Logistics & Post

Fully automatic mail service

PLANET AI’s address recognition capability detects and reads postal codes, cities and street names, house numbers and names quickly and reliably. Our technology can automatically distinguish between sender and receiver, no matter if machine-printed or handwritten text.

PLANET intelligent systems leverages our technology to develop solutions for the logistics industry.


Follow the process step by step

1. Finding Address Areas

The starting point is a picture of the postal item. Using advanced image processing, areas are found, which contain an address with high probability. These areas are further pre-processed: The text alignment is determined and the area is split into lines, resulting in a list of images of single lines.

Address Finding

2. Intelligent Text Recognition

These images are then processed by PLANET AI’s Intelligent Text Recognition which uses state-of-the-art machine learning and artificial intelligence technologies. The result is our patented PerceptionMatrix for each row – containing all probabilities for the given characters. The confidence values for every line of the matrix are used to correct typing and formatting errors. Thus, PLANET AI’s technology gives post and parcel customers a significant advantage compared to classic Optical Character Recognition (OCR).

Address ITR

3. Address Matching

Successful address matching requires a number of essential elements:

  • Rules for describing address components, e. g. “a postal code consists of exactly 5 digits” AND / OR
  • Lists of possible words, e. g. names of places, districts and streets to a postal code

The order of address parts can be flexibly defined with a set of rules. PLANET AI combines rules, databases and the PerceptionMatrix to identify the correct destination address.

Address Matching

4. Result

A list of address data, listing the address with the highest confidence values at the top.

Customs Declarations

At the moment, PLANET AI supports the standard forms CN22, CN23 and CP72.
The following information are extracted:

  • receiver’s and shipper’s address
  • telephone numbers
  • e-mail addresses
  • description of the goods
  • currencies
  • values and total value of the goods
  • weights and total weight
  • purpose of sending goods (gift, document, sample, returned goods, …)
  • HS tariff code

Due to privacy regulations, we had to disguise several address parts.

Click through our demo pictures below.

  • Demo of reading address
  • Demo of reading customs document
  • Demo of reading address
  • Demo of reading customs document
  • Demo of reading address
  • Demo of reading customs document

Supported Countries

Czech flag

Czech Republic

Danish flag


German flag


Luxembourgian flag


Slovenia Flag


UK flag


US-American flag


The implementation of further address systems is easily possible due to the flexible rule-based approach.
Other character systems, e. g. Cyrillic, can also be integrated without any problems.