by the reader, it will be Inspire, send your message to the reader.
RFID can store more information than two-dimensional codes, and the extension can reach tens of KB. In terms of sensing efficiency, RFID is also significantly higher than two-dimensional code scanning. It only needs the tag tag to enter the electromagnetic field range to be sensed. There is no need to scan one by one. The disadvantage is that the price of tag tags is relatively high, the price of passive tags is about a few corners, and the price of active tags is about a few yuan.
Compatibility method for label 1.0: OCR
Although there have been better solutions to replace traditional tags, in actual projects sometimes it is necessary to compatible with 1.0 tags left over from history. The method we think of is to use OCR. The content of the label is collected by taking pictures, and then the text in the label is recognized by the OCR model and converted into structured data.
OCR, or optical character recognation, is a small branch of image recognition. OCR's recognition of plain text is relatively mature, and the recognition rate can reach more than 99%. All factories have corresponding services. The format of labels is generally relatively fixed. Through targeted model training, a model for identifying labels can be formed.
The cost of OCR is also relatively high. One is the cost of pre-model training, and the other is that each tag needs to be photographed before sensing the item and then transmitted to the server for identification, which is relatively inefficient. But as a method compatible with traditional labels, it is still more effective.
Summarize the three types of tags: QR codes and RFID have their own advantages and disadvantages, which can be selected according to the requirements of storage capacity, sensing efficiency and cost, and OCR can be an effective way to be compatible with paper tags.