Skip to content

Case study, sorting facility for separately collected recyclables

Trigironde: characterising rejects to inform investment decisions

SPL Trigironde uses analysis of its rejects line to better understand the materials present and explore additional recovery options. NextWaste’s support provides measurement evidence to inform investment decisions.

Olivier Guilmois, Managing Director, SPL Trigironde

34,000 tannual capacity
11 t/hsorting throughput
550,000residents served
7shareholder local authorities
2024commissioned, Saint-Denis-de-Pile (Gironde)

The context

Understanding the line better to explore recovery options

Greyparrot analyser mounted on its frame above the rejects line at the Saint-Denis-de-Pile sorting facility
The analyser and its bracket, above the rejects line.

At the Saint-Denis-de-Pile sorting facility, rejects pass over a picking table in a cabin, where operatives pull out recoverable materials before the rejects are removed. The SPL wanted a better understanding of the materials on this line, in order to explore additional recovery options.

Understanding the waste stream

Identify the material categories present at the observation point, using the local authority’s characterisation grid.

Testing under real site conditions

Compare the automated readings with manual characterisations to establish a reading tailored to the site.

Informing the decision

Document the recovery options before making any investment.

Setting the optical sorters more aggressively is not a solution in itself: unwanted items would end up on the other tables. The real challenge is finding the right balance.

The system

A six-month audit contract, installed in under two hours

An AI vision analyser is positioned above the conveyor carrying large rejects, upstream of the cabin. It observes the waste stream continuously and identifies the items visible on the belt surface. The SPL opted for simplicity: a six-month audit contract with no change to the process, to obtain measurement evidence before any investment decision. Analysis began in June 2026.

Under €500
for the bracket, made by a local craftsman and fixed to the existing conveyor.
Under 2 h
to install, with the power supply and network close by.
Citeo grid
the measurement uses the characterisation categories the local authority already works with.
5 cm mesh
any item smaller than 5 cm is counted as fines, whatever its material, in line with the site’s convention.

The method

Testing AI under real site conditions

Observations are compared with manual characterisations to establish a reading tailored to the site’s categories and conditions.

Time-stamped cross-calibration

For two to three weeks, NextWaste and the site compared the analyser’s readings with manual characterisations. The protocol is simple but somewhat demanding: at a set time, the operatives at the table stop, and everything on the table is collected, characterised and photographed. These results are then compared with what the analyser recorded over the same period, until the gap is minimised.

Reports and alerts

The SPL receives periodic reports, alerts and a count of materials by category. This continuous data complements sample-based characterisations rather than replacing them.

Bringing the operator on board

The site operator was involved from the start. The measurement provides a shared basis for discussion with the SPL, particularly on optical sorter settings. As Olivier Guilmois puts it, “it’s a tool that lets you both understand your process and take action”.

What the measurement provides

A more precise understanding of the waste stream

The analysis highlights the material categories present at the observation point and provides evidence for comparing recovery options.

Testimonial

In Olivier Guilmois’s own words

“There are always investments to be made in a sorting facility; you have to choose the most relevant ones.”

Olivier Guilmois, Managing Director, SPL Trigironde

Next steps

Documenting the options before investing

The results are feeding into the study of additional solutions. No investment has been decided to date.

An option being explored

The analyser could drive an ejector that sends packaging back into the optical sorting loop, subject to modifying a chute.

The role of the operatives

If this option were adopted, the operatives at the table could focus on fibre-based materials.

Value yet to be confirmed

The technical and economic value of these options has yet to be confirmed before any deployment decision.

Reading guide

Limitations to bear in mind

  • The observation point is upstream of manual sorting. On its own, it therefore does not reflect the facility’s final performance.
  • Accuracy differs between an item count and a mass estimate; that is precisely what cross-calibration is for.
  • These results are specific to this site’s configuration and waste stream. Every sorting facility has its own balance.

Want to understand a waste stream better before investing?

Let’s look at the right measurements and validation conditions for your site.

Discuss my project

Other case studies

Testimonials from the NextWaste webinar “IA en centre de tri: retours d’expériences de l’analyse continue” (AI in sorting facilities: lessons from continuous analysis), held in French on 17 September 2026; quotes on this page are translated from French. Translated from the French version updated on 3 October 2026.