Tevian and DIXY Test Automated Price Tag Checks Using a Robot

Tevian and DIXY conducted a pilot to automate price tag checks on the sales floor. The Shelf Audit retail space recognition system was integrated with a PUDU cleaning robot: as the robot moved through the store, the system captured shelf images, recognized information on price tags, and used it to compare the data against the product matrix.

The main goal of the pilot was to determine whether price tag compliance checks could be automated after daily price changes. The robot moved through the sales floor while the Tevian system processed the captured images and recognized product names and prices.

During preparation for the tests, the team also adjusted the robot’s movement scenario. To obtain images of sufficient quality, the robot needs to stop near the shelves, as images captured while moving are not always suitable for recognition. During the pilot, the system selected frames captured while the robot was stationary.

The prototype successfully passed testing. Following the tests, DIXY noted the high accuracy and reliability of the video analytics and recognition system.

Further development of Shelf Audit involves working not only with price tags but also with images of the products themselves. This will make it possible to determine product availability on shelves, detect empty spaces, and verify whether a product matches the price tag placed next to it.

Following the tests, the teams identified further development areas, including product recognition and increased autonomy of the robotic platform.

More details about the project are available on the case study page.

A technical breakdown of the implementation and the team’s experience is available in Tevian’s Habr article.