TL;DR
- During a single lift lasting a few seconds, D4T quickly analyses more than a hundred images of the waste falling into the hopper.
- The AI computer on the truck turns those images into one lift record: what was in the container, when, where and at which customer.
- That record appears shortly afterwards in the Daisy Portal, with images as evidence.
A lift takes a few seconds. In that time, more happens on a refuse truck with D4T than you might think. Camera images become a lift record, that record becomes a report and that report becomes a conversation with the customer. We follow one container from the kerb to the dashboard.
Take as an example a wheelie bin with commercial residual waste at a bakery.
The first seconds: the container tips
The driver hooks up the container and the lift tips it over the hopper. At that moment the industrial camera above the loading opening is filming. The lighting ensures a clear image, even when it is still dark outside.
While it falls: 40 images per second
As the waste slides down, the AI computer analyses 40 images per second. If the lift takes three seconds, that comes to 120 images. Each image shows a different moment: first the top layer, then what lay underneath.
In each image the model looks for three things:
- which waste fractions are visible, such as cardboard, film, PMD or glass;
- whether there is material in it that does not belong in this stream;
- whether there is a hazardous item in it, such as a gas cylinder or a battery.
In our example, D4T sees a lot of cardboard and some plastic film among the residual waste. That is recyclable material now heading for the incinerator.
After the lift: one record per container
The AI computer combines the analyses of all images into one lift record. This happens on the truck itself, without any video having to go to the cloud. The record contains:
- the time and location of the lift;
- the link to the collection, the customer or the address;
- the recognised fractions and non-conformities;
- the images of every detection.
If there was a hazardous item in the container, the driver immediately hears an audio signal in the cab.
On the way: synchronisation with the Daisy Portal
Via 4G, 5G or wifi, the truck sends the records to the Daisy Portal. There all lifts of the round come together. A planner sees on a map where the truck has been and which containers stood out.
The next day: from record to conversation
The account manager opens the bakery's overview. He sees that their residual waste contained cardboard at almost every lift over the past month. With the images to hand, he calls the customer. Perhaps the bakery needs a cardboard container, or a shorter explanation for its staff.
Since 1 January 2026, businesses in Flanders have paid for their residual waste based on the weight actually presented, and the environmental levy on incineration has roughly doubled. Read in What contamination really costs why such a conversation saves money for both parties.
Why this matters across thousands of lifts
One container says little. Together, the fleet with D4T analyses around 1.75 million lifts per year. In commercial waste, D4T measures on average 40 to 50 per cent recyclable material. On the impact page you can read what that means for recycling and CO2.
Want to know how the five steps work together? See how D4T works or request a demo.