Video as a sensor has been a topic for several posts here at The Manufacturing Connection. One notable case study involved detecting unwanted critters wandering into the facility when gates were opened to allow train cars to enter or leave. Machine vision has been usable since the 1980s to detect flaws, presence, and other quality issues, as well as to guide robots.
Anurag Maunder, CEO and founder, and Subbu Kuchibhotla, VP Growth and Development, of a new company in the video streaming market called Sensable.
They told me this is the first vision platform built for industrial engineering. Almost all current vision and video applications involve narrowly focusing on a part or a piece of a machine. The idea of the Sensable platform is to broaden the focus of the camera, or combine multiple camera, such that an entire operation or segment of the plant can be viewed, captured, and analyzed.
My grandfather told me of the time he was summoned to the front office of the GM plant where he was a production superintendent. The US had entered World War II and his plant was converting to production of aircraft armaments. Production ramp up was slow. Management picked him to organize things and get production up to expectations.
He explained to 8-year-old me how he went up on the mezzanine and watched the process. Guys were performing a process, dragging a crate of parts to the next operation across the department, where the next operation took place, and that operator dragged the crate across the facility to the next operation. He told me how he organized the process to minimize material handling. That and other things boosted production and won the war—well, anyway, he did his part. And I learned a lesson.
Now imagine that you don’t have people to just stand and observe and take notes over three shifts a day for a week or so. What if you could position a few cameras in strategic locations. The video is captured and run through analytics. Engineers, operators, and managers would not have to manually parse through hours of video. They would be presented with data visualization designed to help them get to root causes of problems, assist worker ergonomics, improve safety, and boost productivity.
That is what the Sensable solution does.
Imagine another scenario. You are an operator on a production line. You have been trying to point out bottlenecks to production on your machine. Then engineers install streaming video pointing not just at a specific point on you or the machine but with wide enough scope to see the larger process. The video analytics point out the bottleneck. Voila. Vindicated. Proof in the data.
The video is not for spying on employees. It is designed to help them. Just what true digital transformation is—an aid to decision making and continuous improvement.
- Missed throughput targets—station utilization lower than expected, unplanned downtimes more than planned
- Low process efficiency—cycle time variability, too many interruptions
- Low operations visibility—safety challenges due to best practices violations, missed inspection or assembly steps
- Manage work area or assembly line—real-time feedback, identify bottlenecks, performance reports by shift/day, remote visibility-ideal for managing off shifts
- Perform long duration time studies—data-driven Kaizen setup/changeover analysis, run/analyzed over weeks, compare across time and facilities, store metrics for Kaizen, perform SMED analysis in large areas
- Identify missed inspection steps with 360 degree analysis—rapidly identify root cause of defects, search for video clips associated with product assembly
- Achieve healthier, safer, well trained workforce—capture near misses and best practice violations, capture the impact of fatigue by measuring throughput at beginning and end of shift, capture and share the best practices for training
- Build realistic engineering standards—capture data for the entire shift or multiple shifts before creating a standard to be enforced