Vision aids removal of objects from conveyor

Vision system uses a matrix and plots the location of the objects in each cell, passing the cell ID to the PLC which in turn fires ejectors at the end of the conveyor to remove the objects.

The power of Scorpion Vision has been proven once again with a new application for the food production industry.

In November, DISL demonstrated a new product which observes the progress of multiple objects along a conveyor belt.

The problem was presented to the Scorpion Vision team in the summer after successful launch of the optical sorting system which removes cardboard from the stream of waste paper in a recycling plant.

This particular application uses a matrix and plots the location of the objects in each cell, passing the cell ID to the PLC which in turn fires ejectors at the end of the conveyor to remove the objects in those cells.

This application is great for situations where the objects on the surface of the conveyor are static, not moving from cell to cell.

But what about a scenario where the objects are not simply flat pieces of cardboard, are more rounded in shape and roll about, wandering across the field of view - such as potatoes or eggs? In food processing, substandard product needs to be identified and removed from the stream.

With new potatoes this is simply a matter of identifying the green component using a colour analyser.

Scorpion Vision can do this very quickly (in milliseconds), but then once the object is identified it should be marked for ejection.

But ejection happens at the end of the conveyor - possibly as much as 5 metres or 10 seconds away.

In this time period, the offending item could wander right across the field of view, and perhaps as it rolls over, obscure the target colour.

So, once it's been identified, its position needs to be monitored 100% of the time so that when it finally reaches the ejection mechanism, the object is ejected regardless of whether the reason for ejection is not seen again.

Using Scorpion Vision's advanced scripting capabilities, a tracking algorithm was devised which does just this.

Using high performance firewire cameras, Scorpion is presented with new images at up to 30 times per second.

As the objects progress along the conveyor belt, each one is given a unique ID, and marked with a value which if positive, is passed to the ejection system at the right moment in time.

Any unmarked, good product passes on through to be packaged and despatched.

This works for anything where objects need to be inspected as they turn or roll over - the first application is in chocolate egg production, identifying and rejecting those chocolates that don't have 100% foil coverage.

But of course, this application isn't restricted to just food!.

Back to top