Reduce costs and invest well with help from computer vision

15 July, 2020
Travis Siegfried
IBM

How can businesses survive and thrive during and after a recession? And can cutting-edge computer vision solutions help?

A 2010 Harvard Business Review (HBR) analysis of 4,700 companies over 3 recessions (in 1980, 1990 and 2000) identifies that only 9% of companies flourish by the end of a recession. Most take an additional 3 years to return to growth. The successful companies applied a balanced strategy of reducing costs in a manner that drives ongoing, sustainable operational improvements while making business development investments to provide greater customer value.

There are many ways AI empowers businesses to succeed (even during COVID-19), and the use of computer vision on manufacturing lines provides a great illustration of how technology can help organizations with cost savings and innovation.

The business case for AI

Digitalization of manufacturing provides visibility into real-time operations, and AI adds the ability to predict and respond to operational disruptions. Applying these techniques can help manufacturers lower operational costs, not just today but into the future. Evolving into a resilient manufacturing operation positions you to better serve your customers and invest in the right things, whether that’s by adding more flexibility in your product mix, personalizing offerings or making design improvements based on product usage in the field.

With AI, you can combine multiple data sources to create a clearer picture of how to proceed intelligently in a general business process. In manufacturing, those data sources might be heat detection sensors, indoor/outdoor weather data, PLC data from a machine or conveyor system, or visual data from a camera. You can then use AI to gain insights into the integrity of a weld or to predict potential product failures for warranty purposes. AI can also help with re-manufacturing to reuse parts and materials, or with designing new product possibilities.

Building an AI model for manufacturing lines

I work with various industrial companies that design and build specialty trucks, military vehicles, truck bodies, and airport, fire and access equipment. These businesses find that they’re spending large quantities of money on warranty repair due to failures during the product production process, and they’re looking for ways to more proactively address potential manufacturing line issues.

Some of the most expensive and potentially life-threatening repairs in the manufacturing process are welds and part assembly. This where object detection technologies can help, by using images from cameras on conveyor belts and manufacturing lines to detect anomalies. Computer vision technology can help businesses more easily deploy deep learning vision models for challenges like this. Here’s how it works: First, we work with experts on the assembly line to train a computer vision model using deep learning technology to recognize defects. Once the model is trained, we set up a camera and inference against that model. The manufacturer gets real-time alerts so that they can remediate problems early in the engineering process, before it’s too late and materials are damaged.

Organizations can recover significant revenue if their manufacturing facilities introduce computer vision on manufacturing lines. IBM offers one such solution using IBM software and hardware with deep learning and proven analytics algorithms. Solutions like this can help your business survive and thrive even in difficult times by investing in high-quality products and reducing the amount of waste and rework.

Where to find support for your AI project

If you’re interested in seeing how computer vision solutions could help your business, please contact us at the IBM Artificial Intelligence Center of Competence. We can help you narrow down and define your immediate needs and see how IBM technology can enhance your enterprise with AI.

The post Reduce costs and invest well with help from computer vision appeared first on IBM IT Infrastructure.