Case studies

How to Detect Anomalies on a Production Line

  • Client

    Pulp and paper industry

  • Service

    Artificial Intelligence

  • Industry

    Manufacturing

Pulp and paper industry
Pulp and paper industry

Our client is a Canadian leader in the pulp and paper industry. They produce a variety of products, including household and industrial paper, cardboard, packaging products, and specialty papers.

Challenge

One of their factories uses cardboard boxes to package their products. These boxes are transported throughout the plant on numerous conveyors, some of which are difficult to access. During transport, some boxes may fall or end up in a non-compliant position (tilted, overturned, at the wrong angle, etc.). If these misaligned boxes are not detected in time, they can block the conveyor and sometimes force the company to completely shut down their production line, a procedure that can be very costly.

To address the situation, the company had installed a series of sensors along the conveyor to measure the height of the boxes and detect those that had fallen. However, this more rudimentary equipment was unable to recognize boxes that were simply tilted or positioned incorrectly. As a result, the method proved very imprecise and regularly required human intervention. An employee would be tasked with continuously monitoring all the surveillance cameras placed throughout the plant to confirm any anomalies.

Solution

Solution
Solution

The typical approach for this type of problem would be to install a more traditional, but very expensive (up to $30 000), vision system equipped with specialized cameras, lenses and other components to detect invalid boxes.

Our approach is much simpler and, most importantly, far less costly. First, we connected to the existing surveillance cameras already installed in the plant to deploy a vision system. Then, we trained an AI algorithm to detect boxes that were lying down, tilted, or in any other non-compliant position, while ignoring properly aligned boxes. As soon as an anomaly is detected, the plant receives real-time notifications, allowing an employee to quickly reposition the box and prevent downtime further along the conveyor.

Results

The performance of the vision system is significantly superior to that of their previous system. Moreover, the company did not need to purchase any expensive new equipment: we simply connected to the existing infrastructure and made their cameras "smart," allowing them to save both time and money in the long run. The simplicity of this vision system also made it possible to deploy the solution quickly, in about one month in their case.

Expertise

Your one-stop shop for digital transformation.

Services
Services

By combining the best of data, artificial intelligence and web and mobile development. Solutions designed to optimize your processes and accelerate your growth.

Services

Services

Artificial Intelligence

Previous Learn more about Artificial Intelligence

Services

Web + Mobile Development

Previous Learn more about Web + Mobile Development

Services

Data Engineering + BI

Previous Learn more about Data Engineering + BI

Services

Cybersecurity

Previous Learn more about Cybersecurity

Discover how Vooban can transform your projects with innovative technological solutions.

Read more