AI-based goods recognition, counting and measuring

Scanner-free logistics based on artificial intelligence (AI) saves a considerable amount of time and effort, and increases accuracy of processes

Artificial intelligence in use in real world logistics applications – this is no longer a dream of the future, but already fully operational and implemented in many facilities around the world.

Whether you want to measure the master data of your products quickly and easily, or want to use a 100% check – including suggestions for packaging materials during the packing process without the hassle of barcode scanners – the W2MO identification workstation offers all of this. Even automatic video documentation is included.

Objects are recognized by using existing codes or purely on the basis of their appearance. For the latter, you can easily teach your own products using our self-training functionality.

Artificial intelligence also brings considerable added value in the detection of larger objects, e.g. pallets. Whether as a drive-thru recognition gate or onboard the forklift, our AI not only reads the visible barcodes, but also carries out a consistency check to see whether labels have been read for each container.

The identification workstation can usually be integrated very quickly and easily via a USB interface, e.g. as a keyboard simulation. With this approach, virtually no support from your IT department is required and the integration is basically "plug and play". Of course, interfaces via RestAPI, SAP integration or other technical options can also be implemented.

Despite the use of cameras, privacy is protected. People are not filmed – no personal data is recorded either. In all applications mentioned below, the GDPR is fully considered.

AI-Based Master Data Acquisition

One of the biggest problems in logistics is missing master data. Virtually every company complains that information about product dimensions and weights is incomplete or incorrect. The costs of these deficits are enormous. Deliveries are often delayed or cancelled.

Using conventional 3D scanners is too time consuming. Handling each product takes a long time. Manual measurement is not a solution either; it also takes a lot of time, plus is prone to errors – the measurements could be determined incorrectly or entered into the system incorrectly.

Logivations AI-based master data collection is simple, fast, and can be integrated into existing workstations. Identification of products by scanning a code and the automatic recording of product images are included as standard.

AI-Packstation

Packing shipments is often a time consuming and labor-intensive process. The products must be identified, they must be properly packed in the box and documentation may be required to avoid the customer later contesting acceptance of some or all of the products shipped. All of these functionalities are now available within the W2MO AI Packstation!

The products can be identified by scanning a barcode, QR code or data matrix code. For products without such a code, recognition based on appearance can be trained.

Recognition of Products by Appearance – Training by the User

In order to distinguish products that do not have any printed code, the W2MO identification workstation uses recognition based on their appearance. For this purpose, an artificial intelligence – a neural network – is trained. In addition to appearance, this also uses information about the size of the object. This means that even very similar objects, such as screws of different sizes, can be distinguished.

A user can teach the system to recognize new products. For this purpose, images of the object are taken from different perspectives. A simple dialog guides the user through this data acquisition. Once the data has been recorded, further products can easily be added later. Data acquisition is only necessary at one workstation and can be then reused at any number of others.

Automatic counting of products

Counting is a frequently required and time-consuming task in logistics. Most counting processes can be automated by W2MO.

If the products are not touching each other at the time of counting or can be separated, this can be done without further training of the system. For specialized cases, the respective goods or their types can be trained and then the system can count the learned and visible goods in milliseconds. At the same time, QR codes or barcodes on the objects can be read and processed, layer packing schemes can be identified, or the number of read codes can be validated via the counting result.

Identification gate or goods recognition by onboard cameras

How much time is wasted in logistics when forklift drivers scan the goods on a pallet they are picking up? Or when workers have to identify – pallet by pallet – what has arrived in the receiving area? Two innovative solutions are possible with W2MO AI-based goods detection, counting and measurement: