Detecting pallet pockets using a 3D snapshot camera

Hitting the heights with 3D plug and play machine vision solutions: The AGVs of the manufacturer Transolt move about its customer’s warehouses as if by magic. There’s not really anything supernatural at play here, of course, but automated guided vehicle systems do possess a certain kind of technological magic. A magic to which solutions from SICK contribute. Automated guided vehicles need to be able to see to perform their work reliably. Safety issues also need to be taken into account, especially when large loads are being moved about.

SICK has numerous tried-and-proven laser scanners for mobile platforms in its product portfolio. For autonomous forklift trucks, there’s more to it than merely moving things. In this case the eyes need to detect the location of pallet pockets even at significant heights, and precisely control the storage and retrieval operations. An ideal task for the Visionary-T AP 3D snapshot camera and the Key App that SICK has developed especially for this application.

Transolt had been waiting a long time for this: A solution that allows automated guided vehicles to not only drive around but also pick up and put down loads with great flexibility.

Transolt Managing Director Mahmut Özkan said at present, autonomous forklift trucks can only perform their tasks in warehouses under the condition that the pallets to be transported are always located at the same angle and within a maximum of 20 mm of their predefined position.

“Real-life conditions in dynamic warehouses are not like this, however. Pallets get inadvertently pushed out of their original positions, stacked on top of one another or damaged, which can prevent smooth transport processes using automated forklift trucks.”

Now thanks to the Pallet Pocket Detection SensorApp running on the Visionary-T AP, he can finally offer his customers in the warehousing and intralogistics industry autonomous forklift trucks as well. The trick: This Key App functions directly on the 3D snapshot camera with no further hardware required. That saves not only costs but also the effort of additional programming and integration.

Pallet Pocket Detection SensorApp sharpens the view

This Key App enhances the ability to see gaps. The autonomous forklift truck needs to precisely detect pallet pockets, displaced pallets, and damage. Being able to precisely scan storage spaces and a variety of pallet types is essential for perfectly storing and retrieving loads — even way up in the storage racks. The Pallet Pocket Detection SensorApp provides the data required to do this. It runs directly on the Visionary-T AP 3D snapshot camera. The measured values required to pick up the pallet are pre-processed and evaluated on the sensor, then transmitted to the control system of the automated forklift truck. The 3D camera and vehicle thereby become an inseparable unit thanks to the SensorApp.

That’s something Mahmut Özkan really appreciates.

“The Visionary-T AP and its key App scan the pallet and determine the precise position to enable our autonomous forklift truck to always insert the forks into the pallet holes at the right angle and to lift them from the same correct depth,” he said.

Özkan also said he was delighted with the precise geometric data of the Visionary-T, as it provides the basis for the company’s newly developed maneuvering algorithm.

“With the help of this algorithm, the autonomous forklift truck can pick up pallets that are not located at their predefined position by recalculating the pick-up angle and maneuvering the forks again into the correct orientation.”

SICK AppSpace — the place for the future

The term Industry 4.0 is often used to describe future technologies. But there are also more concrete examples: machine vision solutions based on the SICK AppSpace eco system are taking a real step towards the future and Industry 4.0. Pallet Pocket Detection is successful proof of a completely new and adaptive solution to an automation application.

It also works equally well for positioning dollies. The Dolly Positioning SensorApp provides the data needed for this purpose. It runs directly on the Visionary-T AP 3D vision sensor. The measured values required to automatically pick up the dolly are pre-processed and evaluated on the sensor, then transmitted to the control system of the automated guided vehicle. Thanks to an active developer community, there are numerous further custom SensorApps based on intelligent software tools and algorithms being created. No other app, however, is hitting the heights at the moment with its concrete application like Pallet Pocket Detection.

Originally published here.