Revolutionize the quality control of the production line.

Controlling the output of the production line has always been a complicated operation. We have developed technology that allows you to do this in an automated manner using Artificial Intelligence.

Revolutionize the quality control of the production line.

Controlling the output of the production line has always been a complicated operation. We have developed technology that allows you to do this in an automated manner using Artificial Intelligence.

What's ADR-Flow?

ADR-Flow is the automatic defect recognition system for production lines.

What's an automatic defect recognition system?

Play Video

The problems of
Quality Control

Quality control is a crucial phase of every production process. Sometimes it is possible to identify a specific moment in which it occurs, but now more and more often it is an operation that constantly accompanies the production of the product.

How to carry out quality control, how often, at what times and with what standards requires making some important choices for the line, which consequently affect production in terms of waste or slowness of the process.

Some of the classic quality control problems are:

High costs

Dedicating resources to quality control has a cost that affects the final product.

Missing standards

Depending on who is doing the testing and their state of alertness, there could be very different results.

Data interpretation

Despite the many data collected, it is often almost impossible to use them to prevent or improve the line.

What we propose

ARTIFICIAL INTELLIGENCE

ADR-Flow is software that, once trained, uses artificial intelligence to autonomously recognize defects.

REAL TIME DETECTION

It quickly detects products that do not meet production standards and reports them in real time.

ANALYSIS AND STATISTICS

Show the results of your measurements on a dashboard and create customized data analyses.

QUALITY IMPROVEMENT

The final result is a clear improvement in line quality, with fewer resources used.

What do we do?

1.
We evaluate the manufacturing process at this time
2.
We install the necessary monitoring equipment
3.
We teach ADR-Flow to recognize defects
4.
We provide a dashboard with real-time view of controls

CAR PARTS

Surface inspection of spark plugs, injectors, larger components or final assemblies.

ASSEMBLY LINE

Verification of correct wiring and correct assembly of motorcycles.

What types of defects can we detect?

We can detect all defects that are visible via a photo or video. ADR-Flow and its Machine Learning algorithm instantly analyze the image flow and determine the quality of the product in real time. ADR-Flow can detect: assembly precision, roughness, distances, concentricity, positioning, wiring and much more.

WELDINGS

Checking the correct weldings of the components of a frame

ACCURACY

Verification of the assembly accuracy of multiple components.

How bulky is ADR-Flow?

ADR-Flow has the excellent characteristic of being compact and above all of being able to be implemented on any production line. It adapts, without any type of constraint, to older production lines and the installation is completely customized, based on production and site access needs.

Percentage of
defects recognised

Each system has its own peculiarities, its own strong points and its own weak points, but at the end of the analysis you have to choose based on the data.

The data speaks clearly and says that an automatic defect recognition system, led by an artificial intelligence that uses neural networks, is absolutely the best solution (as well as the cheapest).

Human
1 %
AI
1 %
Hybrid systems
1 %
technique
How artificial intelligence works

The functioning of an artificial intelligence is very interesting. It uses the most advanced technologies in the field of neural networks and machine learning.

As a first step, we need to define which defects the neural network will have to recognize and position the cameras in such a way that they can shoot at exactly the right moment.

The neural network will initially need to be trained. In practice it is very simple: you need to show ADR-Flow a certain number of images of the component and tell it if there is a defect within each image or if the photographed component is fine.

This operation is carried out by an expert operator together with the ADR-Flow team.

Once ADR-FLow has learned to detect defects, the neural network will be implanted directly inside the cameras installed on site. In this way there is no communication latency and defect recognition is immediate.

The neural network, as it works, improves its own ability to find errors, because it constantly learns from the new images it processes.

Our strength is numbers

In the field of quality management we know that the only thing that matters are the numbers. Every process has a numerical result and what interests us is that our technology can improve the numbers.
Precision
1 %
sensitivity
1 %
specificity
1 %
accuracy
1 %
1 %

/

1 %

PRECISIONE

1 %

/

1 %

SENSIBILITÀ

1 %

/

1 %

SPECIFICITÀ

1 %

/

1 %

ACCURATEZZA

FAQ

Frequentl Asked Questions

The system uses machine learning algorithms to analyze images and detect defects. Training is performed on a dataset that contains images of defects and images of objects that show none, resulting in immediate and precise recognition of defective objects.

If a defect can be identified visually ADR-Flow, appropriately trained, is able to detect it.