Marmon's Use of Computer Vision and Video Analytics to Enhance Manufacturing Efficiency

Marmon Holdings, Inc., a part of Berkshire Hathaway Inc., is a global industrial organization that comprises more than 100 autonomous manufacturing and service businesses. Each business entity within Marmon operates independently within a group structure, which provides access to the expertise of other businesses with related products and services, or that serve the same customers or markets. Marmon manages over 400 manufacturing, distribution, and service facilities, and employs more than 20,000 people worldwide.

Marmon Holdings, a global industrial organization with over 100 autonomous manufacturing and service businesses, was facing a challenge in optimizing its manufacturing appliances at the McKenzie Valve Plant. The company sought to integrate a Computer Vision (CV) solution to analyze and enhance the use of its machines. The primary goal was to gain more visibility into the run versus idle time of the machines, thereby eliminating production bottlenecks and improving return on investment (ROI). The challenge was to achieve this while ensuring low implementation costs.

In collaboration with Provectus, Marmon's IT team developed a CV solution capable of detecting the color of indicator lights on the machines and sending notifications when the colors change via the IoT protocol. The video feeds from the machines were processed on the AWS Panorama appliance, an edge device designed for Machine Learning (ML) workloads. The solution was designed with scalability in mind, allowing it to be easily implemented across other Marmon Holdings plants. This innovative approach allowed Marmon to collect more data about the usage of its machines and gain visibility into its workforce.

• The implementation of Computer Vision as part of its manufacturing process has enabled Marmon to collect more data about the usage of its machines and gain visibility into its workforce. The CV solution allows Marmon’s operations team to check the run vs idle time for every machine, and to optimize their processes. The solution is ready to be scaled across the network of Marmon Holdings plants, indicating a significant improvement in operational efficiency and potential for future scalability.

• Highly accurate ML model that detects machine utilization

• Critical insights on operations at the facility

• Project and required installations delivered in 5 weeks